Language and Interaction Discussions with John J. Gumperz
This book features a fascinating and extended focal inter...
209 downloads
1198 Views
2MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
AUTHOR ""
TITLE "Language and Interaction: Discussions with John J. Gumperz"
SUBJECT ""
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Language and Interaction Discussions with John J. Gumperz
This book features a fascinating and extended focal interview with Professor John J. Gumperz, who ranges over his long career trajectory and reflects on his scientific achievements and how they relate to the contemporary linguistic scene. In this way, the reader is presented with a snapshot introduction to Gumperz’s work in a contemporary context. A number of commentaries provide a stimulating and illuminating series of theoretical and applied encounters with Gumperz’s work from different perspectives. In so doing, they shed new light on Gumperz’s seminal contribution to the study of language and interaction. In his Response Essay and in a final discussion, Gumperz clarifies his views on many of the topics discussed in the volume, as well as sharing with readers his views on some other approaches to language and interaction that are closely aligned to his own. Sociolinguistics, the ethnographic approach to language, language and social interaction, intercultural communication, communicative conventions, contextualization — these are some of the key terms which Professor John J. Gumperz discusses in this wide ranging and searching interview about his career as an anthropological linguist and sociolinguist interested in cultural diversity and intercultural communication. Contributions by: Aldo di Luzio, Carlo L. Prevignano, Stephen C. Levinson, Paul J. Thibault, Afzal Ballim, Susan L. Eerdmans, and John J. Gumperz.
John J. Gumperz, Professor Emeritus of Anthropology, University of California, Berkeley, is one of the founders of Sociolinguistics whose early work on speech communities and on the relationship of linguistic to social boundaries helped lay the basis for much current work in the field. Since the 1970s he has concentrated on a theory and methods of discourse analysis that can account for the intrinsic diversity of today’s communicative environments. His publications include: Language in Social Groups (1962); Ethnography of Communication (1964) and Directions in Sociolinguistics (1972/2002), both coedited with Dell Hymes; Discourse Strategies (1982); Language and Social Identity (1982); and Rethinking Linguistic Relativity (1996), coedited with Steven Levinson. He is currently working on a collection of studies New Ethnographies of Communication (coedited with Marco Jacquemet); and Language in Social Theory. Photograph by Paul Bishop with kind permission.
Language and Interaction Discussions with John J. Gumperz
Edited by
Susan L. Eerdmans University of Bologna
Carlo L. Prevignano University of Bologna
Paul J. Thibault University of Venice and Lingnan University, Hong Kong
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 Language and interaction : discussions with John J. Gumperz / edited by Susan L. Eerdmans, Carlo L. Prevignano, Paul J. Thibault. p. cm. Includes bibliographical references. 1. Gumperz, John Joseph, 1922- 2. Sociolinguistics. 3. Social interaction. I. Eerdmans, Susan. II. Prevignano, Carlo. III. Thibault, Paul J. P85.G84 L36 2002 306.44-dc21 2002028005 isbn 902722594X (Eur.) / 1588113043 (US) (Hb; alk. paper)
© 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 Preface
vii
Chapter 1 Presenting John J. Gumperz Aldo di Luzio
1
Chapter 2 A discussion with John J. Gumperz Carlo L. Prevignano and Aldo di Luzio
7
Chapter 3 Contextualizing “contextualization cues” Stephen C. Levinson
31
Chapter 4 Contextualization and social meaning-making practices Paul J. Thibault
41
Chapter 5 On Gumperz and the minims of interaction Carlo L. Prevignano
63
Chapter 6 A commentary on a discussion with John J. Gumperz Afzal Ballim
79
Chapter 7 A review of John J. Gumperz’s current contributions to Interactional Sociolinguistics Susan L. Eerdmans Chapter 8 Response essay John J. Gumperz
85
105
vi
Contents
Chapter 9 Body dynamics, social meaning-making, and scale heterogeneity: Re-considering contextualization cues and language as mixed-mode semiosis Paul J. Thibault
127
Chapter 10 Continuing the discussion with John J. Gumperz Carlo L. Prevignano and Paul J. Thibault
149
Bio-bibliographical note
163
Index
165
Contents vii
Preface The aim of the present collection of papers is to highlight some aspects of John Gumperz’s analysis of human interaction, that which he calls Interactional Sociolinguistic Analysis. This choice has meant that a number of themes that are central to his work are less focussed on. Our principal concern has been to propose ways in which Gumperz’s research may speak to the theory and practice of both communication analysis and interaction analysis, present and future. Moreover, we seek to promote discussion and debate between the work of Gumperz and other analytical and theoretical approaches to human interaction. At the same time, we would like to situate this volume within a by-nowconsolidated tradition of interviews and peer commentary on the research Wndings of scholars in the human sciences. Herman Parret’s Discussing Language (1974) and the journal Behavioral and Brain Sciences come to mind as signal contributions conceived in this spirit. The present volume is proposed to the reader with these considerations in mind. In his interactional sociolinguistic perspective, Gumperz is interested in the communication ecologies where participants’ interpretive acts and conversational inferencing reXect societal power relationships and ideological processes in the course of human interaction. The role of contextualization cues, which Gumperz considers to be pure indexicals (Prevignano and di Luzio, this volume: 9), is crucial here. On the one hand, contextualization cues are the means whereby social actors retrieve presuppositions in order to make sense of what they see and hear in interactive encounters; on the other hand, they interact with symbolic, fully-coded, lexical and grammatical signs in the processes of constituting speech events. We shall now brieXy introduce the speciWc contributions to the volume. Di Luzio’s introductory essay is an overview of Gumperz’s research program since his earliest dialectological study of the German Swabian dialect of a rural community in Michigan in 1954. Di Luzio examines the main contours of Gumperz’s development as both interactional sociolinguist and anthropologist, focussing particularly on the development of Gumperz’s research on the social motivations for linguistic variation, the interpretative basis of interactional sociolinguistics, and the notion of contextualization cues. The discussion between John Gumperz, Carlo Prevignano and Aldo di
viii Preface
Luzio is the center piece of this volume. It provides a forum in which Gumperz elaborates on and explains many of the central tenets of his approach. In particular, the discussion examines the ethnographic basis of Gumperz’s research in sociolinguistics, the need to distinguish linguistic forms from the communicative practices in which these are embedded and in which they have their meaning. This is a constant theme throughout this discussion and it is refracted through an interesting series of reXections on intercultural communication, social action, communicative conventions, and contextualization. Levinson’s essay draws on the author’s early association with Gumperz as a Ph.D. student in Berkeley. Levinson develops an internalist approach to contextualization cues, arguing that utterances specify their own contexts of interpretation in often very implicit ways, through the use of a range of verbal and non-verbal resources. Context in this reading is not something which is externally imposed on utterances. Levinson’s essay shows how the explicit propositional meaning of utterances is often signiWcantly modiWed by their cross-modal connections with non-verbal and more implicit, topological modalities of semiosis. This further suggests that many of the implicit aspects of human communication cut across apparent surface distinctions between semiotic modalities, and that this may lead one to make overhasty and superWcial distinctions between various meaning-making resources. Thibault’s essay is complementary to Levinson’s in this sense. Thibault argues for an approach which suggests a way of cutting across these apparent ‘surface’ distinctions in an attempt to develop a conceptually uniWed approach to the question of communicative — meaning-making — practices, irrespective of the semiotic resources that are deployed. To achieve this, he argues for a three-way distinction between indexical, intertextual, and metatextual meaning-making practices. This distinction is proposed as a means for analyzing the ways in which agents access and co-ordinate their deployments of semiotic resources in the activity structure-types and discourse genres of a given culture. Prevignano’s contribution sheds light on a diVerent aspect of this problem with reference to what he deWnes as the “minims of interaction”. In this, he contrasts Gumperz with Grice, Leech, and other pragmatists, whom Prevignano characterizes as “maximists”. In particular, Prevignano examines some “interaction minims”, which participants deploy in order to signal to each other what they are doing in the course of the interaction. Importantly, such interaction minims are embedded in activity structures and entail inter-
Preface
pretive principles that are historical rather than universal in character. The analyst’s interpretation, insofar as it is a hermeneutical reconstruction of the participants’ interpretation, is historical in two fundamental senses. First, participants are embedded in a history of their applications and interpretations of each others’ interactional minims. Secondly, the analyst, too, is embedded in a history of ethnographic and analytical reconstruction and understanding. This history itself acts on and aVects our understanding as both participants in and analysts of the object of study — human interaction. Ballim approaches the same problem of interpreting utterances in context, though from a diVerent perspective — that of computer-mediated communication from the point of view of applied computational linguistics. A central problem for Ballim is how utterances are interpreted in the light of the knowledge, beliefs, desires, and intentions of discourse participants. Consequently, he is interested in how discourse participants create models of each other in the course of interaction. Ballim indicates that the overall system forms top-down constraints on the ways in which participants make inferences and reason about each other in the course of interaction. His discussion is informed by his research on the use of language as a communicative medium between humans and computers. Eerdmans provides a critical account of two exemplar case studies conducted by Gumperz in the area of intercultural communication. She shows how participants involved in inter-ethnic communication interpret and negotiate their interlocutors’ contributions by embedding what they see and hear in interpretative frames which may conXict with or misconstrue those of the interlocutor. Interactional sociolinguistics, with its hermeneutic basis, can provide a powerful tool for explaining such failures on the basis of its expanded understanding of the interaction between the internal dynamics of agents — their structures of understanding and inferencing — and the largerscale dynamics of the unfolding activity-structure, along with its relations to the wider social and cultural contexts at play in the interaction. In this way, we begin to see some meaning-based answers as to why interactants act and behave in the ways they do, and how misunderstanding and communication failure occur. Eerdmans concludes her discussion with some comments on the usefulness of interactional sociolinguistics for both teachers and learners of a second language. In particular, she draws attention to the practical and theoretical insights which a critical awareness of the impact of diVerent cultural backgrounds can have on teaching and learning contexts.
ix
x
Preface
In his Response essay, Gumperz ranges over his career trajectory as sociolinguist before clarifying and expanding on a number of points raised in the individual contributions as he responds to these. In particular, he makes further pertinent observations on how his current thinking about language and interaction builds on his earlier ethnographic Weld studies. In so doing, he also clariWes the ways in which linguistic and cultural diversity and sociocultural boundaries are represented in and shape linguistic interaction. In a second essay, Thibault gives voice to some further developments of the theoretical issues raised by the individual contributors. In particular, Thibault returns to the notion of action and interaction as a unifying principle for the analysis of the multimodal co-deployment of semiotic resources. He argues for the embeddedness of action in higher-scalar ecosocial environments which interact with the lower-scalar embodied dynamics of the agents who participate in discursive interaction. He also explores the relationship between the notions of fully-coded message content and contextualization cues, as discussed by both Gumperz and Levinson. In particular, he suggests how these notions can be brought together in a more theoretically-uniWed framework by considering the typological and topological dimensions of language as a form of mixed-mode semiosis. A Wnal discussion between Gumperz, Prevignano and Thibault rounds oV the discussions which have taken place throughout this volume. In this way, Professor Gumperz provides further clariWcation of a number of key issues in interactional sociolinguistics, the role of inferential processes in interpreters’ understandings of each other’s meanings in interactional events, and the importance of these insights for understanding the contribution of crosscultural factors in communicatively-diverse environments. Gumperz also provides further comments on the centrality of ideological processes in human interaction, as well as sharing with his readers his views on a number of other approaches with similar aims to interactional sociolinguistics. A short bio-bibliographical note detailing the most salient moments in Professor Gumperz’s career concludes the volume. The present is a revised edition of a volume which was issued as a prepublication by Beta Press (Lausanne CH, 1997) with the title Discussing Communication Analysis 1: John J. Gumperz. This Preface, as well as Chapters 8, 9 and 10, have been specially written for this new edition.
Preface
Bertie Kaal of John Benjamins provided us with constant editorial advice and support from the outset. To Bertie Kaal, all our thanks and deepest appreciation. To conclude, we would like to express our gratitude to John Gumperz for his interest, patience, and generosity throughout all the stages of preparation of this volume, which we wish to oVer him on the occasion of his 80th birthday, 2002. Susan L. Eerdmans Carlo L. Prevignano Paul J. Thibault
xi
xii Contents
Presenting John J. Gumperz
Chapter 1
Presenting John J. Gumperz Aldo di Luzio
As a linguist and anthropologist, John Gumperz has successfully combined and integrated diverse yet complementary issues: theoretical construction with empirical research, linguistic with sociocultural facts, cognitive constraints with communicative ends, text with context, discourse with ethnography, analyst interpretation with member interpretation. In this way, he has developed a new method of considering and explaining intra- and intercultural linguistic processes in multilingual and multicultural communities in three continents. Gumperz dealt with languages and cultures in contact in 1954 with his Ph.D. dissertation at the University of Michigan. The topic was the German Swabian dialect of a rural community of the third generation of immigrants in Washtenaw County, Michigan. In this study, Gumperz showed something very new and innovative for the dialectology of the time, which was still strictly diachronically and geographically oriented. He was able to demonstrate that the actual distribution of dialectal variables was a consequence of social and religious groupings after the settlement in the USA in the 19th century. The connection or functional relation between the type of group construction and that of language shift and variation represents a recurrent topic in Gumperz’s subsequent research in very diVerent kinds of speech communities. Thus, in the second half of the 1950s, in his two years’ Weldwork together with an interdisciplinary research team in a North Indian village with strict caste diVerentiations, Gumperz again dealt with the very important question of the relation between the diversity of sociocultural groups and language variation. Contrary to what one would have expected, he proved that the distribution of linguistic varieties or variables in that North Indian commu-
1
2
Language and interaction
nity did not reXect either caste diVerentiation or other social factors, such as, for example, income, educational level, or social status. On the contrary, and as a decisive factor for the distribution of the linguistic variables, norms were operating which governed and constrained the social quality of interpersonal communication. In his analysis of code-switching between Hindi and Punjabi in Delhi, he made it clear that change of code represents a style which depends on social norms and may express solidarity and group membership. Accentuating the fact that code-switching is not only a grammatical but also and especially a discursive process, he showed that this could not be analyzed with traditional linguistic methods. His analysis of the mutual adaptation of Kannada (a Dravidian language), Marathi and Urdu (two Indo-Aryan languages) in interactions among their speakers in a multilingual community in the region of Mahrastra represents a milestone for the analysis of creolization and language convergence studies. In his analysis, he made clear that under the inXuence of very frequent code-switching and consequent simpliWcation and reciprocal assimilation, especially at the syntactic level, the local dialects of the three typologically very diVerent languages become creole-like varieties of the same language. They present a similar linear structure and can be reciprocally translated word for word. What in earlier structuralist and psycholinguistic approaches was analyzed as interference is now interpreted by Gumperz as socially and communicatively aVected convergence. Some very important results of Gumperz’s earlier research are that (i) diVusion and limitation of linguistic variables are determined and displayed in interaction; (ii) speakers’ perceptions or deWnitions of language equivalence or diversity do not depend on genetic aYliation; (iii) speech communities are not linguistically homogeneous; (iv) linguistic variation and alternation are not random or arbitrary, but communicatively functional and meaningful. These earlier achievements already evidenced new directions in sociolinguistics. Gumperz showed how multilingualism or multidialectalism, as well as style variation, with their dialectical and dialogical aspects, represent a prototypical paradigm or condition of human communication. Gumperz’s conception, which strongly contrasts with the traditional and structuralist one of idealized and homogeneous languages and speech communities, is strictly bound to his new concepts of speech community, language repertoire and sociolinguistic co-occurrence rules. Members of the same speech community do not necessarily speak the same language variety. The only requirement is that the speakers have in common a linguistic reper-
Presenting John J. Gumperz
toire, that is, the whole of the linguistic resources and communicative strategies at their disposal and the rules for using them in communication. These are conceived as co-occurrence rules or implications, which determine at the paradigmatic and the syntagmatic levels the choice of units from the repertoire in building utterances; they also deWne the stylistic proWle of the discursive units according to the communicative goals of the interaction. In this way, the apories derived from the dichotomy between langue and parole, competence and performance, are resolved. For Gumperz as well as for Hymes, with whom he collaborated in the 1960s in order to deWne some aspects of his approach to the ethnography of communication, linguistic activity is always situated in a dialogical or interactional context and is revealed within the ethnographic context of the speech community of which the interactants are members. However, Gumperz tends to consider the communicative units in which communication is enacted as types of speech activities or communicative genres rather than speech events. The interactional sociolinguistics which Gumperz developed in the 1970s implies a new hermeneutical orientation. It is critically opposed to correlational methods because these are external to discourse and cannot explain the relation between linguistic facts and social structure. Interactional sociolinguistics may be seen as a sociolinguistic analysis of dialogue and of the social meanings displayed in the dialogical interaction against the background of the ethnography of the sociocultural context of the interactants. Regularities of linguistic interactions are no longer considered as reXecting independent social norms; on the contrary, these norms are considered as interactively (re)produced in the communicative situation. The task of the interactional sociolinguistic analysis of determinate communicative situations is to reconstruct the interactive methods by which speakers use discursive means from their sociolinguistic knowledge (for example, language choice, code-switching and stylistic variation according to determined cooccurrence rules) for the production and interpretation of the social meaning intended. An illuminating example of an interactional sociolinguistic reconstruction of how, in a determinate sociocultural situation, discourse is structured and social meaning is produced, is already oVered by Gumperz’s study of code-switching in the small speech community of Hemnesberget in northern Norway. A very important element of interactional sociolinguistics is a generalized theory of contextualization and interpretation in communicative interactions.
3
4
Language and interaction
According to this theory, the interpretation of meaning is interactively negotiated and constituted, taking into account the knowledge of context which the interactants have at their disposal. In this way, the Lebenswelt enters into the discourse. In order to constitute meanings, speakers activate interpretative frames or schemata from their experience and from their grammatical, lexical and pragmatical knowledge. The enacting of these schemata is called contextualization. Contextualization is thus the process by means of which speakers relate what is said in an interaction to the context of the background knowledge which is presupposed. It is a necessary element of communication. Conversational interpretation or inference goes beyond the decodiWcation of lexical referential meaning. Presuppositions of context are activated and indicated by means of contextualization cues, such as, for example, intonation, rhythm, tempo, loudness, gestures, language or register choice and code-switching, etc.. They are all means whereby intended meaning is signaled. Contextualization cues are produced and organized according to cultural conventions. In modern multiethnic society, in spite of eVective referential transfer of information, a communicatively adequate contextualization of discourse can fail because of diVerent cultural conventions. The negotiation of meaning necessarily presupposes assessment and understanding of the culture and communicative habits of the interlocutors. The assessment of the meaning in the ongoing interaction is also culturally conditioned. The inXuence of language on culture and of culture on language are for Gumperz bi-directional. Culture is strictly interwoven with language and is displayed in the typiWed and conventionalized ways of signaling speech activities through contextualization cues in interaction. Thus, for example, Gumperz has shown in one of his studies how, in the intercourse between Hindi English speakers and native English speakers, diVerent prosodic and rhetorical conventions lead to problems in the interpretation of the meaning. According to Gumperz, the communicative diYculties do not depend on lack of grammatical or lexical competence, but on diVerences in the rhetorical styles which are learned and used in the diVerent networks of interpersonal relations. Problems of intercultural understanding are, according to Gumperz, much more serious, and put the non-dominant interactants of minority groups at a disadvantage in institutional settings and interactions, e.g., in counseling and job interviews, public debates and discussions, formal hear-
Presenting John J. Gumperz
ings, law proceedings, formal interactions in the work place or at school. In many studies in which Gumperz analyzed such situations, he has shown how the diVerent ways in which interactants realize speech activities, leading to communicative failure, mostly depend on (i) diVerent cultural assumptions about the situation and the choice of appropriate speech behaviour; (ii) the diVerent ways of structuring and developing information and argumentation; (iii) the diVerent ways in which contextualization cues are used and made functional. From the study of how communication broke down in particular cases, he was able to gain new insights into the communicative rules and forms of sociocultural knowledge which in general govern communication and constrain the forms of its realization. Intercultural communication, according to Gumperz, is not conWned to interethnic encounters, but is prototypical for modern industrialized and bureaucratized western societies when diVerent subgroups live and interact with one another in large urban agglomerations. In many institutional encounters, one very often has to speak with people outside one’s own group in order to defend one’s own needs and rights and to win personal and social control. In such situations, in general, two diVerent forms of speaking and communicative strategies contrast with one another. These denote and connote not only or not necessarily social identity and group membership, but also symbolize diVerent attitudes and support commitment to diVerent systems of values. The use of in-group symbols in such situations contrasts with the need to control access to important resources. From this situation, diVerences in the assessment and interpretation of the rhetorical strategies employed by the other interactants may follow. These can negatively aVect the result of the interaction with representatives of institutions or majority groups for the interactant who has failed to master the appropriate rhetorical forms of communication. Gumperz shows how power asymmetry is hidden under the surface of equality, since on the assessment side, the performers are people who know what to expect in the interaction as well as the rules of their rhetorical attainment, but who do not make them explicit in such interactions. Starting from “a microsociological point of view”, Bourdieu has advised us not to forget that “social structure” is present in every interaction and that communicative intercourse symbolizes and enacts power structure. Gumperz has made manifest and accountable the realization of these power structures. He has empirically reconstructed how language as social behavior produces and reproduces power structures in the course of interaction. He draws our
5
6
Language and interaction
attention to the fact that discrimination in interaction has a linguistic dimension. In this case, it is not important how communication is grammatically structured, but in particular what it does socially and how it is socially governed and constrained. Gumperz’s interactional sociolinguistics has crucially contributed to the recognition of the ways in which language (or speech) and society (or culture), the cognitive and communicative aspects of language, the theory and praxis of both speakers and analysts, are strictly interwoven. He has also shown us how to account for the links between these. The methods of analysis of communicative or dialogical interactions which are based on Gumperz’s interactional sociolinguistics are hermeneutical and necessarily imply ethnographic analysis and the inclusion of the situational and ethnographic context which is reXexively reproduced in the interaction. The criteria used for the interpretation of the social meaning intended and interpreted in communicative interactions are not a prerogative of the analyst, but are based on those of the participants in the interaction. The analyst’s interpretation is a reconstruction of the participants’ interpretation. The assessment of the analyst who has to reconstruct, that is, to know and understand, the ethnographic context as if he were a participant member, is a reconstruction of the participants’ assessments and may possibly be conWrmed by them. Such assessment is based, as Vico would say, on sensus communis, or on judgement without reXection. Gumperz’s method and his factual criteria for the analysis of communicative interaction remind us of Vico’s hermeneutical method and the criteria he stated in his Scienza Nuova:verum est factum and verum et factum convertuntur.
A discussion with John J. Gumperz
Chapter 2
A discussion with John J. Gumperz1 Carlo L. Prevignano and Aldo di Luzio
Carlo L. Prevignano: I’d like to begin by asking you about your current research projects. John J. Gumperz: My interests are both theoretical and applied. Currently I’m engaged in empirical research on classroom interaction with eight- and nineyear-old Spanish/English bilinguals. I use methods of conversational and discourse analysis to formulate hypotheses about the indirect reasoning processes students employ in problem-solving. These processes are for the most part indirect in that they rely on implicit, taken-for-granted presuppositions to convey information that adults would expect to be overtly lexicalized. By formulating explicit, veriWable, analytical assumptions about what these processes are, and about how children of that age rely on them and talk about them with their peers, we hope to help teachers gain better insights into their students’ ways of dealing with the learning task. Ultimately, we intend to make training Wlms illustrating such reasoning processes, using audiovisual techniques similar to those employed in the B.B.C. Crosstalk Wlms (Gumperz et al. 1979) to reveal the possible linguistic causes of misunderstandings in intercultural communication. As with my earlier research, such empirical studies are directly related to my theoretical interest in showing how detailed analyses of communicative practices can illuminate basic issues in social theory. I’m about to begin working on a book which I hope will bring out the import of sociolinguistically-oriented pragmatic analysis for our understanding of social process. Theorists of many persuasions argue that the social environments in which we live and act are dialogically constituted. The main question I pose is: how does verbal communication aVect such dialogic processes and to what extent do these processes depend on shared linguistic knowledge? My perspective on verbal communication is grounded in earlier studies
7
8
Language and interaction
in the ethnography of communication. The key insight here is that ethnographically-based sociolinguistic analysis, if it is to be empirically viable, must focus on speciWc speech events, deWned as interactively constituted, culturally-framed encounters, and not attempt to explain talk as directly reXecting the norms, beliefs and values of communities seen as disembodied, hypothetically uniform wholes. To look at talk as it occurs in speech events is to look at communicative practices. Along with others, I claim that such practices constitute an intermediate and in many ways analytically distinct level of organization. A sociological equivalent here is Erving GoVman’s “interaction order”, a level of organization which bridges the linguistic and the social. GoVman’s work on this topic has greatly inXuenced the conversational analysts’ argument that conversation is separate both from grammar and from macro-social structures and must be analyzed at the level of “activity” (their use of this term is diVerent from my own; see below). In my approach to interaction, I take a position somewhat between that of Erving GoVman and Harold GarWnkel. The former looked at encounters from an ethologist’s perspective, while the latter was concerned with the interpretive processes that make interaction work. The current trend in my work began with the papers in the 1982 book Discourse Strategies (Gumperz 1982). Apart from its empirical focus on interethnic and intercultural communication, this book can also be seen as a Wrst attempt to explore the role of typiWed communicative practices in interaction, what levels of linguistic signaling they reXect, how they relate to speakers’ communicative and social background and how they aVect interactive outcomes in key encounters. It is the focus on the interactive and therefore social import of the Wne details of verbal communication that distinguishes my work from others. One of my main concerns is with how we can analyze communicative practices in such a way as to account for participants’ ability to create and maintain communicative involvement and to achieve their communicative ends. By background I am a linguist trained in the tradition of Saussure, Sapir and BloomWeld. Structuralism has been severely and, on the whole, convincingly criticized by Bourdieu and a host of others. Yet, I believe that the structuralists’ basic insights into linguistic, that is, phonological and syntactic, competence and their approach to speaking as a partially subconscious process, continue to be useful. The problems arise in analyses of everyday talk. In re-analyzing my ethnographic Weld data on communicative practices for Dis-
A discussion with John J. Gumperz
course Strategies, I began to realize that Saussurean phonological and grammatical structures deWned in terms of Wnite sets of oppositions and truth condition semantics could not account alone for the relevant dialogic and discursive facts of everyday talk. These insights are of course not unique. Along with others in my Weld, I became aware of the semantic importance of context. Gregory Bateson had long talked about communication being both context-creating and context-dependent. Conversational analysts provide impressive empirical evidence to show how interpretations shift as part of the on-going sequential ordering of an interaction. But I argue that sequential ordering cannot be taken as a structural given. It presupposes active conversational involvement on the part of speakers, listeners and audience members. The ability to create and maintain such involvement rests on shared conversational inferences. I have proposed the notions of contextualization cues and contextualization processes as a way of accounting for the functioning of linguistic signs in these inferential processes. Contextualization cues are a class of what pragmaticians have called “indexical signs”, which serve to retrieve the contextual presuppositions conversationalists rely on in making sense of what they see and hear in interactive encounters. They are pure indexicals in that they have no propositional content. That is, in contrast to other indexicals like pronouns or discourse markers, they signal only relationally and cannot be assigned context-free lexical meanings. Yet they play a major role in transforming what linguists refer to as “discursive structures” into goal-oriented forms of action. A main aim of my current work on discourse and conversation is to show how indexical signs, including prosody, code- and style-switching, and formulaic expressions, interact with symbolic (i.e. grammatical and lexical) signs, sequential ordering of exchanges, cultural and other relevant background knowledge to constitute social action. This, in brief, is my long-term research program. In my current writing, I am attempting to integrate the various strains of thought I have alluded to above, together with my empirical Wndings on urban communicative practices, into the outline of a coherent theoretical framework. Sociolinguistic explanation, if it is to be relevant to today’s concerns, cannot implicitly accept traditional categories of language, culture and society. I believe that interaction at the level of discursive or “communicative practices”, as Hanks (1996) calls them in his important book by that title, must be seen as separate from either linguistic or socio-cultural processes. It is constituted by the interplay of
9
10
Language and interaction
linguistic, social and cultural/ideological forces and governed or constrained by partly universal and partly locally-speciWc organizational principles. My argument is that systematic investigation of these principles can provide a vantage point for an empirically-based reworking of the established traditions that continue to follow structuralist practices of separating the linguistic from the social. The recent linguistic anthropologists’ move from a Saussurian to a more inclusive, broader, Peircian semiotics that distinguishes between symbolic and indexical signs is a Wrst step in that direction. Symbolic signs communicate via well-known grammatical and lexical rules. Indexical signs (and among them contextualization cues), on the other hand, communicate by virtue of direct conventional associations between signs and context, established or transmitted through previous communicative experience. Conversational inferences build on both signaling processes. A major issue in my own research is to show, without abandoning what we have learned from structuralism, how and under what conditions discursive practices work (a) to create communicative conventions and (b) to aVect interpretation. C.L.P.: Can you tell us some more about your idea of contextualization cues? How do you see these verbal and non-verbal constituents in relation to, for example, their function? Do you assume there is a one-to-one mapping between cues and cued elements? Because, in my opinion, there is a problem here. If you can Wnd a sort of one-to-one mapping or relationship, everything is Wne, but, in pragmatics, there are usually many-to-many relations. Another problem: you introduced the idea of discourse strategies more than Wfteen years ago. GarWnkel and others were talking about inference and... J.J.G.: Inference and sequentiality. C.L.P.: Inference and procedures. Do you see a diVerence between ethnomethods or ethnoprocedures and your discourse strategies? J.J.G.: If we accept that (a) interpretation is always context-dependent and (b) contextual presuppositions shaping interpretations are themselves subject to constant change in the course of an interaction, then we cannot expect to Wnd one-to-one mappings of form to meanings. There are always several possible interpretations. Contextualization cues, along with other indexical signs, serve to retrieve the “frames” (in GoVman’s sense of the term) that channel the interpretive process by “trimming the decision-making tree” and limiting the range of possible understandings. In talking about the functioning of indexi-
A discussion with John J. Gumperz
cal signs in interpretation, it becomes necessary to distinguish between meaning in the linguist’s sense of reference, and situated inferences. The latter are crucial in communicative practice. In everyday talk, situated inferences always take the form of assessment of what a speaker intends to convey by means of a message and these are often quite diVerent from propositional content. For example, if you and Paolo had been talking, and I asked you, “What did you just do?”, you will not answer, “I made a statement” or “I performed a speech act.” A more likely answer would be, “I asked him for a favor” or “I asked him if he was free this evening.” Moreover, even if you had said, “I asked him a question,” I would not take this as referring to grammatical or speech act categories, but rather as telling me what you wanted from him. Communicative practices are actions, and conversational inferences are made by human agents, acting in the real world. With respect to GarWnkel: in arguing that all communication is intentional and based on inferences, I build on GarWnkel’s notion of “inferential process”. But GarWnkel is not speciWc as to what he means by the “historical method” by which members’ interpretive processes can be retrieved, apart from saying that we need to resort to background information on how and why a particular inference came about. How do we know what aspects of background knowledge are relevant at any one time, and is extra-communicative background knowledge enough? We must assume that information about contextual frames is communicated as part of the process of interacting, and therefore it becomes necessary to be clearer about the speciWcs of what happens in the interaction as such to assess what is intended. Conversational analysts set out to do this, and their work has brilliantly shown what can be learned through turn-by-turn sequential analyses. But, as I suggested, sequential analysis alone cannot account for situated interpretation. It describes just one of the many indexical processes that aVect inferencing. I’d like to argue that assessments of communicative intent at any one point in an exchange take the form of hypotheses that are either conWrmed or rejected in the course of an exchange. That is, I adopt the conversational analysts’ focus on members’ procedures, but apply it to inferencing. The analytical problem then becomes not just to determine what is meant, but to discover how interpretive assessments relate to the signaling processes through which they are negotiated. But how can we overcome the inherent ambiguity of inferential processes? In my empirical studies, I have worked out a set of procedures along the
11
12
Language and interaction
following lines. Analysis begins with turn-by-turn scanning at two levels of analysis, content and rhythmic organization. The aim is to isolate sequentially-bounded units, marked oV from others in the recorded data by some degree of thematic coherence and by beginnings and ends detectable through co-occurring shifts in content, prosody, tempo or other formal markers. Lectures, ceremonies of various kinds, interviews, that is, named units of the type normally studied by ethnographers of communication, are instances of such events. But event sequences can also be isolated in everyday conversations and other casual encounters, where, for instance, narrative sequences may alternate or be interspersed with discussion, argument, banter and the like. In performing this segmentation, we seek to discover natural units of interaction that contain empirical evidence that conWrms our analyst’s interpretations, evidence against which to test assumptions about what is intended elsewhere in the sequence. Such event sequences then form the basic units for the analysis of conversational inferencing. They vary in length from longer sequences, which in turn can contain many sub-events, to brief three-part exchanges of move, countermove and conWrmation or disconWrmation. In phase two of the analysis, events are transcribed. The goal here is to prepare “interactional texts” by setting down on paper all those perceptual cues — verbal and nonverbal, segmental and nonsegmental, prosodic, paralinguistic and other cues — which past and ongoing research shows speakers and listeners demonstrably rely on as part of the inferential process. This enables us not only to gain insights into situated understandings, but also to isolate recurrent form-context relationships and show how they contribute to interpretation. These relationships can then be studied comparatively across events, to yield more general hypotheses about members’ contextualization practices. To return to conversational inference and its role in communicative practice. Let me give you some concrete examples to show how I view the process of understanding. Some time ago, while I was driving to the oYce, my radio was tuned to a classical music station. At the end of the program, the announcer, a replacement for the regular host who was returning the next day, signed oV with the following words: “I’ve enjoyed being with you these last two weeks.” I had not been listening very carefully, but the extra-strong accent on “you” in a syntactic position where I would have expected an unaccented pronoun caught my attention. At Wrst the speaker’s words seemed to suggest that he intended to produce the Wrst part of a formulaic exchange of compli-
A discussion with John J. Gumperz
ments. But since there was no one else with him on the program, I inferred that by the way he contextualized his talk, he was indirectly — without putting it ‘on record’ — implicating the second part, “I hope you have enjoyed listening to me” (see Gumperz 1996). A second, somewhat more complex example, comes from my analysis of the cross-examination transcript of the victim in a rape trial. Counsel: “You knew at the time that the defendant was interested in you, didn’t you?”. Victim: “He asked me how I’d been... just stuV like that” (see Gumperz 1995a). In both cases, I had to search my memory of past communicative experience to construct a likely scenario or narrative plot that might suggest possible interpretations. My initial hypothesis in example one conXicted with what I knew about the radio program, and this triggered a search for a diVerent, more plausible, scenario. In the second case, I relied on what I knew about cross-examinations as adversarial proceedings, where the attorney attempts to expose weaknesses in the defendant’s testimony. But while these general facts tell us something about participants’ motives in their choice of verbal strategy, we need to turn to what they actually said to understand what they intended to convey. By the words he chose, and by the way he contextualized his talk, the attorney raised the possibility that defendant and victim had had a prior relationship. The victim’s move, on the other hand, positioned as it is immediately after the attorney’s question, implicitly argues for a diVerent scenario, one where the two were merely casual acquaintances. In this way, she sought to deny and in a sense ward oV the questioner’s potential attack on her testimony. I use the term “activity type” or “activity” to refer to the above type of constructs or “envisionments”, to borrow Fillmore’s term. My claim is that all interpretation rests on such constructs. Activities are an aspect of GoVmanian frames and are subject to constant change in the course of the exchange. That is, they do not apply to events as wholes, they apply to each component move. I argue that ultimately all interpretation at the level of discursive practice relies on these constructs. This view of understanding has some similarity to Fillmore’s notion of “scene” that he discusses in his work on the semantics of understanding. But whereas Fillmore is concerned with physical settings, I take more of a social perspective. I see activities as evoking the actions of actors engaged in strategically formulating and positioning their moves in order to accomplish communicative ends in real-life encounters. In so doing, they rely on their
13
14
Language and interaction
presuppositions about mutual rights and obligations, as well as on ideologies of language and individual personalities, to get their message across. This implies that, in addition to meaning assessment in the established sense, there are always social relationships that are continuously negotiated and renegotiated by means of the same interpretive processes by which content is assessed. It is useful to distinguish between two levels of inference in analyses of interpretive processes: (a) global inferences of what an exchange is about and what mutual rights and obligations apply, what topics can be brought up, what is wanted by way of a reply, as well as what can be put into words and what is to be implied, and (b) local inferences concerning what is intended with any one move and what is required by way of a response. In this way it becomes possible to account for changes in frame as a function of the sequential positioning of moves. Both levels of interpretation involve activities as cognitive constructs: the Wrst is related to what GoVman calls “framing”, while the second deals with something like the conversational analyst’s “preference organization”. While contextualization cues assist in retrieving the knowledge on which activity constructs are based, they do not work in isolation. Interpretation always relies on symbolic, lexical and indexical signs. But as pure indexicals, contextualization cues are usually produced and interpreted without conscious reXection, and are therefore particularly useful in revealing frequently unnoticed aspects of the interpretive process that tend to be highly sensitive to cultural variability. I’m not claiming of course that these methods solve the problem of interpretive ambiguity. The aim is to Wnd likely solutions that are plausible in that they show how component actions cohere in the light of the event as a whole, be it a three-part string of moves or a longer encounter. This is of course quite diVerent from assessing the truth or falsity of speciWc interpretations. The method resembles conversational analytic procedures of reconstructing the general procedures members employ in formulating speciWc actions. I diVer from conversational analysts in that my concern is with situated, on-line interpretation. I want to show both what the most likely inferences are and how participants arrive at them. In studies of intercultural and interethnic communication, these methods have been useful in detecting systematic diVerences in interpretive practices aVecting individuals’ ability to create and maintain conversational involvement. Aldo di Luzio: How do you consider coherence? Is coherence dependent on the type of activity you are engaged in?
A discussion with John J. Gumperz
J.J.G.: Not exactly. Our assumptions about relevant activities determine our view of how the interaction coheres. A.d.L.: By “type of activity”, do you mean “genre”? J.J.G.: Not quite. I use the term “genre” for another level of pragmatic analysis, the ideological level. For me genre is not an analytical category, because, as I understand it, genre, at least in the Bakhtinian sense of the term, is better treated as an analyst’s, or for that matter also a lay-person’s, concept for referring to or labeling texts or speech exchanges. A.d.L.: A mode of speaking about something. J.J.G.: It’s not just a “mode” of speaking. When we speak of moral discourse as a speech genre, for example, we are making an ideologically-charged metapragmatic assessment. That is, we are engaging in a form of talk about talk. As I said above, I use the notion of activity to refer to conversationalists’ and analysts’ hypotheses about what cognitive processes are involved in understanding, hypotheses that can then be validated by methods such as those I have outlined. C.L.P.: Could you add some remarks about your idea of linguistic convention and typiWcation? J.J.G.: Yes, I don’t want to use the term “convention” in the linguist’s sense of grammatical convention. I use convention as a general term in the lay sense, as the outcome of a process of typiWcation. A.d.L.: In the Schutzian sense? J.J.G.: In the Schutzian sense. For example, Liliana Ionescu-Ruxandoiu this morning was talking about “pre-sequences”.2 She said certain expressions can serve as pre-sequences. This reminded me of a Weld-work experience in a South Asian village where I lived for almost two years. People would come up to me in the street and say, “Dinner is ready, come on over.” Once, when I seemed hesitant, the speaker said, “Don’t you want to come? We’ve made dinner, why don’t you join us?”. I still hesitated. I didn’t know what to do. Finally, thinking I had perhaps actually been invited and had forgotten about it and might oVend them if I didn’t come, I went along. Well, it turned out that I was... C.L.P.: Invited?
15
16
Language and interaction
J.J.G.: I was not invited, no. It was just a greeting. An invitation is a very solemn matter in the village. To invite someone, you send an emissary, a young man in your household, to give the invitation. And even then, one doesn’t go to the host’s house until a second emissary comes to call and asks the guest to come over. This means that a special meal has been prepared and is now ready. So that to say, “Dinner is ready”, to someone in the street, simply counts as a form of greeting. C.L.P.: This is a convention, the outcome of a typiWcation within a community. J.J.G.: One Wnds similar usages all over South Asia. In some parts of the continent they say, “Have you eaten yet?”, as a way of conveying something like, “How are you?”. Such conventions arose over time as outcomes of culturally-speciWc processes of typiWcation. They may reXect a time when people in farming communities did not have everything they wanted to eat. C.L.P.: So when you are confronted with this kind of utterance, if you come from another community, like myself, you take it as an invitation. J.J.G.: Exactly, as I did. In fact, once I went, and they had to prepare the food. And I don’t know whether they had enough to feed me. They were not poor people — it was a wealthy village — but still, you know, the women had to get to work and make some more food. C.L.P.: To come back to your notion of contextualization cues. You said that you are not interested in the propositional content, but in the indexical function in the Peircian sense, versus the symbolic, is that right? J.J.G.: Yes, yes. C.L.P.: So you see cues as orienting people to... J.J.G.: Certain interpretations. C.L.P.: And what about the orienting relationship? You have an addressee who has to be oriented to, who has to be instructed via cues... J.J.G.: That’s right. C.L.P.: However, there are diVerent competences, repertoires, but let us assume that you have two people with just the same repertoire or competence. Faced with the same cue, they are supposed to be oriented in the same way, in
A discussion with John J. Gumperz
the sense that they are supposed to reach the same cued element ... J.J.G.: The same inferences. Be able to draw similar inferences, yes. C.L.P.: But this is not always the case. Given these two people belonging to the same community, with the same repertoire of cues, are we sure that there are no cases of ambiguity, contrary to the one-to-one relationship? J.J.G.: There always are some ambiguities. C.L.P.: There always are, okay. J.J.G.: There usually are. Ambiguities always exist. This is why I emphasize that the basic issue is not whether or not people understand factual information, but whether or not participants in an encounter are able to attune to each other’s interpretive processes. For example, ambiguities and misunderstandings always occur, but you need to be able to repair them. And since conversational repairs must, for reasons inherent to conversing, rely in large part on indirectness, repairs always make unusually high demands in the way of shared inferences. In the cross-examination example I gave before, it is not that the victim and the attorney were unable to draw similar inferences. On the contrary, given the way jury trials work, they chose intentionally to rely on shared interpretive conventions to convey conXicting accounts of what might have happened. C.L.P.: During the IADA Round Table, Edda Weigand spoke about coming to an understanding in dialogue, a sort of Habermasian idea. However, you are never sure you have reached a total agreement, a convergence of inferences, of schemata and so on. So misunderstanding is a human condition, and you are alone, in a certain sense, with your cues, facing the other person. And this is the reason why I see individuals as a bit like solipsistic entities. They just send and get cues, and cues about cues, trying, via them, to instruct the other’s inferences or schema use. J.J.G.: But we are not really talking about individuals agreeing on what something means. Nor is it so easy to say when people have the same communicative background. Let me give another example. One of the very best of my former students, a Nigerian, Niyi Akinnaso, was born in a Nigerian village. He was the Wrst literate person in the community. His father, a local chief, was secretary of the farmers’ cooperative and he used to keep accounts using a local mechanical device, something like an abacus. As a child, Niyi used to
17
18
Language and interaction
keep records for his father. He went to missionary school in the region and ultimately to the University, became an instructor there and then came to Berkeley in the U.S. for his Ph.D. He now has a professorship in the United States and specializes in issues of literacy. Niyi and I now move in the same circles and are part of the same academic community. I talk with him about a range of academic and other issues in ways that I cannot with many others in the U.S., and would certainly never be able to talk to his father or anyone else in Nigeria, although Niyi can. Through participation in similar “networks of relationships” over time, we have been socialized into similar network-speciWc communicative practices. Although our backgrounds are about as diVerent as they could be, we share certain communicative conventions and interpretive practices. It is long-term exposure to similar communicative experience in institutionalized networks of relationship and not language or community membership as such that lies at the root of shared culture and shared inferential practices. In most people’s lives, community membership is of course directly linked to participation in such networks of relationships, but in our post-industrial worlds, it is less and less possible to take this for granted. Apart from cultural sharing in the face of diVerences in background, there are now more and more cases of people brought up in what by ordinary criteria counts as the same community, but in whose case surface similarity of language and background hides deep underlying diVerences. To go back to what I said about shared interpretations. It is not merely a question of what something ‘means’. Ultimately, agreement on speciWc interpretation presupposes the ability to negotiate repairs and re-negotiate misunderstandings, agree on how parts of an argument cohere, follow thematic shifts and shifts in presuppositions, that is, share indexical conventions. The more basic issue is to show how these tasks are accomplished. And it is for this reason that my analysis puts so much stress on contextualization processes. C.L.P.: You start from the idea of “metamessages” à la Bateson or Watzlawick, and you seem to give great importance to these repair/adjustment procedures via metacues, cues concerning other cues, or metapragmatic cues. J.J.G.: Actually, the Wrst paper I did on this topic was a paper I gave in Urbino. C.L.P.: It was published in 1974. J.J.G.: In the 1974 series, yes (Gumperz 1974). That was my Wrst, but I didn’t have the theory then. But now...
A discussion with John J. Gumperz
C.L.P.: As to the metapragmatic activity, I think it is important in order to survive. If you don’t have metapragmatic strategies ... J.J.G.: It is not only that you don’t survive, you don’t learn. You cannot proWt from your own misunderstandings. And this clearly has to do with verbal ability at the level of communicative practices, because it is on such practices that our ability to assess and evaluate the signiWcance of what we perceive rests. C.L.P.: In psychiatric cases, it’s just these metapragmatic procedures that fail. You try, you try, and the other... J.J.G.: Exactly, exactly. I have had students who worked with family psychiatry transcripts, from the Watzlawick school. And although no complete analysis was done, it would seem that even in those cases, much could be learned from close, turn-by-turn analysis of communicative practices. C.L.P.: Another point. You mentioned GarWnkel and you said, “I am a little bit in between GarWnkel...” J.J.G.: And GoVman. C.L.P.: And you said that from GarWnkel you took the problem of inferential processes. J.J.G.: That’s right. C.L.P.: So let us discuss, if you agree, the problem of dialogue and reasoning, the reasoning required by dialogue. Because dialogue requires reasoning. J.J.G.: That’s right, you’re always reasoning. Conversational inference always involves reasoning in some form. You start — I’ll use the term ‘assessing’ or ‘assessment’ — you start by assessing in order to be able to recall a message. You always ask yourself both, “What are they saying?” and “What do they intend?”. Assessments begin at the level of phonetics, and segmental phonemes are of course important — you need to be able to assess whether for example you are hearing ‘g’ or ‘k’. But rhythmic organization and prosody may be equally important in making the relevant assessment. In fact, in my Urbino paper, which formed the basis for the Wrst chapter of Discourse Strategies, I discussed the example of a graduate student who asked me after class, speaking in standard-like style, “Can I come to see you?”.
19
20
Language and interaction
I replied, “Sure, come to the oYce,” whereupon he went on to say in what seemed like old-fashioned Afro-American English, “Ahma git me a gig.” At Wrst I wondered why he chose to switch codes in this academic context. He usually spoke much like everyone else, so the switch must have been intentional. Was he rejecting the ‘academic norm’ of Standard English? It was only after discussing the incident with several other graduate students familiar with Afro-American interpretive traditions that I recognized that, by speaking the way he did, he was signaling that he was no longer addressing himself just to me, but to anyone who shared his interpretive conventions, that is, mainly — but not exclusively — the other African-Americans in the group. What he was trying to convey can roughly be paraphrased as: “I am doing something that minority people like myself have to do, get support where I can.” In other words, when his utterance is interpreted in terms of what we know about the positions of African-Americans in the urban U.S., and in terms of cues such as the undiphthongized lengthened ‘ah’ instead of ‘I’ and the lengthened vowel in ‘gig’, as well as the highly-contoured intonation he used, it becomes clear that he was employing a formulaic expression, which, in the context at hand, indexically conveyed his message. He was not ‘code-switching’, but using metapragmatic strategies to convey a message. In a related example from Discourse Strategies, a young elementary school student, when asked to read, replied, “Ah cain’t read.” The teaching aid thought he meant to say that he was not able to read. Examples like this have often be cited in the literature on classroom learning in support of assertions that African-American students have more diYculty with literacy learning than others. But when we discussed this example with a group of African-American graduates, they pointed out that the expression carried contoured intonation and, given the expression’s positioning after the tutor’s question, they would interpret the student as saying essentially, “I don’t want to do it right now. I want company in reading.” On the basis of experiences like these, I became alerted to the fact that for those familiar with AfricanAmerican conventions, the opposition between contoured and non-contoured intonation may be information-carrying. Such uses of prosody are not only found in African-American speech. Consider the following example I recently heard in Cambridge from someone talking about King’s College: “Fellows of King’s are well-known, to fellows of King’s.” The second (italicized) phrase is set oV from the Wrst and therefore foregrounded by lowering of pitch and volume, and this suggests that “well-known” is restricted so as to highlight the interpretation that the compliment phrase applies only to
A discussion with John J. Gumperz
other fellows of King’s, not to the public at large. Understanding, as structuralists have taught us, always relies on selective perceptions based on our knowledge of oppositions. This is true for indexicals as well as for symbolic relationships. All interpretive assessments are relational (note the Saussurian heritage here). They are made with reference to something else, not necessarily directly represented in talk. But nevertheless, as the examples show, assessing involves reasoning which is intrinsically dialogic in the sense that positioning within an exchange is crucial. C.L.P.: Two other questions. What do you think about what could be called “corpus pragmatics”? And the second question, what do you think about the notion of “pragmatic creativity”? J.J.G.: Corpus pragmatics? C.L.P.: Or corpus linguistics, in this case for pragmatic analysis. J.J.G.: Okay, we always rely on a corpus, in the sense that to analyze anything at all in any depth we must prepare written transcripts. My problem with corpus linguistics is that it treats talk as if it were a literary text. And as to corpus pragmatics, pragmatics always requires us to take context seriously. This suggests we cannot base our analyses on a single corpus, we must work comparatively. We must Wnd ways of systematically contrasting our analysis with other comparable ones carried out under diVerent but comparable contextual conditions. This requires us to address the question of the criteria we use for determining similarities and diVerences in and across contexts. A second question that arises, if we take seriously the need for comparative analyses in pragmatics, is the distinction between communicative practice as a form of action and language as linguistic form. This is an issue that has only begun to be addressed in current comparative research in linguistic anthropology, but so far it has not, as far as I know, been systematically considered in corpus linguistics or, for that matter, in pragmatics. C.L.P.: When talking about corpus linguistics, I have in mind the rather recent wave of electronic corpora. J.J.G.: But you mean ...? C.L.P.: Sinclair, for example, John Sinclair’s work. J.J.G.: John Sinclair, yes, and the Lund people.
21
22
Language and interaction
C.L.P.: For example, yes. But the problem, from your perspective, is how to record the contexts as well... J.J.G.: That’s right, exactly. C.L.P.: That is the question. J.J.G.: And also, it is not the corpus as such that is the problem. It is a question of context, as well as of more basic questions of analysis. Take transcription. We have as yet no generally agreed-upon, universally applicable system of transcription. Not even something like the international phonetic alphabet. And transcription, as recent work has shown, must always be related to analysis. For conversation, the so-called Gail JeVerson transcription system has come to be widely used throughout the world. But in my terms, it neglects communicatively signiWcant prosodic and paralinguistic aspects of speech. It cannot, for example, account for the interpretive import of phonetic variability and so on. There is also no agreement on methods of analysis. Quantitative methods are widely used, but in the absence of agreement on what is signiWcant and countable, quantitative methods are limited in value. Therefore, I currently do not Wnd it productive for the issues that concern me. There has to be a division of labor. C.L.P.: Sure. It’s a problem I’ve found, too. When I have a number of cases before me, I am faced with a set of pairs of text and context. But this is much more than a corpus of texts only. Instead, when you have an electronic corpus, you have a set of utterances only, you can enumerate their constituents, their co-occurrence, and so on. J.J.G.: There are quite sophisticated methods of computer-based scanning and retrieval that work on raw data. But we need agreement on what to look for and what the goals of the analysis are. Most discourse analysts work with lexicon and clause level grammar and are concerned with issues of structure. C.L.P.: Okay, it’s possible to carry out all kinds of search procedures, but given this possibility, in your opinion, what’s the sense of these procedures with regard to the problem of context? J.J.G.: Such analyses cannot account for the eVect of context on interpretation. You always need a prior analysis of context. You need independent, ethno-
A discussion with John J. Gumperz
graphically-based analysis and information on the relationship between linguistic forms and discursive structures. C.L.P.: I think that as an analyst, an imperfect analyst, I always have to resist the temptation of omniscience and to try to reconstruct. But what does it mean to be an analyst, even if consciously not omniscient, and to re-construct? It depends on the cues which are at your disposal. J.J.G.: You need more background information than you ordinarily have as an investigator who simply elicits and transcribes talk. This information is best collected through ethnographic observation and participation in everyday routines such as those we relied on in the classroom interaction study I talked about at the beginning of this interview. It is also useful to make preliminary analyses and test one’s assumptions about interpretive conventions through informal discussions of analyzed texts along the lines of the procedures I discussed in Discourse Strategies. Ideally, such ethnographic procedures should accompany or precede large-scale text analyses. C.L.P.: So, can you have a corpus of texts with contexts? J.J.G.: I don’t know if I would use the term “corpus” here. There are ethnographic studies of communicative practices. Marco Jacquemet’s (1996) book on the Naples Camorra trial is such a study and should provide an analyst with at least some background information on discursive practices in the area of Naples. Another book in my Cambridge series, Linda Young’s Crosstalk and Culture in Sino-American Communication (1994), is also very useful in showing how the Chinese frame their interactions. C.L.P.: But if you try to reconstruct, what kind of generalization can you reach? For instance, there is a very interesting example of yours quoted by Deborah SchiVrin in her book Approaches to Discourse (1994:100), the example of the teacher and the child. When the child replied, “I don’t know,” it was a sort of cue, as if to say, “Please give me time.” J.J.G.: “I don’t know”? C.L.P.: “I don’t know,” in this sense, a sort of hesitation, conveying the wish to be encouraged. But the teacher took this kind of utterance as meaning: “I really have no knowledge concerning that.” J.J.G.: Exactly, yes.
23
24
Language and interaction
C.L.P.: Okay, this is one case. You were able to reconstruct the dynamics of this case... J.J.G.: Only because I happened to have a graduate student with me who was familiar with the relevant conventions. But one or two instances are only suYcient for initial hypotheses. After that I had to go on to Wnd other instances of the same phenomenon in order to construct a more general picture. That is the way ethnography works. It is a matter of observation, hypothesis formation, hypothesis testing on the basis of further information and then critiques from locals who are familiar with the situation, and so on. C.L.P.: Yes. But how can I continue? For example, I have tried to examine a set of already published examples, already studied by other people. J.J.G.: That’s very hard. C.L.P.: And the problem is that you can reach diVerent conclusions, diVerent hypotheses, even if you are still attached to the same presuppositions and assumptions, because you can attend to or focus on diVerent subsets of cues within the same overall set of cues. J.J.G.: One solution is to have a local assistant who works with you on the analysis and who will tell you what it is in what he/she hears or perceives that leads to the interpretation. That gives you information at two levels, content and form. The more one works with such interpretive analyses, the more native-like one’s interpretations can become. C.L.P.: But you can reach diVerent conclusions, given the same inputs or cues, even if you are a native interpreter. So I think that the enterprise of pragmaticians is a sort of game whose rules still have to be determined. J.J.G.: It’s only a game if you share the rules, otherwise it can’t be a game. It’s the kind of game ethnographers must learn to play. And many of them fail to do it. The trick is to learn to ask the right kind, i.e. productive, questions. Questions that lead to answers that give one the feeling that progress is being made even though many problems remain. C.L.P.: In your opinion, what kind of generalizations can we reach, starting from these cases? Because they are single cases. J.J.G.: Ultimately, large-scale quantitative studies are necessary. But current studies of this kind tend to work with many unexplicated assumptions. Good,
A discussion with John J. Gumperz
valid ethnographic information can lead to more sophisticated and informative quantitative analyses. A basic problem in sociolinguistic research is how to deWne the domain of our analysis. In the case of standard languages like English or Italian, the problems are relatively minor. But how do we deWne what is or is not Romansch or Piedmontese, apart from basing our sampling on a geographically-bounded region? In discourse analyses of such genres as classroom lecturing, similar problems arise. Initially, students of classroom discourse simply asked teachers to lecture on speciWc topics and recorded what they did. Obviously, that kind of procedure raises problems about the generalizations one can derive from one’s data. At the other extreme, some researchers simply turn on the machine and record everything that is said. This raises other analytical problems. All these are problems that can be avoided or at least clariWed given basic background information and theoretical insights into how communicative practices work. C.L.P.: I think that you adopt a conception which is sociocognitive. Aren’t you afraid of the mentalistic objection? J.J.G.: No, because my analyses are based on empirical and I hope replicable data. I talk about inferences, but these are grounded inferences, based, as I said before, on two sorts of data, at the level of form and at the level of content, as well as on background information on interpretive practices, and symbolic and indexical signs. If it can be empirically shown that the inferential processes I postulate also hold true in other comparable contexts and are similarly grounded in verbal signs, I can make Boolean generalizations about relations between classes of objects. But I cannot make valid predictions about speciWc interpretations. Dialogue analysts are right. We are limited in some ways, but maybe not in the way they say we are. C.L.P.: As to the second question, what do you think about pragmatic creativity? J.J.G.: Pragmatic creativity has to refer to innovation in the context of certain constraints. It’s akin to creativity in poetry or writing. Writers can be creative or crazy. C.L.P.: Craziness is maybe a kind of creativity. J.J.G.: We often assume that people are “crazy” when we fail to see order or coherence of some kind in what they say or do. Creativity bends boundaries without violating our sense of order.
25
26
Language and interaction
C.L.P.: So, in your sense, pragmatic creativity can be a sort of violation of conventions? J.J.G.: Yes, it can be understood in terms of a broader set of conventions. These apparent violations can be understood in terms of our broader sense of order. C.L.P.: Okay, here I think Mr. Grice enters the scene... J.J.G.: Yes, given a broad interpretation of Gricean inference. C.L.P.: ...with “violation” and “Xouting”. From your perspective, how is it possible for conversationalists and analysts facing some inputs, some cues and so on, to decide whether they are creative or not? J.J.G.: That always depends on your background and speciWc circumstances. C.L.P.: I see it as the problem of the existence of unconventionality. You are always faced with conventions, maybe conventions coming from another set or system of conventions. You can take them as new things or violations, but in fact they belong to other sets of conventions. But consider someone who wants to be unconventional within the same community, that is, within a set of established conventions. In your opinion, what does it mean to be unconventional? J.J.G.: To do something that seems to violate one rule, but that can often be explained in terms of another, broader principle. C.L.P.: In your opinion, can we capture violation with procedures, with strategies? J.J.G.: Yes. Irony and humour, as Grice points out, often depend on such violations. I gave an example of irony in my article in Auer and di Luzio’s book, The Contextualization of Language (Gumperz 1992). Three graduate students are talking about why they chose the subjects they are studying and one of them, hearing that the other is going into engineering, exclaims, “We Asians.” To me as an analyst, the utterance seemed strange and certainly not on topic. I only found out afterwards, when I inquired, that his reaction to his friend’s statement was an ironic allusion to a stereotype about Asians, who are aware of the so-called glass ceiling which keeps them from being promoted to important managerial positions and therefore prefer to stick to technical
A discussion with John J. Gumperz
occupations. In a way, the graduate student was being creative in calling his fellows’ attention to a problem they all face. C.L.P.: A last question. In your career, have you ever tried to describe a frame, a script, or something like that? J.J.G.: Bateson likens framing in communication to framing in paintings. GoVman talks about how frames aVect the way we perceive an interaction and how they change in the course of the interaction. For me framing is another way of talking about the activity level presuppositions that aVect interpretations at any one point in an exchange. In my studies of South Asians living in England, I have shown that frequently they perceive and frame encounters quite diVerently from their native-speaking interlocutors. In one particularly striking case, an interviewee seemed quite unreasonably to be denying or contradicting information that the interviewer told him, even though only the interviewer had Wrst-hand knowledge of the topic. It took us a long time to discover that the interviewee was simply pleading for sympathy or understanding. Based on these and a series of other miscommunications, I was able to argue that many South Asian speakers have ways of framing formal interview-like encounters which are quite diVerent from those Westerners are familiar with. A “frame”, as I use the term, can be described as a class of related presuppositions that guide conduct in certain situations and that are directly related to choice of topic and verbal strategies. I gave an example of this in my paper in Mutualities in Dialogue (Gumperz 1995a). C.L.P.: You said (in Gumperz 1995b) that Foppa, one of the editors of Mutualities in Dialogue, asked you something about your paper. J.J.G.: He asked me a question, one that I frequently hear: “How do you know if people understand each other? When do I know that my interpretation is the right one?”. And I said, “Well, only through context do you have the right interpretation, and you’ll reach an agreement, but only as the result of a dialogical process.” C.L.P.: How can one study the cognitive aspects of this process? J.J.G.: This is actually what I will talk about tomorrow.3 I will take two examples. There’s a rape case, a re-analysis of Paul Drew’s example from a rape case, and I’ll talk about some of the grammar that’s involved there. I’ll talk about what each person has to know in order to do what they’re doing. And
27
28
Language and interaction
then I’ll take my example from a job interview to show again that they did not share that knowledge, that therefore they couldn’t understand, and the more they tried, the more things went wrong. C.L.P.: That’s very interesting. A.d.L.: Just to go back to the notion of genre for a minute: it is not very clear to me. Is it a metapragmatic category? J.J.G.: We use the term “genre” to label, that is, to refer to people’s modes of talking. In a way genre is similar to what some people call speech style. But genres are always associated with speciWc ideological values. As Hanks and others have pointed out, by using speciWc genres, we allude to the values that are associated with them. A.d.L.: Yes, that’s what I thought. J.J.G.: But the only thing is that I talk from the speciWc standpoint of somebody who has either heard about it or analyzed it, and I am aware of what other people think about it. A.d.L.: I understand it as a sort of art of organizing activity in a special way, as Luckmann says, “to solve a communicative problem”. J.J.G.: Yes, that’s right. We organize activity to solve a communicative problem, but I would limit it. I would say that we do that in an activity, but genre is the way we talk about it, and the way it gets transmitted and passed on. And in fact I would say, when we talk about dialect, for example, Konstanz dialect versus Kreuzlinger dialect, we’re making a judgement. That is, a judgement at the level of genre. And what I would argue is that any attempt to delineate the notion of dialect and style empirically, in terms of some kind of Saussurean or empirical analysis, is not going to be very successful. You don’t need agreement on the linguistic facts of what a genre is, as long as there’s some sharing of certain key elements that would be the prototypical judgements that we’re making, not a systematic analysis, that’s the point.
Notes 1. This discussion took place on March 31, 1995, at the Department of Linguistic and Oriental Studies, University of Bologna, Italy, on the occasion of the 1995 Bologna Round
A discussion with John J. Gumperz
Table on Dialogue Analysis: Units, Relations and Strategies beyond the Sentence, organized by the International Association for Dialogue Analysis (IADA) in honour of its President, Sorin Stati. Transcribed and edited by Susan Eerdmans. Revised by John J. Gumperz and received February 1996; revised for the present edition February 2002 [Editors’ note]. 2. Gumperz is referring to the paper which Ionescu-Ruxandoiu Wrst presented at the 1995 IADA Round Table as “Pre-sequences as a conversational strategy”, now published as Ionescu-Ruxandoiu (1997) [Editors’ note]. 3. Gumperz concluded the 1995 IADA Round Table with a seminar on his work in interactional sociolinguistics (see Eerdmans, this volume) [Editors’ note].
References Eerdmans, Susan L. “A review of John J. Gumperz’s current contributions to Interactional Sociolinguistics”. This volume. Gumperz, John J. 1974. The Sociolinguistics of Interpersonal Communication. Urbino, Italy: Centro Internazionale di Semiotica e di Linguistica. Working Papers and Prepublications, 33, serie C. Gumperz, John J. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Gumperz, John J. 1992. “Contextualization revisited”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 39–53. Amsterdam/Philadelphia: John Benjamins. Gumperz, John J. 1995a. “Mutual inferencing in conversation”. In Mutualities in Dialogue, I. Marková, C.F. Graumann and K. Foppa (eds.), 101–123. Cambridge: Cambridge University Press. Gumperz, John J. 1995b. Personal communication to C.L. Prevignano, Bologna, 31 March. Gumperz, John J. 1996. “The linguistic and cultural relativity of conversational inference”. In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 374–406. Cambridge: Cambridge University Press. Gumperz, John J., Jupp, Tom C. and Roberts, Celia. 1979. Crosstalk: A Study of CrossCultural Communication. London: National Centre for Industrial Language Training in association with the B.B.C.. Hanks, William F. 1996. Language and Communicative Practices. Boulder, CO: Westview Press. Ionescu-Ruxandoiu, Liliana. 1997. “Pre-exchanges as a conversational strategy”. In Dialogue Analysis: Units, Relations and Strategies beyond the Sentence, E. Weigand (ed.), 121–128. Tübingen: Niemeyer. Jacquemet, Marco. 1996. Credibility in Court: Communicative Practices in the Camorra Trials. Cambridge: Cambridge University Press. SchiVrin, Deborah. 1994. Approaches to Discourse. Oxford: Blackwell. Young, Linda W.L. 1994. Crosstalk and Culture in Sino-American Communication. Cambridge: Cambridge University Press.
29
30
Language and interaction
Contextualizing “contextualization cues”
Chapter 3
Contextualizing “contextualization cues” Stephen C. Levinson
I would like to focus my remarks on John Gumperz’s notion of a contextualization cue, which I think is central to his current program. However, before I begin, I would like to make some more general remarks on that program and its place both in Gumperz’s overall oeuvre and in discourse studies generally. As di Luzio sketches in his contribution to this volume, Gumperz had a previous incarnation before becoming the founding father of interactional sociolinguistics; the previous avatar was of course one of the foundational spirits in the broader Weld of sociolinguistics itself. As the title of his collected papers of that vintage suggests (Gumperz 1971), that previous self was (amongst other things) interested in how social groups express and maintain their otherness in complex societies. Gumperz started as a dialectologist interested in tracking down the forces of standardization and particularly those of diVerentiation, and it was the search for where these forces are located that has led him inexorably from the macro-sociological to the micro-conversational perspective; it was a long journey from the study of regional standards, to ethnic groups, to social networks, to the activation of social boundaries in verbal interaction, to discourse strategies. Readers will not understand his work if they view it just as the study of conversation. He has not abandoned his interests in the macro — it is the large-scale sociological eVects of multitudes of small-scale interactions that still partially fuel his preoccupations with conversation (see his contributions to Gumperz and Levinson 1996), most evident perhaps in his concern with the plight of the individual caught up in these large-scale forces. Those seeking to understand his attempt to link macro and micro aspects of language use may still usefully turn to his studies of code-switching, where many of his current ideas were adumbrated (see especially Blom and Gumperz 1972). It was clear to him, for example, that social situations partially determine the choice of code, and yet that code-
31
32
Language and interaction
choice can partially determine situation — and this play between the presupposing and creative aspects of linguistic choice (to employ Silverstein’s terminology) still dominates his work. This background will also help to explain what may otherwise be taken as a demerit: Gumperz’s analyses of conversations have nothing of the theoretical cleanliness to be found e.g. in conversational analysis. His tools are eclectic, and the toolbox cluttered, on the one hand with pragmatic notions like implicature, speech acts, frames, activities, cues, indices, and the like, and on the other with sociological notions like network, ethnicity, gatekeepers, habitus and so on. He is trying to depict processes that still defy understanding with the best tools that come to hand from whatever school of analysis. And he is trying to connect levels of analysis, from macro to micro, rather than to develop an isolated level of conversational analysis. This is the real diVerence, it seems to me, between his approach and e.g. conversation analysis, rather perhaps than the one he oVers in the interview (in terms of his own preoccupations with “situated on-line interpretation”; elsewhere he oVers a characterization of his own work in terms of the “social import of the Wne details of verbal communication”, which seems more accurate). With that as background, let me turn to the subject of Gumperz’s notion of contextualization cues, which seems to me the central innovation in his analysis of discourse. First, some remarks on the motivation. At the time he was developing these ideas, I was lucky enough to be a graduate student in the Language Behavior Research Lab at Berkeley where he worked. In Berkeley at that time there was a rare and wonderful conXuence of ideas from diVerent disciplines concerning the study of meaning — in philosophy, Grice and Searle were expounding the ideas about implicature and speech acts now associated with them, Fillmore was preoccupied with indexicality in language, Kay with its sociological import, Robin LakoV with contextual meaning, and George LakoV was attempting to wrap it all up in a uniWed theory of generative semantics. It was an era of optimistic open-mindedness in which, for example, Harvey Sacks could give an extended series of lectures to the Linguistics Institute of the Linguistic Society of America. In Gumperz’s lab, we were taping whatever we could get access to, and taping also in Weld locations around the world, and we were trying to apply these theoretical ideas to the analysis of actual snippets of conversation. It wasn’t easy. There was a yawning gulf between what, on a simple-minded analysis, ‘the words meant’ and what we took the participants to be self-evidently doing with their words. The gulf was
Contextualizing “contextualization cues”
not to be bridged completely with the apparatus at hand — implicatures, indirect speech acts, frames, and the like. This was the problem that preoccupied Gumperz and his students at the time.1 In many ways, Gumperz’s approach to discourse analysis still has this interpretative gulf as its primary theoretical target. He has since tried to narrow the gap in a number of innovative ways, which are the hallmark of his current approach. One line of attack was the careful analysis of prosody, the neglected acoustic cues that might help to explain how we can possibly mean so much by uttering so little. Another line was the apparently paradoxical idea that utterances could somehow carry with them instructions about how to build the contexts in which they should be interpreted. The two were combined in the idea of a contextualization cue, often (but not necessarily) a prosodic trigger that in conjunction with lexical material will invoke frames and scenarios within which the current utterance is to be interpreted as an interactional move. I would like now to attempt to further clarify the notion of a contextualization cue, which is perhaps better exempliWed rather than analytically explicated in Gumperz’s work.2 There are two issues in particular worth exploring. One is what the notion presupposes about the nature of context. Another is the role of an implicit distinction between foreground and background information in messages. Out of these considerations should emerge a clearer delimitation of the notion of a contextualization cue. For current purposes I will assume a deWnition of context as a set of propositions taken for granted by the participants3 (See especially Duranti and Goodwin 1992). Let me turn then Wrst to the apparent paradox that utterances can create their own contexts. The paradox would be: if it takes a context to map an interpretation onto an utterance, how can we extract a context from an utterance before interpreting it? The idea that utterances might carry with them their own contexts like a snail carries its home along with it is indeed a peculiar idea if one subscribes to a deWnition of context that excludes message content, as for example in information theory. Context is then construed as the antecedent set of assumptions against which a message is construed. But it has long been noted in the study of pragmatics that this dichotomy between message and context cannot be the right picture. We may want to say that use of an expression of the form ‘The so-and-so’ presupposes the mutual knowledge that there is one relevant so-and-so, but then Wnd that in fact a standard way of informing an interlocutor that there is such a so-and-so is to say e.g.,
33
34
Language and interaction
“The King of Tonga will begin his state visit on Monday.” Some have argued that this is as it were an abuse of the conventions governing the use of deWnite descriptions which forces the interlocutor to “accommodate” the utterance by interpreting it as if it were mutual knowledge that there is a King of Tonga. But in fact there are many other (non-presuppositional) devices that work in this way — conventional implicatures associated with expressions like ‘but’ or ‘even’ force the creative construction of a context for the interpretation of the utterances they occur in. “Even Harry will come” projects a context in which it is assumed that there is a ranking of people in terms of their likelihood of coming, and Harry is very low on that scale. Or suppose I say, “What are you doing tonight?” — this projects some kind of invitation or request sequence in the oYng. So the idea that utterances can carry their contexts with them, that is, the set of assumptions necessary to unpack their interpretation, is not as outlandish as it may seem at Wrst sight. The paradoxical quality of the idea of a contextualization cue is as much due to our wanting to hang on to the simple information theoretic view of what a context is. But still, there is a puzzle: how does it work? Here both the phenomenon and the theorists part company. Perhaps we can distinguish diVerent species of context-invoking aspects of utterances:
1. Context-importation by conventional coding devices Examples would be presupposition triggers like deWnite articles, expressions which carry conventional implicatures like ‘even’ or honoriWcs in those languages that have them. Contrastive stress also Wts here: “It wasn’t me that did it” projects a context in which someone is supposing I was indeed responsible. Many languages have particles or other morphemes that serve the same sorts of function: indicating that something is already presumed, or that no-one assumes it, or that it is being introduced for discussion, or that its veracity is in doubt, etc.. Interacting closely with the central meaning (e.g. truth-conditional content) of the utterance, these conventional codings signal extra propositions that form the background to the interpretation of the utterance.
Contextualizing “contextualization cues”
2. Context-invocation by inference alone Classical Gricean implicatures would be the prototype here: (1)
A: Hey, how about supper together? B: I have a jealous husband
where in order to infer a (negative) response to the invitation, it is necessary to invoke a large series of contextual assumptions of the sort: “One does not choose to make one’s husband jealous”, “Eating supper is an intimate act” and the like.
3. Context-invocation by “cue” or “Xag” This is, I take it, the Gumperzian notion, in which the term “cue” denotes an encoded or conventional reminder, like a knot in a handkerchief, where the content of the memo is inferentially determined. Thus the “cue” cannot be said to encode or directly invoke the interpretative background, it’s simply a nudge to the inferential process. Moreover, the interpretative process is guided more by a series of nudges now in one direction and now in another — thus “cues” come as complex assemblages where the result of the whole assemblage cannot be equated with the inferential results that each part alone might have. The interpretive process may be guided by general pragmatic principles of a Gricean sort, and thus be in many ways universal in character; but the “cues” are anything but universal, indeed tending toward sub-cultural diVerentiation. Hence the Gumperzian perspective on communication: at once potentially possible across cultural divides and inevitably thwarted by cultural nuances. Further insights into how contextualization cues work, and their place in an overall pragmatic scheme, may be found by turning to the second issue, the relation between “foreground” and “background” in message structure. By these terms I mean something entirely pretheoretical, the opposition between central message content, coded propositional information, and peripheral, more loosely associated and less clearly formulatable information, a sort of informational penumbra. The opposition has aspects at diVerent levels: form, content, and cognitive saliency (Table 1).
35
36
Language and interaction
Table 1. The opposition between central and peripheral information. Foreground
Background
Form
lexico-syntactic
Content
propositional
Cognition
communicative salient conscious
particles, modiWers, prosody, kinesics general/vague or non-propositional meta-communicative inconspicuous unconscious
This set of alignments is the intuition in the Batesonian analysis of communication, with the diVerentiation of channels between conscious ‘verbal’ coding and less conscious, less fully-coded prosodic and paralinguistic channels, carrying with them some diVerentiation of function. Silverstein (1981) has oVered some further elaboration of the relation between formal factors and cognitive saliency: he suggests that certain factors tend to render aspects of messages “out of awareness”. Thus non-segmentable, discontinuous, noniconic forms are likely to be inaccessible to native intuition; they will also tend to be associated with non-referential, contextual content. The message vs. context opposition is, as we have noted, a false opposition: the message can carry with it or project the context. But we tend to hang on to the opposition because we focus on the foreground or message content, and it is the background that tends to project the context. “Contextualization cue” is one of a number of terms of art that attempt to explicate this relation between message-background and context-projection. The hypothesis may now, I think, be clariWed by suggesting that contextualization cues form a natural class with the following cluster of properties:
Contextualization cues Formal properties 1. a tendency towards non-segmentable, prosodic, paralinguistic, or kinesic features; 2. if cued in lexico-syntax, then by lexical alternate (register) or minor grammatical class (e.g. particles); 3. any one clear function associated with a whole cluster of disparate features (cf. Silverstein’s discontinuous feature), such features often being cross-
Contextualizing “contextualization cues”
channel (a constellation of e.g. kinesic, prosodic and lexical features). Content properties 1. “out of awareness” background features; they are context-invocative, and cannot therefore be easily directly responded to; 2. non-propositional content, e.g. aVectual, rhetorical, social or metalinguistic; 3. tendency to invoke holistic bodies of assumptions (contextual “frames”), which then play a role in the interpretation of the utterance — e.g. help select reference, clarify rhetorical structure, indicate illocutionary force, etc.; 4. content not really coded, but “cued” — i.e. reliant on large dose of inferential reconstruction; thus the inferred content of the same cues can be diVerent in diVerent utterances. The reasons why Gumperz has made this notion central to his analysis of discourse should now I think be clear. It is because he is interested in the relation between the micro- and the macro-sociolinguistic that contextualization cues have a special interest. First, their “out of awareness” features, coupled with their essentially arbitrary but loose association with formal cues, mean they can only be learnt by rich exposure to a communicative tradition, a deep immersion in social networks. This takes us back to the earlier Gumperz with his interest in social groups and their networks. Second, in the workings of contextualization cues can be found the springs of dialect and group diVerentiation: we pseudo-speciate by slow degrees that start here in subtle miscommunications. In Labov’s view, in contrast, sociolinguistic diVerences are little carriers of prestige or stigma. In Gumperz’s view, the smallest formal diVerences may carry with them a chasm of incomprehension, because contextualization cues invoke the essential interpretive background for the foregrounded message. This is where the barriers Wrst come up, later to grow into the saplings of dialects or the oaks of languages. This is also the last place for the barriers to come down, as demonstrated in his analyses of the minitragedies — the failed job application, the lost welfare beneWt — which have been a favourite theme of his recent work, and which can be seen to be an almost inevitable outcome of diVerentiated socialization. Gumperz is Americanized enough to put forward here an optimistic message about the educability of the gatekeepers in multi-ethnic societies; others will Wnd here a deeply
37
38
Language and interaction
pessimistic story about how elites guard access to social advancement. A Wnal remark. In the interview in this volume, one of the interviewers persists with a line of questioning that Gumperz seems to sidestep. If understanding is so complex, is it not ineVable? In particular, what gives the analyst the right to say, “A intends this, but B thinks he means that”? Gumperz suggests that it is all a matter of good ethnography: one asks A about what he meant, and one notes that elsewhere C said something similar with similar intent, and so on. But the interviewer, in good postmodernist style, is not so easily assuaged: after all, if contextualization cues are often ambiguous, even properly socialized ‘natives’ may understand diVerent things by the same utterance. Gumperz assents, but does not draw the postmodernist conclusion: he thinks that in certain cases, things are reasonably clear; he thinks he can demonstrate a recurrent tendency by those who share the right kind of network to associate a class of interpretations with a highly speciWc set of linguistic cues invisible to those who belong to other networks. Radical doubt does not assail him. In this, Gumperz is again a cheery optimist, and perhaps that stems from his own Wrsthand experience as a network hopper, who landed on American shores a refugee in his late teens. In the twenty years since we sat together in the basement that served as the Language Behavior Research Lab and pondered the gulf between the said and the unsaid, Gumperz has surely but steadily tried to build out the piers of an analytical bridge across it. He will be the Wrst to admit that the gap is still very much there: we have today at best only the feeblest of understandings of inferential processes in conversation. In a way, his own contributions only deepen the puzzle about how human communication is possible, by drawing attention to yet further layers of hidden, self-invoking machinery of the kind sketched in his contextualization cues.
Notes 1. My own response, with Penelope Brown, was to develop a theory of politeness that we hoped would help to bridge the gap between the said and the unsaid. Other students of his, like Susan Gal and Deborah Tannen, have developed their own responses. 2. One of the clearest expositions may be found in Gumperz (1992). 3. There are all sorts of things wrong with this, but that is another issue.
Contextualizing “contextualization cues”
References Blom, Jan-Petter and Gumperz John J. 1972. “Social meaning in linguistic structure: Codeswitching in Norway”. In Directions in Sociolinguistics: The Ethnography of Communication, J.J. Gumperz and D. Hymes (eds.), 407–434. New York: Holt, Rinehart and Winston. Gumperz, John J. 1971. Language in Social Groups. A.S. Dil (ed.). Stanford, CA: Stanford University Press. Gumperz, John J. 1992. “Contextualization and understanding”. In Rethinking Context: Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 229–252. Cambridge: Cambridge University Press. Gumperz, John J. and Levinson, Stephen C. (eds.). 1996. Rethinking Linguistic Relativity. Cambridge: Cambridge University Press. Silverstein, Michael. 1981. The Limits of Awareness. Sociolinguistic Working Paper 84. Austin, Texas: Southwest Educational Development Laboratory.
39
40
Language and interaction
Contextualization and social meaning-making practices
Chapter 4
Contextualization and social meaning-making practices Paul J. Thibault
1. Preliminary considerations My perspective is in broad sympathy with Gumperz’s attention to “the detailed analysis of communicative practices.” However, my purpose in what follows is both to suggest possible areas of theoretical convergence, as well as to suggest alternative theoretical solutions where I see these as appropriate. Gumperz draws our attention to the need for a theory of communicative practice, or, in the terms that I use, “social meaning-making practices” (Lemke 1990; Thibault 1991). This emphasis on the communicative practices of a community highlights the need for a theory of how linguistic and other semiotic forms are used by the members of a given community, and why, in the enacting of the social life of particular communities. To achieve this goal, Gumperz is primarily concerned with the analysis of instances of social interaction, or “speciWc speech events” (Prevignano and di Luzio, this volume: 8). He posits the need for some intermediate level of analysis which is able to bridge the gap, both theoretical and methodological, which has all too often seen the ‘linguistic’ and the ‘social’ banished to separate, seemingly incommensurate domains of inquiry. In so doing, Gumperz shows that the description of a community’s communicative practices cannot adequately be accomplished within the conWnes of any single discipline in the human and social sciences. Such an enterprise is necessarily a transdisciplinary one, drawing on the insights of sociology, ethnology, linguistics, anthropology, social psychology, and so on, in order to develop a uniWed conceptual framework for talking about social meaning-making (Gumperz 1992). Gumperz’s focus on communicative practices is always conducted on the basis of “independent, ethnographically-based analysis and information”
41
42
Language and interaction
(Prevignano and di Luzio, this volume: 22–23). This underscores the “semantic importance of context” (Prevignano and di Luzio, this volume: 9). The notion of “communicative practice” refers, then, to the real-time processes whereby participants construe semantic signiWcance in the unfolding speech event. Gumperz emphasizes how the participants — speakers, listeners, audience members — actively and jointly construct the speech event. Meaning-making is fundamentally dialogic in character. He criticizes the structuralist tradition of phonological and syntactical analysis in so far as this has failed to provide the analytical tools and conceptual frameworks necessary for the analysis of “everyday talk” (Prevignano and di Luzio, this volume: 8–9). In pointing out that “Saussurean phonological and grammatical structures, deWned in terms of Wnite sets of oppositions” (Prevignano and di Luzio, this volume: 9), cannot by themselves provide an account of communicative practice, Gumperz pays somewhat less attention to the importance of accounting for the semiotic resource systems which members deploy in their enactment of particular communicative events. Gumperz has written of the ways in which “the localized choice of one or the other of the paradigmatic set suggests diVerent implicatures in terms of which the utterance at hand is to be interpreted” (1992:46; di Luzio, this volume). However, there is no detailed speciWcation of these paradigmatic resource systems and their relations to actual instances. I shall now discuss this question.
2. Social meaning-making: Resource systems and practices How may the gap between the linguistic and the social in both linguistics and social theory be overcome? In order to answer this, the relationship between the semiotic resource systems of the community and its communicative practices requires rethinking. This is especially so if we wish to avoid being trapped by the abstract and de-contextualized formalisms in terms of which language has been characterized as a closed and self-regulating system of abstract forms. If the detailed analysis of communicative practices is to “illuminate basic issues in social theory” (Prevignano and di Luzio, this volume: 7) — that is, if it is to show, for instance, how social ideologies are constructed and maintained in and through speciWc meaning-making practices, then we also need an account of what it is possible to mean in a given community. In other
Contextualization and social meaning-making practices
words, we need a complementary account of the social semiotic resource systems that are available to the members of the community — the systems of relations of semiotic forms of all kinds to their possible meanings. That is, how do lexicogrammatical and other semiotic forms enable us to mean, and also, what is it possible to mean in and through their use in discourse? From the system perspective, the early attempts by Sapir and Whorf to construct an account of language form and function which is informed by the insights of cultural anthropology and Saussure’s theory of the system of value-producing diVerences in langue, far from subscribing to context-free rules according to the formalistic tendencies of early structuralism, already go in this direction (Thibault 1997). In his work on speech genres, Bakhtin emphasized that the study of “grammar and lexicon” (1986:66–7), on the one hand, and the study of the stylistics of the contextualized utterance, on the other, should not be opposed to each other. This is so, Bakhtin points out, because “the speaker’s selection of a particular grammatical form is a stylistic act” (ibid:66). The choice of particular options from the system of possible selections and their combinations always constitutes a response to speciWc contextual contingencies. In acknowledging this, Bakhtin further argues for the need to “organically combine” the two perspectives, without, however, losing sight of the methodological distinctions between them (ibid:66). Bakhtin’s solution to this problem lies in his notion of speech genres. Speech genres specify the typical ways in which the linguistic and other semiotic resource systems are deployed as regular, repeatable, semiotic action formations in the social practices of some community (Lemke 1988a; Thibault 1990). These fundamental insights of Bakhtin’s have been further developed in recent functional theories of lexicogrammar and discourse (Halliday 1994; Martin 1992). Bakhtin does not himself provide any detailed account of these resource systems. Further, the recent reception of Bakhtin in literary and cultural theory has tended to emphasize heteroglossic diversity and stylistic speciWcity at the expense of the typical and the systematic. Grammar and lexicon, or lexicogrammar, belong to the system of language. The stylistics of the individual utterance, to use the Bakhtinian terminology, refers to the speciWc text or occasion of discourse. Actual speech events — cf. utterances — are always speciWc semantic construals of the contexts which, as Gumperz points out, they help to construct and deWne. As such, speech events respond to speciWc semiotic and material contingencies, and in ways that are, at some
43
44
Language and interaction
speciWable level of delicacy, unique to the particular context. There is not, however, a direct, unmediated translation of the order of system — lexicogrammar — into speciWc occasions of discourse. Rather, genres are types. But they are types in a rather particular way. Genres do not specify the lexicogrammatical resources of word, phrase, clause, and so on. Instead, they specify the typical ways in which these are combined and deployed so as to enact the typical semiotic action formations of a given community. The ‘translation’ from lexicogrammatical system to speciWc speech event is always mediated by the typical patterns of language use — the primary and later the secondary genres — in the community. The notion of speech genre is an intermediate level of analysis in this sense. How are the resources of the system integrated in the joint enactment of some semiotic performance, or speech event? My starting point here is that all communicative events are multimodal semiotic performances (Lemke 1998; Thibault 1994; Van Leeuwen 1992). That is, they co-deploy in an integrated way the resources of diverse semiotic modalities — lexicogrammar, prosody, posture, kinesics, rhythm, and so on. The meaning of the overall discourse event cannot be localized in any single modality, seen as isolated from the others. Gumperz rightly rejects the notion of discourse as a mere string of clauses. Taken literally, such a view would amount to a linear and additive view of meaning-making. However, meaning-making does not occur on this basis. Rather, the resources of diverse semiotic modalities mutually contextualize each other in a multiplicative way (Lemke 1998). Language and other semiotic modalities are not, for this reason, independent of each other to start with. This is so both synchronically and diachronically: language both co-evolved with the other modalities of meaning-making and is co-deployed with these. A given discourse genre consists of typical copatternings of selections from a diversity of semiotic resource systems. In drawing attention to “communicatively signiWcant prosodic and paralinguistic aspects of speech” (Prevignano and di Luzio, this volume: 22), Gumperz points to the ways in which selections in one semiotic modality may aVect the meaning of a selection in some other, or how the new contextual redundancy which results from the co-patterning of the two may result in a newly contingent contextual meaning. That is, one which is not predicted by the usual patterns of the two selections, taken ‘separately’.
Contextualization and social meaning-making practices
3. Context-dependent and context-independent meanings Gumperz’s solution to these issues lies in what he calls a “move from a Saussurean to a more inclusive, broader, Peircian semiotics” (Prevignano and di Luzio, this volume: 10). In my view, the distinction between symbolic and indexical signs is itself problematic. Gumperz deWnes the notion of “contextualization cue” as a class of indexical sign. He contrasts the ‘pure’ indexicality or context-dependent character of these to the purportedly context-free lexical and propositional meanings of the symbolic sort. The distinction between symbolic and indexical signs thus rests on the assumption that there are two very general classes of meanings, viz. those of the context-independent sort and those which are context-dependent. Broadly speaking, the former rest on the assumption that there is such a thing as a context-free or ‘literal’ propositional meaning. There is an assumption of a Wxed, literal correspondence relation between the propositional content of the form and some extralinguistic state of aVairs to which the proposition refers. Indexicals, on the other hand, are seen as being, to varying degrees, non-propositional and context-dependent. But the distinction between symbolic and indexical signs is itself problematic in the following two ways. First, the notion of “context-free lexical meanings”, based on “well-known grammatical and lexical rules”, derives, in actual fact, from a speciWcally western linguistic folk-theory of reference (Silverstein 1987; Rumsey 1990) and the various philosophical reWnements of this.1 According to this folk-theory, some classes of surface segmentable lexicogrammatical forms — e.g. nouns — when abstracted from context, may be assigned a Wxed word-object correspondence relation. This is a system view of sorts in the sense that the word-object correspondence relation is seen to apply to abstract noun classes, abstracted from their contexts-of-use. Furthermore, it is a view which recognizes just one kind of structuring principle in language, viz. constituency, along with the bias towards typological-categorial meanings that this entails. In this view, the meanings of linguistic forms can be glossed according to context-free principles of referential equivalence between language form and extralinguistic states of aVairs. Language forms so deWned are explicitly meaningful for the insider-members of the culture, as witnessed by the way in which such forms may be assigned a dictionary meaning as lists of formal class items, irrespective of context-of-use. Now, meanings of the so-called context-dependent or indexical sort tend
45
46
Language and interaction
not to be so readily segmentable into discrete particles or constituents of isolable linguistic form such as words, phrases, clauses, and so on. They tend to be “analogic” and continuous, rather than “digital” and segmental, in character (Bateson 1973:261–2). Thus, they conform to quite diVerent organizational principles. These tend to be topological and continuous in character, rather than typological and discrete.2 From the perspective of insider-members, the resources of these semiotic modalities are ‘felt’ or ‘intuited’ as having implicit meaning in some speech event, even though members Wnd it diYcult to recognize or to gloss such meanings as being explicitly communicative. SpeciWcally, they cannot usually be analyzed as explicit chunks of digitalized experiential content. However, and as Gumperz’s work shows, prosodies and paralanguage are always communicative from the point of view of the analyst, who is committed to reconstructing the underlying principles of contextualization in discourse. The distinction between symbolic and indexical signs serves in my view to highlight the need to give equal analytical and theoretical value to both the typological-categorial distinctions in terms of which language, seen as autonomous with respect to other semiotic modalities, has usually been deWned, and the topological-continuous character of not only many other, frequently neglected dimensions of linguistic organization, but also of the other semiotic modalities with which language is always integrated in social meaning-making. Thus far, our theories of both language and other semiotic modalities have been overwhelmingly shaped in terms of the former. This has meant that linguistic structure has been seen as prototypically typological-categorial and digital in character, as witnessed by the almost exclusive reliance on partwhole constituent hierarchies in the analysis of linguistic form. At best, this is a singularly partial view of the object of study.3 The Wrst problem resides, then, in the implicit assumption that “contextfree lexical items” may be described from the perspective of the system of linguistic forms, abstracted from their speciWc contexts-of-use, whereas other semiotic modalities may not be so described because these are merely contextdependent indexes of the given situation. In this view, the latter would have no internal systematicity of their own. I doubt very much that this is Gumperz’s own position, though his focus on instances may be seen to invite this interpretation. From the system point of view, the meaning-making resources of all semiotic systems are analytically reconstructable as abstract systems of diVerences which cross-classify with each other in potentially very many diVerent
Contextualization and social meaning-making practices
ways (see above). There is no reason in principle why, for example, the gestural semiotic modality, intonation, or depiction cannot be described as systems of diVerences, or as systems of possible form-meaning relations, seen as the resources for making context-dependent meanings in these modalities. These may have proved intractable to the analytical methods and procedures of traditional (structuralist) linguistic analysis. However, there is no reason why a system of gestural diVerences, say, cannot be postulated so as to specify the possibilities for meaning-making in that semiotic modality. In any case, the assumption that lexical items can be speciWed on the basis of Wxed form-meaning relations is itself questionable. From the point of view of the system, semiotic forms of all kinds — linguistic, gestural, and so on — have a meaning potential (Halliday 1978). In fully contextualized discourse, there is a great deal of semantic variation in the use of the ‘same’ lexical item according to factors which may relate to genre, for example, or to the semantic relations that may be made between items in speciWc contexts. The system of diVerences, seen as abstracted from speciWc contexts, is variably deployed according to speciWc contextual requirements. The second problem concerns the folk-linguistic view in our (western European) cultures and their historical oVshoots that linguistic forms are only explicitly communicative for insiders if they can be glossed as having a Wxed, de-contextualized propositional meaning which functions to “refer to” designated states of aVairs in the real-world (Silverstein 1987).4 Now, from the point of view of the outsider-analyst of social meaningmaking, all semiotic forms are both meaningful and communicative, irrespective of whether they are explicitly communicative for insider-participants or not. Once we drop our folk-theoretical Wction that some classes of lexicogrammatical forms, prototypically nouns, may be assigned a Wxed referential meaning, irrespective of context, then it should become clearer that, in the context of a speciWc speech event, it makes no sense to uphold a distinction between context-independent symbolic signs and context-dependent indexical ones. If all semiotic modalities may, from the system point of view, be speciWed as systems of diVerences, abstracted from speciWc contexts, then it is no less important to show how it is through the co-deployment of these resources that participants jointly enact the speech event. It is generally accepted that intonation, say, can be analyzed as a system of potential meaning-making diVerences which systematically contrast with each other, viz. rising, falling, falling-rising, rising-falling tone, and so on. Abstracted from speciWc dis-
47
48
Language and interaction
course contexts, these can only have a meaning potential rather than a fully contextualized discourse meaning.5 On the other hand, all semiotic forms are indexical when deployed in some context. Further, the resources for indexing a particular form in a contextually speciWc way may themselves be speciWed systemically. For example, this is so of the often very complex systems of deictic categories in all languages. In English, for example, the systemically speciWable diVerences between ‘the’, ‘a’, ‘this’, ‘that’, and so on, specify how the Head noun in the nominal group is grounded in relation to some material or other feature of the situation which is made semiotically salient in the interaction. If I say, for instance, “Look at that man over there”, to index a particular man in the spatio-temporal purview of the speaker and listener, then I am using a selection from the system of deictic categories in English — ‘that’ — to index the perceived object as an instantiation of the type-category speciWed by the Head noun ‘man’ in the nominal group in question. Furthermore, I am doing so in ways which index the man in question as a semiotically salient feature of the interaction. The man is made a relevant part of the discourse context. From the system perspective, the noun ‘man’ construes a type-category which is internal to the grammar and semantics of the English language. As such, it does not refer to any speciWc ‘real-world’ man. It is only when it is grounded by the resources of deixis that the noun speciWes particular states of aVairs, objects, events, and so on as instantiations of the type-category. Further, there may be degrees of conformity to the criterial features speciWed by the typecategory (Langacker 1987:68–71). A noun, as a grammatical class item, speciWes a type-category of a semantic ‘thing’ which belongs to a given language system. Its indexical grounding with respect to a given material or imagined object, event, etc. means that the perceived or Wctive phenomenon so indexed through the resources of the nominal group is construed as an instantiation of the type-category. There is no direct word-object relation to start with. Nouns do not simply refer to objects, states of aVairs, etc. in a direct, unmediated way. Rather, they specify an experiential category which is internal to the language system. The diVerence between signs with experiential content (e.g. nouns) and those with none (e.g intonation) is one of degree. It is not an absolute distinction. In discourse, all signs function in one way or another to index the wider situational context. Both lexicogrammar and intonation can be analytically reconstructed as systems of contrasting form-meaning diVerences. In discourse,
Contextualization and social meaning-making practices
what is important are the cross-coupling patterns between (i) the diverse semiotic modalities in operation, and (ii) these and the material world in the enactment of the wider situational context (Lemke 1993:250; Thibault 1996:342). Informally, some examples are: 1. (nominal group / perceived object or event) // situation) 2. (lexicogrammar / intonation) // aVective-attitudinal orientation of speaker) 3. (phoneme / morpheme category) // semantic concept) These examples illustrate, albeit informally, how local cross-coupling patterns between, for example, a given nominal group and some perceived phenomenon in the material world, or between some choice in lexicogrammar and intonation, combine to index some aspect of the wider context which the combining of the two modalities helps to constitute. In each case, there is a Wrst-order contextual redundancy pattern between the selections from the two modalities. This is represented by the single slash, viz. ‘/’. Yet, in each case it is necessary to talk about a second-order relation which speciWes the wider context in which this particular combination has the meaning it does. This is represented by the double slash. But notice that the third of these three examples refers to the language system in the narrow sense. The same basic principle applies here too. Thus, the co-patterning of a particular phoneme category with a morpheme is a Wrst-order contextual relation on the basis of which a second-order semantic concept is construed in the language in question. In each case we see an important principle at work. This is the principle of ‘what goes with what else’ in the enacting of the wider situational context in which a particular combination has its meaning. An index was deWned by Peirce as a sign which is “physically connected” to some object (Nöth 1990:123–4). If indexicality is centrally concerned with the making and specifying of contextual relations, as Gumperz’s work shows, then this deWnition is no longer very satisfactory for a theory of discourse. The general problem of “indexicality” has to do with how particular cross-couplings among diVerent semiotic modalities and between these and selected aspects of the phenomenal world instantiate a local discourse context by specifying in various ways the parameters of the interaction itself.
49
50
Language and interaction
4. Discourse, context, and meaning-making practices: Towards an integrated account Gumperz draws attention to the shifting nature of participant interpretation in the ongoing speech event. In doing so, he identiWes a number of ways in which participants interpret and understand, in contextually constrained ways, the nature of the discursive activities they jointly enact and engage in with others. For my present purposes, I shall single out the following three categories, as proposed by Gumperz (Prevignano and di Luzio, this volume): 1. contextualization cues, seen as propositionless pure indexicals; 2. the inferencing of extra-communicative background information; 3. genre as an ideologically-charged metapragmatic assessment for referring to or labeling texts or speech exchanges.6 Each of these categories plays its role in explaining how the participants in some communicative event interpret and understand the unfolding interaction. Gumperz’s eclectic approach highlights two critically important issues. First, no single approach or tradition in the human and social sciences can on its own provide all of the analytical tools and theoretical categories that are required in order to answer the kinds of questions that Gumperz raises about discourse processes and social practices. Secondly, the attempt to integrate the insights from various approaches and traditions means that an act of reconstitution of their theoretical categories and analytical procedures is called for. This is necessary in order to produce a conceptually uniWed framework — one which can overcome the ideologically disabling disjunction between the “macro” and the “micro” approaches to the analysis of social phenomena (Thibault 1991:229–40). In the process of borrowing and adapting insights from various frameworks, these cannot simply remain as they were. Rather, they must themselves be changed in the process of their theoretical reconstitution. In my view, the above three categories refer, in actual fact, to three essential parameters in terms of which all acts of social meaning-making may be deWned. This is so irrespective of the speciWc semiotic modalities that are deployed. On this basis, I now propose three generalized meaning-making, or context-building, strategies, as follows: I. Indexical meaning-making practices; II. Intertextual meaning-making practices; III. Metadiscursive meaning-making practices.
Contextualization and social meaning-making practices
The three categories are analytically, though not constitutively, separable dimensions of all acts of social meaning-making. The fundamental question at stake here is: how do discourse participants integrate the diverse semiotic resource systems that are deployed into a single semiotic performance? Or, what are the contextualizing parameters in terms of which such integration is achieved? In order to provide a coherent theoretical response to these questions, we need to specify a number of very general meaning-making strategies that apply to all modalities of semiosis. I shall argue that the three parameters proposed above satisfy this requirement. Further, they do so in a conceptually uniWed way, as I shall now show.
I. Indexical meaning-making practices Indexicality, or “grounding” (Langacker 1987:126–8) refers to the fact that all semiotic forms and their functions in discourse contribute in some way to the ongoing enactment and maintenance of the wider situational context of which these forms are a constitutive part. From the instantial perspective of discourse, there is not, then, some special class of semiotic forms whose function it is to indicate the context. Instead, all the forms used in conjunction with each other function in various ways to enact and to specify the overall discourse event. From the point of view of discourse, the distinction between context-independent symbolic signs and context-dependent indexical ones is meaningless. All of the semiotic forms which are co-deployed in some communicative event serve a plurality of semiotic functions in that event. In so doing, they act on and aVect each other in complex and multiple ways. In my view, the distinction between propositional and propositionless signs is better handled by the distinction that Lemke (1984:69) makes between content-speciWc and type-speciWc semiotic acts (see also Thibault and Van Leeuwen 1996:567–8). A content-speciWc act in some way categorially construes or makes salient some speciWc features of the situation as being an instance of an event-type, act-type, etc. that is categorized by the ideational resources of the linguistic system and is, for this reason, recognizable in that community. If I say, for example, “The cat is drinking the milk,” with reference to a particular material happening, then I am not only indexing a speciWc cat in the situation through the combined use of speciWc deixis (‘the’) and the type-category speciWed by the noun ‘cat’. I am also construing, in my use of this clause, the perceived happening as an instantiation of a particular activity-
51
52
Language and interaction
schema, as modeled by the semantic conWguration [Actor^Process: Material Action^Goal] in the grammatical-semantic resources of the English clause. In so doing, I am making some perceived phenomenon in the purview of speaker and listener semiotically salient to the discourse event in a content-speciWc way. In indexing this phenomenon as an instantiation of a particular activitytype, I am constructing a contextually speciWc relation between the perceived event and the particular clause selection that I use. The clause-level semantics categorize or construe the material event as being the instantiation of a certain type-category of event which is modeled in the semantics of English clause grammar (Halliday 1994: chap. 5). A type-speciWc act, on the other hand, is an act which is in some way appropriate to the situation, irrespective of whether its ideational content, if it has any, indexes an actual state of aVairs or not (Lemke 1984:69). Consider the following, discussed in Levinson (this volume: 35): (1) A: Hey, how about supper together? B: I have a jealous husband. B’s response is content-speciWc if it is interpreted as referring to an actually jealous husband, at least from the point of view of B in the exchange. In this case, B provides a Reason-Explanation for declining A’s invitation. On the other hand, B’s response is type-speciWc if, for example, she uses it as a conventional strategy for forestalling the strategically calculated plan which she may interpret as motivating A’s invitation. In this case, B’s speech act is a conventional strategy which may be used to decline such invitations, and without giving oVence to the proposer, irrespective of whether her husband is actually jealous or not. B’s response to A may be interpreted as being either content-speciWc or type-speciWc or, perhaps, both. Whatever interpretation prevails is always a question of the ways in which the contextually constrained meaning potential of an utterance is taken up and negotiated in the unfolding discourse (Thibault and Van Leeuwen 1996:577–82). Semiotic acts, whether content- or type-speciWc, always contribute to the joint enactment of the relevant context. The particular examples I have focussed on may appear to exemplify the symbolic or propositional type of meaning discussed above by virtue of their exhibiting clause-level propositional content. In actual fact, these examples do not easily conform to the notion of context-free lexical and grammatical rules. I have tried to show
Contextualization and social meaning-making practices
that these, too, always function to index some aspect of the relevant contextual parameters whereby a particular communicative event is enacted. II. Intertextual meaning-making practices Gumperz refers to the extra-communicative “background information” which discourse participants draw on in order to make appropriate interpretative inferences. This raises two orders of problems. First, the notion of background information is not, analytically speaking, close enough to the order of discourse to be really useful. In my view, Gumperz is talking about a most important dimension of all social meaning-making, and one which is not speciWable at the level of, say, lexicogrammar per se in the linguistic system. Secondly, there is a danger that the notion of background information is reiWed as an abstract psychological property of the individual. In other words, an essentially social order of meaning relations is seen as simply already existing ‘in’ the minds of individual members of the culture in question. On both counts, the speciWc examples which Gumperz discusses belie these assumptions. With this in mind, I shall now propose an alternative to the notion of background information — one which puts social meaning-making in the center of the picture. The central question to tackle is: how do discourse participants construe meaningful relations between one occasion of discourse and another? Or, how does one occasion of discourse serve as the context for the interpretation of another? Following the work of Jay Lemke (1985), I shall discuss these questions in terms of the intertextual meaning-making practices (IMMP’s) of some community (see also Thibault 1986, 1990). Two of the examples that Gumperz himself discusses — viz. the announcer on the classical music radio station and the rape trial transcript — provide a good starting point. Gumperz points out how, in both cases, he had to “search [his] memory of past communicative experience to construct a likely scenario or narrative plot that might suggest possible interpretations” (Prevignano and di Luzio, this volume: 13). In both cases, Gumperz’s own observations show how the speciWc meanings he assigns to these instances are construed in and through their relations to still wider, more abstract, typical patterns of meaning relations which the members of a particular community may share, access and draw on to varying degrees and in varying ways. Thus, the contrastive accent on ‘you’ in the radio announcer’s speech is
53
54
Language and interaction
assigned its speciWc meaning in relation to expectations concerning the typical patterns of relations that are construed between intonation and lexicogrammar, and context-type. In this example, Gumperz draws our attention to local features of the interactional context. It is the lack of expectational ‘Wt’ between the two semiotic modalities which provides the basis for the speciWc meaning that Gumperz assigns to the event. This example shows how often very subtle co-patternings of selections across semiotic modalities may be crucial to the meaning of the discourse event, or to some local part of this. In his second example — the rape trial transcript — Gumperz starts from assumptions concerning the type of social activity which is taking place. In contrast to the Wrst example, the perspective he adopts on this occasion is a global one. SpeciWcally, he interprets the speech event through the genre of “cross-examination as adversarial proceeding.” Again, Gumperz draws our attention to the ways in which discourse participants relate the speciWc meaning selections that are made — viz. the speech act strategies of attorney and victim — to their expectations about the particular discourse genre through which they interpret the activity which is taking place. It does not follow, of course, that there is always consensus about this, as this example shows. Both of the perspectives — local and global — illustrated in these two examples draw attention to a second most fundamental dimension of social meaning-making, viz. the IMMP’s of some community. Intertextuality does not simply refer, as in much of literary theory and criticism, to the relations between a particular text and some other, usually historical prior text (Thibault 1991:135). The fact that it may do so is not at issue here. Literary citation, allusion, paraphrase, and so on, are all intertextual in this more restricted sense. More fundamentally, it concerns the multiple ways in which a community’s meaning-making practices enable discourse participants to link two or more texts or occasions of discourse on the basis of some more abstract pattern of meaning relations that they are seen as having in common. Intertextual meaning relations are, then, abstract patterns of relations that discourse participants draw on both to construct and to interpret speciWc instances. Thus, no text or occasion of discourse is ever understood in isolation from the wider systems of intertextual meaning relations in the community (Lemke 1985:275). Rather, all discourse occasions are made and interpreted in and through wider systems of IMMP’s. This does not mean that these are shared, accessed, or interpreted in the same way by all members of the community. Bakhtin’s (1981) notion of “social heteroglossia” represents
Contextualization and social meaning-making practices
an important early attempt to explore the diversity of IMMP’s, and the relations of alliance, conXict, co-optation, and so on among these in particular texts and occasions of discourse (Lemke 1988b; Thibault 1989). As Gumperz’s examples show, IMMP’s entail more abstract, or higherorder, relations of relations of relations ... in Bateson’s sense. They are neither systems of ready-made text-types nor recipes showing how to construct particular types of texts. Rather, they provide an interpretative resource for construing culturally salient patterns of meaning in speciWc instances. In neither case is it a simple question of whether the speciWc instance matches the intertextual pattern or not. Rather, discourse speciWc meanings are interpreted on the basis of whether and to what extent they conform to or contrast with a given intertextual pattern. IMMP’s are neither homogeneous nor consistent across the diverse practices that constitute a given community. That is why they are said to be heteroglossic in Bakhtin’s sense. The notions of “background information” and “inferential procedure” risk both de-semioticizing what are in fact the systems of IMMP’s in a community and also reifying their essentially social nature as if these were properties of individual mind. Neither of these notions adequately speciWes the local (instantial) resources whereby discourse participants reconstruct the relevant IMMP’s as they deploy these in their own meaning-making practices. The notion of IMMP’s better enables us to reconstruct in our analyses the diverse ways in which members indicate which intertextual relations are relevant, and how these relate to which other occasions of discourse (Lemke 1985:276). Neither “background information” nor “inferential procedure” is adequately grounded in a theory of the semiotic forms and their modes of deployment in discourse to bring about the conceptually uniWed account — one which can reconstruct the links between resource systems and social meaning-making practices — to make this possible. The idea of a ‘background’ certainly has the merit of suggesting the diVuse, taken-for-granted, and outof-focus assumptions that characterize the lifeworld or the habitus of values, practices and orientations that inform the speciWc discursive event (see also Levinson, this volume). As it stands, however, I remain unconvinced that it is able to provide us with the means for analytically reconstructing the links between the “macro” and the “micro” perspectives in a conceptually uniWed way (see Thibault 1991:229–40 for further discussion).
55
56
Language and interaction
III. Metadiscursive meaning-making practices Gumperz also refers to Bakhtin’s notion of speech genres. This is the metadiscursive dimension of social meaning-making. For Gumperz, genre is “an ideologically charged metapragmatic assessment”, to echo a Silversteinian turn of phrase, or “forms of talk about talk”, in the GoVmanesque parlance. From this point of view, discourse participants make interpretative assumptions about the meaning of their own and others’ contributions to the ongoing discourse event. They self-reXexively interpret to varying degrees of explicitness and awareness the meaning of particular discourse events or parts of these, relative to some social viewpoint. To do so, members deploy the meaning-making resources of the system so as to assign signiWcance to their own and others’ local instantiations of these same resources. Social meaningmaking practices may be said to be metadiscursive in this sense. Whereas IMMP’s connect the patterns in the speciWc instance to still wider patterns of meaning in a given culture, metadiscursive meaning-making practices (MMMP’s) serve to articulate the links between the local, or occasion-speciWc, meaning selections in some speciWc communicative event to the paradigmatically-organized contrast sets of context-types in relation to which the former have their meaning. That is, MMMP’s link the speciWc instance or meaning choice to the metasystem of possible alternative choices from which it was selected. This may be so to varying degrees of completeness and explicitness. MMMP’s bring into focus the meaning of the particular act by providing a connection between this and the higher-order metasystem of possibilities which gave rise to it. However, it is not enough simply to say that metapragmatic assessments involve “talk about talk”. There are many diVerent ways in which members talk about talk. As analysts, we need to be able to specify which forms of talk about talk occur, and in which context-types. What are the speciWcally ‘meta’ relations which specify the relations between the particular act, the context in which it occurs, the always social viewpoints of the participants in the discourse, and the metasystem — the contrast sets — which any occasion of metadiscourse always invokes? Whenever we engage in talk about talk, we always invoke or reconstruct, to varying degrees of completeness, the metasystem in terms of which the given act has its meaning. MMMP’s make explicit in some way the fact that the metasystem is always immanent in speciWc occasions of discourse. Thus, MMMP’s interpret or gloss the meaning
Contextualization and social meaning-making practices
of the instance from the global standpoint of the metasystem. In making the link between local occasion and system, participants show how they have their own socially-constrained metatheories — folk or otherwise — concerning the ways in which occasion-speciWc acts of meaning are relatable to the global system of possibilities which engendered them.
5. Conclusion: Towards a functional semantics of action The above three-way distinction refers to diVerent aspects of the constant dialectic between local (instantial) and global (systemic) relations in discourse. These may be summarized as follows: I. Indexicality is concerned with the ways in which selections from the global resource systems of the various semiotic modalities cross-classify and integrate both with each other and with selected features of the material-phenomenal world so as to enact a local, occasion-speciWc, discourse event; II. Intertextuality speciWes the global systems of intertextual meaning relations that are seen as relevant to the interpretation of a given local occasion of discourse; III. Metadiscursivity is concerned with the ways in which the global system of possible sets of alternatives — the metasystem — is made locally salient in some interaction. It does so by providing the means by which participants construct a self-reXexive connection between the metasystem and some discourse event. This means that discourse participants draw on this system in the process of assigning meaning to some discourse event. They do so by explicitly focussing the metasystem onto the immediate interpretative problem. In this way, the relevance of the global possibilities of the metasystem is made an explicit object of metasemiotic consciousness, relative to some participant viewpoint. The notion of contextualization is a fundamental one in Gumperz’s theory of discourse. Gumperz has shown us how the ongoing interactional event is constructed and construed in and through the ways in which the localized choices — utterances, speech acts, and so on — of the interactants relate to
57
58
Language and interaction
each other in the real-time unfolding of the discourse. This depends on the ways in which interactants integrate all of the diverse contextualization cues into their sense of the overall activity which is taking place (Gumperz 1992:44). Rather than trying to interpret the other semiotic modalities from the exclusive standpoint of linguistic analysis, the task should be to develop ways of grounding the multimodal nature of such events in a functional semantic approach to action as a whole. The three-way distinction I made above attempts to provide a suYciently generalized and integrated framework for analyzing the local and the global dimensions of social meaning-making. Gumperz’s work on contextualization makes a seminal contribution to the development of a semiotics of action in which the functional integration of all of the semiotic modalities which are involved is accounted for.
Notes 1. Thus, Wittgenstein in his Tractatus: “3.203 A name means an object. The object is its meaning” (1969 [1921]:23). 2. See Kenneth Pike’s (1967) wave and Weld, as distinct from his particle, perspectives on language, prosodies of all kinds, suprasegmentals, and so on. 3. It may well turn out, as further research is conducted on this question, that what we conventionally refer to as ‘language’ and treat as a uniWed semiotic system will turn out to be a diverse set of possibilities whereby linguistic forms are integrated with other semiotic modalities. In the conventional view, speech and writing have been seen as diVerent material realizations of the same lexicogrammatical forms in the oral-aural and opticalvisual channels. In my view, speech and writing are themselves folk-linguistic glosses on a diverse set of multimodal meaning-making practices in which diVerent possible integrations of lexicogrammatical, prosodic, paralinguistic, graphological, pictorial, kinesic, spatial, rhythmic, and other resources characterize the discourse genres — the semiotic action formations — of a given community. In such a view, genre will, increasingly, be deWned as the typical modes of deployment of the multimodal meaning-making resources available to the members of a given community. 4. Alan Rumsey (1990) points out how, in the Ungarinyin language of Australia, the meanings of linguistic locutions are glossed, not on the basis of de-contextualized criteria of reference, but, instead, on ‘pragmatic’ criteria concerning the ways in which the locution, or some alternative, locally construed as equivalent to it, is used in the various context-types of the culture. 5. Bolinger, for instance, recognized long ago that entire intonational conWgurations may be related to speciWc context-types in culturally speciWc ways (1951:209). Further, and just like lexicogrammatical forms, these are “ineVable” in the sense that they cannot be reliably glossed in terms of some extra-semiotic absolute standard (Halliday 1988). Rather, the
Contextualization and social meaning-making practices
meaning of a particular intonation selection itself derives in the Wrst instance from the place which any given selection is felt to occupy in an overall system of contrasting options which deWnes its semiotic value. 6. In each case, Gumperz acknowledges the diverse sources which have inXuenced his thinking on these issues. In relation to (1) above, he singles out Peirce’s notion of indexical signs and GoVman’s frames; he relates (2) to GarWnkel’s inferential processes; and (3) derives in part from Bakhtin’s theory of speech genres, as well as Silverstein’s (1992:60) “metapragmatics”.
References Bakhtin, Mikhail. 1981. “Discourse in the novel”. In The Dialogic Imagination: Four Essays by M.M. Bakhtin, C. Emerson and M. Holquist (trans.), M. Holquist (ed.), 259–422. Austin and London: University of Texas Press. Bakhtin, Mikhail. 1986. “The problem of speech genres”. In Speech Genres and Other Late Essays, V.W. McGee (trans.), C. Emerson and M. Holquist (eds.), 60–102. Austin: University of Texas Press. Bateson, Gregory. 1973. “The logical categories of learning and communication”. In Steps to an Ecology of Mind, 250–279. London and New York: Paladin. Bolinger, Dwight L. 1951. “Intonation: Levels versus conWgurations”. Word 7 (3):199– 210. Di Luzio, Aldo. “Presenting John J. Gumperz”. This volume. Gumperz, John J. 1992. “Contextualization revisited”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 39–53. Amsterdam and Philadelphia: John Benjamins. Halliday, M.A.K. 1978. Language as Social Semiotic. The Social Interpretation of Language and Meaning. London: Arnold. Halliday, M.A.K. 1988. “On the ineVability of grammatical categories”. In Linguistics in a Systemic Perspective, J.D. Benson, M.J. Cummings, W.S. Greaves (eds.), 27–51. Amsterdam and Philadelphia: John Benjamins. Halliday, M.A.K. 1994. An Introduction to Functional Grammar. Second edition. London and Melbourne: Arnold. Langacker, Ronald W. 1987. Foundations of Cognitive Grammar, Vol.1. Theoretical Prerequisites. Stanford, CA: Stanford University Press. Lemke, Jay. 1984. “Action, context, and meaning”. In Semiotics and Education. Victoria University, Toronto: Monographs, Working Papers and Prepublications of the Toronto Semiotic Circle 2:63–93. Lemke, Jay. 1985. “Ideology, intertextuality, and the notion of register”. In Systemic Perspectives on Discourse, Vol. 1. Selected Papers from the 9th International Systemic Workshop, J.D. Benson and W.S. Greaves (eds.), 275–294. Norwood, NJ: Ablex. Lemke, Jay. 1988a. “Genres, semantics, and classroom education”. Linguistics and Education 1 (1):81–99.
59
60
Language and interaction
Lemke, Jay. 1988b. “Discourses in conXict: Heteroglossia and text semantics”. In Systemic Functional Approaches to Discourse. Selected Papers from the 12th International Systemic Workshop, J.D. Benson and W.S. Greaves (eds.), 29–50. Norwood, NJ: Ablex. Lemke, Jay. 1990. Talking Science. Language, Learning, and Values. Norwood, NJ: Ablex. Lemke, Jay. 1993. “Discourse, dynamics, and social change”. In Language as Cultural Dynamic. Cultural Dynamics VI (1–2):243–275. Lemke, Jay. 1998. “Multiplying meaning: Visual and verbal semiotics in scientiWc texts”. In Reading Science: Critical and Functional Perspectives on Discourses of Science, J.R. Martin and R. Veel (eds.), 87–113. London and New York: Routledge. Levinson, Stephen C. 1997. “Contextualizing ‘contextualization cues’”. This volume. Martin, James R. 1992. English Text. System and Structure. Amsterdam and Philadelphia: John Benjamins. Nöth, Winfried. 1990. Handbook of Semiotics. Bloomington and Indianapolis: Indiana University Press. Pike, Kenneth L. 1967. Language in Relation to a UniWed Theory of the Structure of Human Behavior. Second, revised edition. The Hague and Paris: Mouton. Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This volume. Rumsey, Alan. 1990. “Wording, meaning, and linguistic ideology”. American Anthropologist 92 (2):346–361. Silverstein, Michael. 1987. “The three faces of ‘function’: Preliminaries to a psychology of language”. In Social and Functional Approaches to Language and Thought, 17–38. Orlando and London: Academic Press. Silverstein, Michael. 1992. “The indeterminacy of contextualization: When is enough enough?”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 55– 76. Amsterdam and Philadelphia: John Benjamins. Thibault, Paul J. 1986. “Thematic system analysis and the construction of knowledge and belief in discourse”. In Text, Discourse, and Context: A Social Semiotic Approach. Victoria University, Toronto: Monographs, Working Papers and Prepublications of the Toronto Semiotic Circle 3:44–91. Thibault, Paul J. 1989. “Semantic variation, social heteroglossia, intertextuality: Thematic and axiological meaning in spoken discourse”. Critical Studies 1 (2):181–209. Thibault, Paul J. 1990. “Questions of genre and intertextuality in some Australian television advertisements”. In The Televised Text, R. Rossini Favretti (ed.), 89–131. Bologna: Pàtron. Thibault, Paul J. 1991. Social Semiotics as Praxis. Text, Social Meaning Making and Nabokov’s ‘Ada’ [Theory and History of Literature Series 74]. Minneapolis and Oxford: University of Minnesota Press. Thibault, Paul J. 1994. “Text and/or context?”. In The Semiotic Review of Books (Toronto) 5 (2):10–12. Thibault, Paul J. 1997. Re-reading Saussure. The Dynamics of Signs in Social Life. London and New York: Routledge.
Contextualization and social meaning-making practices
Thibault, Paul J. and Van Leeuwen, Theo. 1996. “Grammar, society, and the speech act: Renewing the connections”. Journal of Pragmatics 25:561–585. Van Leeuwen, Theo. 1992. “The schoolbook as a multimodal text”. Internationale Schulbuchforschung 14:35–58. Wittgenstein, Ludwig. 1969 [1921]. Tractatus Logico-Philosophicus / Logisch-Philosophische Abhandlung. D.F. Pears and B.F. McGuinness (trans.). London: Routledge and Kegan Paul.
61
62
Language and interaction
Gumperz and the minims of interaction
Chapter 5
On Gumperz and the minims of interaction1 Carlo L. Prevignano
1. Introduction The aim of my contribution is to focus on what I call the old problem of human interaction (“How is human interaction possible?”), and on what I propose to deWne as the minims of interaction, as distinct from the maxims. I see Gumperz as a minimist, as compared to Grice, Leech et al., the maximists. I argue that everyday human solutions to the interaction problem can be explained by invoking (specifying and analyzing), not only a set of maxims, but Wrst of all a set of minims, corresponding to minimal tactical requirements for the application of interaction-types and activity-types. Here I conWne myself to considering only some of these minims, in particular the ones I connect to “activity-type cues”. I shall call these activity-type minims. They work as interaction minims as well. One of them says: “Signal what you are doing,” or “Signal the activity-type you are performing.” Other activity-type minims, which function also as interaction minims, say: “Signal the beginning of your performance according to an activity-type,” “Signal its end,” and “Signal and conWrm your activity-type involvement.” Minims govern what Duranti (1985:206V; 1992:48) has called “boundary markers”, that is, conventional cues marking the beginning and closing of an interaction or an activity-type application, or starting and concluding them tout court. Protocols for human interaction, as well as for activity-types, include both minims and maxims, but it is the application of maxims which is constrained by the application of minims, and not vice versa (minims come Wrst). The execution of minims by human agents gives cues, or clusters of cues, as output, which in turn are processed as input by other persons, according to input minims. In the latter case, minims are used as control and monitoring
63
64
Language and interaction
goals (see the goal approach in Castelfranchi 1994), or as plan constituents (see the plan approach in Prevignano 1993). In other words, if the output minim, “Signal what you are doing”, has, as its corresponding input minim, something like “Look for what the other is doing,” or “Look for the other’s activity-type cues,” it is under this input minim (goal or plan) that some cues as input are tentatively ascribed to an activity-type, thereby triggering the connected protocol and the included constraints on input maxim application, as well as activity-speciWc inferences (à la Levinson 1979). If we agree that activity-type cues “are usually produced and interpreted without conscious reXection” (to use the words Gumperz has adopted for “contextualization cues”; see Prevignano and di Luzio, this volume: 14), then we have to assume that applications also of minims usually occur without conscious reXection. Usually, but not always. To Grosjean’s remark that “la non-survenue d’un indice attendu ou sa survenue dans une position non habituelle vont être interprétées” (2001:147), it is therefore possible to add that the absence of the cues corresponding to an expected activity type may indicate an intentional non-application of an output minim or an outputminim violation (e.g. the intentional violation of the “signal the activity-type you are performing” minim), and what is interpreted as an intentional outputminim violation may induce implicatures. As to minims, here I shall conWne myself to indicating a list of open questions. Do they correspond to interactional universals in human cultures? Do the minims of interaction include more than the activity-type minims? Are there speciWc minims corresponding to cultural and historical interaction-types? What about the relationship between interaction minims and participation minims? Regarding human act/action/activity types, while Gumperz has considered and studied the verbal ones only (what I call the “Gumperz level” in section 3), I do not exclude, in my present considerations, any possibility, from entirely and predominantly non-verbal, to predominantly and even exclusively verbal types (see Table 1). Interestingly, for Gumperz, activities or activity types now correspond to participants’ or conversationalists’ “cognitive constructs” or “activity constructs” (Prevignano and di Luzio, this volume: 14), which do not apply to interactions as whole, but to individual moves, and are characterized as being “subject to constant change in the course of the exchange” (Prevignano and di Luzio, this volume: 13). If it is possible to assume that there is usually a constant change in participants’
Gumperz and the minims of interaction
application of verbal activity types during an exchange (see also Gumperz 1979:274, where everyday conversation is presented as being “marked by constant transitions from one mode of speaking [or type of verbal activity, CP] to another; shifts from informal chat to serious discussion, from argument to humor, or narrative to rapid repartee, etc.”), then activity constructs seem rather to be considered today by Gumperz as participants’ models for producing and interpreting moves. I think it is useful to distinguish between participants’ constructs and analysts’ representational and explanatory constructs for those of participants, as well as between act/action/activity tokens and types. At this point, I propose another distinction between folk or cultural typologies and tokenologies, on one side, and analysts’ or culturological ones, including praxeological, action-theoretical and activity-theoretical, of all kinds and schools, on the other. A further remark. To be explicit, here I assume not only a 3-, 5-, or 10-level theory of (joint, we- or collective) activity and interaction, but an n-level one, and consequently a corresponding n-level analysis.
2. Some preliminaries After the diVerent interpretations (objectivist, subjectivist, experientialist, constructivist, dialogic or interactionist, inferential or heuristic, and even skeptical, see Berge 1994) of communication, communication about communication, communication about and for action and interaction, perhaps it is now the turn of the analysts — the communication and interaction analysts — to be analyzed. What type of analyst does Gumperz correspond to? Pseudo-omniscient? Imperfect? Diabolical? I agree with Levinson (this volume:38): “cheery optimist”. But what is left over from a cheery optimist’s analytical game and models of communication and interaction, after the assault of “radical doubt” (Levinson, this volume:38)? A “postmodernist conclusion” (again in Levinson’s words), a skeptical idea of both communication and interaction, of the analytical game itself? I will resist the temptation of arguing that a skeptical analyst is not self-defeating and what he or she needs is, precisely, Gumperzian optimism and openness for nourishment. I will take that for granted. But what about a skeptical meta-analyst? Here I try to play that role.
65
66
Language and interaction
If communication and interaction analysts, as human believers (for artiWcial ones see Ballim and Wilks 1991), are today in a postmodern condition — la condition postmoderne is not yet posthuman! — and their rejection of grands récits, big theories and macromodels, seems to be a must, their analysis reports are still récits, i.e. discourses, narrated argumentations. And how can we reconstruct and assess analysts’ argumentations, their research goals and procedures? How can we reconstruct their itineraries, starting from their selfimages, pre-theoretical assumptions, questions and doubts, and proceeding to prediction-making and dis/conWrming, to elicitation and re-elicitation procedures? How can we accomplish all this, if not via their reports? The metaanalyst has to move between terra incognita and fable convenue (as Koerner 1973 called them, and cf. Prevignano 1979:23–26) in order to reconstruct analysts’ programs and results. The Galilean Eppur si muove (“And yet it moves”) here becomes Eppur si comunica (“And yet we communicate”). However skeptical, a communication analyst cannot reject this assumption, on pain of self-dismissal. The story is the same for the meta-analyst. The methodological solipsism of both (together with Ballim and Wilks’ cognitive solipsism) cannot but be considered methodological, that is, an antidote or a means of defence against reiWcations. And yet we communicate in communication research as well. Meanwhile, Babel has turned electronic, and along with the preWx e-, not only do we have ecommunication (as well as e-orality and e-literacy), but also e-miscommunication and e-misunderstanding. Inevitable misunderstanding (Blum-Kulka and Weizman 1988)? Looking for the locus, or the loci, of (pre-electronic) misunderstanding and misinteraction during intercultural encounters, the Wsh turned ichthyologist and reached contextualization cues, as well as interactional meanings (i.e. micro), after he had studied “social meanings” (i.e. macro), and their not merely “segmental” attachments (see Blom and Gumperz 1972:417–18). Some of the contextualization cues could now be redeWned as “scripticals” in relation to activity protocols or scripts. Here I use protocols and scripts interchangeably. They correspond to both representational and procedural constructs of conversationalists, who need models of their own activity and interaction types and tokens, as well as the respective know-how. For their part, human analysts need sets of assumptions, constraints and procedures as constructs for representing not only cultural and historical types and tokens of human activities (from largely unscripted to rigidly scripted ones), but also
Gumperz and the minims of interaction
the activity models and the know-how which are used by participants themselves to categorize and monitor “what’s going on” and to act in life. Following Gumperz, Svennevig (1999:11–12) prefers to call the latter “activity frames”.
3. The Gumperz level As an analyst of the “fabric of our social life” (1982a:7), sub specie communicationis, Gumperz has contributed to the identiWcation of an interaction principle which I intend to present here in three diVerent speciWcations, each on a diVerent scale (see level i. to level iii. in Table 1). These range from a general action-theoretical or praxeological interpretation to a linguistic one. A caveat. Gumperz has contributed to the individuation of that principle in the sense that there is a complicity in the crime (and his accomplices include, at the very least, Bateson, GarWnkel, GoVman, Sacks, Erickson and Levinson — an invisible college à la Winkin (1981/1995:21), now quite visible; see Kendon 1990 and Auer 1992:22–23).2 However, it is Gumperz who has provided its most extensive analyses at the linguistic levels of speciWcation, in particular studying what he deWnes as “contextualization cues”. Here is the principle, which I propose to call the “semiotic principle of interaction”: human agents do not simply act, but are also accustomed to signaling or cueing the types of act/action/activity they are engaging in; in other words, they make them interpretable at least via what could be called “Gumperzian cues”, rather than (only) with overtly lexicalized or verbalized explanations and rationalizations. Among other reasons, they do so in order to economically facilitate coordination, to simulate, and to lower or raise the anxiety level of the other agents present. “Uncertainty reduction” (Giles et al. 1979:372; Berger and Bradac 1982) is accompanied by certainty induction, and certainty reduction by uncertainty induction. Table 1 presents three speciWcations of the semiotic principle of interaction. I cannot resist the temptation of calling the Wrst level of speciWcation of the principle, the “GarWnkel-and-Sacks level”, the second, the “Levinson level” (see Levinson 1979), and the third, the “Gumperz level”.3 While presenting “interpretive sociolinguistics”, Auer and di Luzio have anticipated a deWnition of what I call the Gumperz level, since they have referred to “the participants’ continuous task and accomplishment to make interpretable and interpret each other’s linguistic activities” also via cues (1984:viii). A distinct, but
67
68
Language and interaction
Table 1. Three speciWcations of the semiotic principle of human interaction. In order to be made intelligible/interpretable/recognizable: i. all human ii. entirely and predominantly nonverbal as well as predominantly verbal iii. even exclusively verbal
act/action/activity types are signaled at least by cues
parallel, set of tasks (i.e., signaling and identifying or detecting signaled pragmatic functions of diVerent linguistic activities or activity-types) has been considered by Stati (1982, 1990) and others from the perspective of dialogue analysis, and in relation to dialogue constituents other than contextualization cues. In my view, the Gumperz level corresponds to both cueing itself (as the process of proposing the use of x as a cue for y as a cued entity, i. e. of a cueing function and its functors; see Kendon 1996) and cue processing (emission and reception of both coded and still uncoded or unshared cues), with a distinction (which has already been used by Richard Petty and John Cacioppo in their research on persuasion) between central and peripheral processing.
4. A cue/cueing model and analysis Gumperz has adopted GarWnkel’s program for studying human activities “through language” (1982a:172), i.e. via verbal signals, and in particular via what he has termed “cues” (entre autres, in his experiments at Berkeley with Frederick Erickson et al. on interactional synchrony/asynchrony, see Gumperz 1979:278), rather than “markers” (see the markers of Brown and Fraser 1979, as well as of Giles et al. 1979, and the latter’s in nuce “marker [and marking] theory”). Gumperz begins by working back from participants’ and other members’ assessments to the surface of the verbal interaction (cf. the notion of “surface” in Orletti 1994). He has managed to combine a top-down approach with a bottom-up one, studying both surface constituents and highlevel inferencing, as well as their couplings. In this way, he has provided in advance a solution to the debate on the predominance/priority of surface structure vs. deep structure in approaches to human verbal interaction (see Castelfranchi 1994; Orletti 1994). Yet Gumperz’s surface is populated by more than just triggers of one-to-one relations (as studied in SchiVrin 1987 or
Gumperz and the minims of interaction
Bazzanella 1990, 1994; see also Dascalu Jinga and Vanelli 1996), if, importantly, Gumperzian signals also include clusters of co-occurring cues. As both output and input, the latter require strategies, not simply processing rules. In order to capture and represent the many-to-many semiotic circuitry which Gumperz has characterized as “contextualization” (both in output and in input), Peircian semiotics is not enough, though this question cannot be explored here. Gumperzian cues can be coded and uncoded, shared and unshared, but also plural, as multimodal clusters of co-occurrent triggers. Further, there is another research problem, which is also pointed out by Bateson’s metaphor of a shower of signals (1971/1981:125; for Bateson’s cues and their functions, cf. Rawlins 1987), when these signals are considered as an output from and an input for human agents. It is the problem of possibility conditions for their human processing, and in general for the human processing of semiotic multimodality (cf. Thibault, this volume). If from a “structureless barrage-of-signals view” of human interaction (i.e. an idea of interaction as an “unstructured exchange” of signals, again in the words of Brown and Levinson (1979:293)), via a “cross-shower of signals” view of it (à la Bateson), we come to a structural approach to interaction as a multichannel event, it is important to avoid studying it only as a turn structure. An antidote against today’s triumphant reduction of human interaction to a turn structure and its variants (from a turn-by-turn sequence to a bundle of n tracks, for n participants, i.e. a multitrack space with tracks being multichannel and turns occurring on each of the multichannel tracks; I propose this variant in Prevignano and Thibault, 2002:166), an antidote, I say, can be an nlevel theory and analysis of interaction, as well as a corresponding n-level processing model. In this direction, one of Gumperz’s contributions has been to analyze occasions of human interaction on a move level, rather than on a turn level, seeing “actors engaged in strategically formulating and positioning their moves,” as well as interpreting and reinterpreting other participants’ moves. Gumperz has distinguished “moves”, “countermoves”, “conWrmations”, and “disconWrmations”, i.e. move sequences (see Prevignano and di Luzio, this volume:12). What could be called “move processing” is analyzed by Gumperz as conversationalists’ execution of “rhetorical strategies” and interpretive procedures, with positional (etc.) constraints on their applications. Among interactional abilities, it is “rhetorical ability”, as the command or control of a set of rhetorical strategies being developed in networks of interactions, that
69
70
Language and interaction
Gumperz (1990b:42) has studied at the move level. A multilevel analysis of a dissertation defence has been provided by Gumperz (also in collaboration with his wife, Jenny Cook-Gumperz, and in relation with the so-called Multiple Analysis Project) in a number of studies (Gumperz 1989b, Cook-Gumperz and Gumperz 1994, 1996), where an idea of conversational politics (or, as it is called, “politics of conversation”) is presented and used to capture a set of strategies and decisions governing the entire dynamics or history of an interaction, and in particular “what can and cannot be said, and […] how and in what tone information is to be presented” (1996:171). A further remark. Interestingly, another multilevel approach to talk-ininteraction is proposed by Catherine Kerbrat-Orecchioni (1997), whose linguistique interactionniste or interactionnelle, starting from the Gumperzian motto, “speaking is interacting” (Gumperz 1982a:29), reaches entre autres a sparkling and realistic idea of interaction as “bricolage interactif incessant” (1998:61), while in interaktionale Linguistik (Selting and Couper-Kuhlen 2001), one of the Gumperzian foci is on conversational or interactional prosody, e.g. the rhythm and tempo of spoken interaction (Auer, CouperKuhlen and Müller 1999). Cued entities à la Gumperz — activity-types in this case — require a “large dose of inferential reconstruction” (Levinson, this volume:37) by both the participants and the analysts. Moreover, Gumperz-the-analyst’s methodology is hermeneutical and reconstructive: “The analyst’s interpretation is a reconstruction of the participants’ interpretation. The assessment of the analyst, who has to reconstruct, that is, to know and understand, the ethnographic context [in our case, activity-types, CP] as if he were a participant member, is a reconstruction of the participants’ assessments, and may possibly be conWrmed by them” (di Luzio, this volume: 6, and cf. Ensink 1987). In this way, a “singsong rhythm” (Gumperz 1974:11; 1982a:34), that is, the rhythmic mode of intonation (Gumperz 1995b) of the well-known, (1) Ahma git me a gig [Gumperz’s rough gloss: I’m going to get myself some support]
was discovered by Gumperz, working back to the utterance surface from some participants’ and other members’ assessments, to be a cue accompanying the utterance which was addressed by a black student to other black students. Gumperz was able to consult four groups of “judges” (1982a:31–32). According to the members of the last group, it is the rhythm which gives the whole
Gumperz and the minims of interaction
utterance “a formulaic character” (1982a:34). Elsewhere, Gumperz considered (1) as an instance of “formulaic phrases identiWable through co-occurrent selections of phonological, prosodic, morphological and lexical options” (1982a:133). In any case, because of its formulaic character, (1) was interpreted as corresponding to (what I call) an “activity-type minim”, if it was glossed as (a) below, to use Gumperz’s own words (cf. Prevignano and di Luzio, this volume: 20): (a) “I am doing something that minority people like myself have to do, get support where I can.”
Other glosses, quoted by Gumperz, were: (b) “I’m still in control” (Gumperz 1982a:31); (c) “I am just playing the game as we blacks must do if we are to get along in a white dominated world” (Gumperz 1982a:31–32).
According to (a), (b), and (c), both the output minim (“Signal via cues what you are doing”) and the corresponding input minim (“Look for what the other person is doing, i.e. for his or her activity-type cues”), seem to have been applied by the student and his addressees, respectively. But can the results of a reconstructive methodology be considered as a post-factum ‘explanation’ of cases such as (1)? Postdictions (vs. predictions) presuppose the resolution of postdictability and postdeterminability problems (Prevignano 1991, 1993).
5. The past and the present There is an extraordinary game which Gumperz has been playing for more than forty years: integration. Its risk is eclecticism, and Gumperz knows that he takes “a somewhat eclectic approach” (1996a:359) to interaction via languages. From when he was Wfty (in 1972), he has subsequently presented his research as speech event analysis (Gumperz 1972:17–18; cf. Gumperz 1984:281), conversational analysis of social meaning and sociolinguistic analysis of conversations (Gumperz and Herasimchuk 1973), sociolinguistics of interpersonal communication (Gumperz 1974), the interpretive approach to conversation (Gumperz 1982a:6, 35–7; for a presentation of interpretive sociolinguistics, as Auer called it, see Auer and di Luzio 1984), and interactional sociolinguistic analysis (Gumperz 1982a:210) or interactional socio-
71
72
Language and interaction
linguistics (Blom and Gumperz 1972:432; Gumperz 1982a was published as the Wrst volume of the series Studies in Interactional Sociolinguistics; see also SchiVrin 1994: ch. 4). Gumperz’s trajectory has cut across the ethnography of speaking, GarWnkel’s ethnomethodology (Gumperz and Hymes 1972a), Sacks and SchegloV’s analyses (Gumperz and Hymes 1972b and 1972c, respectively), Grice’s cooperation principle (Gumperz 1990a), GoVman’s interaction order and Levinson’s analysis of activity-types (1979), but not his idea of interactional intelligence (1995), and may now be seen also as a foundational contribution to an inferential model of human interaction and to what, with Grundy (1995:123), we can call an “activity-type theory”, with Thomas (1995:293), an “activity-type approach”, and with Sarangi (2000:6), an “activity type analysis” with an “integrated model.” I see the peculiarity of Gumperzian socio-cognitivism as consisting in the application of an interpretive approach, to use Gumperz’s own words, to “the analysis of real-time processes in face-to-face encounters” (Gumperz 1982a:vii). More than thirty years after its presentation, this research program continues to be actively pursued and developed (see also Gumperz 1993, 1995a, 1999, 2001a, 2001b, and Gumperz and Berenz 1993).
6. The future If to activity-types there correspond activity scripts, describable as “joint protocols”, i.e. socio-procedural entities, the already rich toolbox of an approach to human interaction should include a scriptionary, i.e. a repertoire of scripts, as both a procedural and representational construct completely diVerent from a script-based lexicon à la Raskin 1981 (Prevignano 1994, 1995). A scriptionary would be useful for the members of the Turing club, as a procedural way of dealing with cultural sets of human activity-types, but also for participation analysts (Goodwin and Goodwin 1992; Duranti 1997:280–330), as a representational way of capturing the “participatory quality” (Duranti 1997:46) of human activities, or what are classiWed as “participatory actions” (Clark 1999:149). From output and input minims to scriptical cues, from participants’ strategies to those of analysts, a research agenda for studying human intelligent interactions and interactional intelligence (see Levinson 1995 and Berger 1997, but also Justine Cassell’s implementations) cannot forget Gumperz’s inspiring contributions. And here ends my piccolo omaggio to a grande maestro.
Gumperz and the minims of interaction
Notes 1. Revised and augmented version of a text previously published as Prevignano 1997. Paul Thibault’s comments were valuable in the preparation of this version. 2. See also Murray (1994) for a group-sensitive Gumperzologist’s report. 3. See Gumperz 1974:18V, 1977:205V, 1980:119, on a distinction between speech acts and speech activities; Gumperz 1982b:329V, and Gumperz and Cook-Gumperz 1982:11V, on a distinction between activities and tasks; see also Gumperz 1989a:78–79, 1990a:439, 1992:43V.
References Auer, Peter. 1992. “Introduction: John Gumperz’ approach to contextualization”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 1–37. Amsterdam and Philadelphia: John Benjamins. Auer, Peter and di Luzio, Aldo. 1984. “Introduction”. In Interpretive Sociolinguistics: Migrants, Children, Migrant Children, P. Auer and A. di Luzio (eds.), vii-x. Tübingen: Narr. Auer, Peter and di Luzio, Aldo (eds.). 1992. The Contextualization of Language. Amsterdam and Philadelphia: John Benjamins. Auer, Peter, Couper-Kuhlen, Elizabeth and Müller, Frank. 1999. Language in Time: The Rhythm and Tempo of Spoken Interaction. Oxford: Oxford University Press. Ballim, Afzal and Wilks, Yorick. 1991. ArtiWcial Believers: The Ascription of Belief. Hillsdale, NJ: Lawrence Erlbaum. Bateson, Gregory. 1971. “Communication”. In The Natural History of an Interview, N.A. McQuown (ed.). MicroWlm collection of manuscripts on Cultural Anthropology 95 (XV):1–40. Chicago: University of Chicago Library. Bateson, Gregory. 1981/1995. “Communication”. In La Nouvelle Communication [French translation of Bateson 1971], Y. Winkin (ed.), 116–144. Paris: Editions du Seuil. Bazzanella, Carla. 1990. “Phatic connectives as interactional cues in contemporary spoken Italian”. Journal of Pragmatics 14:629–647. Bazzanella, Carla. 1994. Le Facce del Parlare: Un Approccio Pragmatico all’Italiano Parlato. Firenze: La Nuova Italia. Berge, Kjell L. 1994. “Communication”. In The Encyclopedia of Language and Linguistics, R. E. Asher (ed.), 614–620. Oxford: Pergamon Press. Berger, Charles R. 1997. Planning Strategic Interaction. Mahwah, NJ: Lawrence Erlbaum. Berger, Charles R. and Bradac, James J. 1982. Language and Social Knowledge: Uncertainty in Interpersonal Relations. London: Arnold. Blom, Jan-Petter and Gumperz, John J. 1972. “Social meaning in linguistic structure: Code-switching in Norway”. In Directions in Sociolinguistics: The Ethnography of Communication, J.J. Gumperz and D. Hymes (eds.), 407–434. New York: Holt, Rinehart and Winston.
73
74
Language and interaction
Blum-Kulka, Shoshana and Weizman, Elda. 1988. “The inevitability of misunderstandings: Discourse ambiguities”. Text 8(3):219–241. Brown, Penelope and Fraser, Colin. 1979. “Speech as a marker of situation”. In Social Markers in Speech, K.R. Scherer and H. Giles (eds.), 33–62. Cambridge: Cambridge University Press. Brown, Penelope and Levinson, Stephen C. 1979. “Social structure, groups and interaction”. In Social Markers in Speech, K.R. Scherer and H. Giles (eds.), 291–341. Cambridge: Cambridge University Press. Castelfranchi, Cristiano. 1994. “Ma non dica idiozie!”: Per un modello delle interazioni verbali al di là della conversazione”. In Fra Conversazione e Discorso: L’Analisi dell’Interazione Verbale, F. Orletti (ed.), 143–170. Rome: NIS. Clark, Herbert H. 1999. “On the origins of conversation”. Verbum XXI (2):147–161. Dascalu Jinga, Laurentia and Vanelli, Laura. 1996. “Mi raccomando eh!: A pragmatic and phonetic analysis of the Italian interjection eh”. Lingua e Stile 31 (3):393–431. di Luzio, Aldo. “Presenting John J. Gumperz”. This volume. Duranti, Alessandro. 1985. “Sociocultural dimensions of discourse”. In Handbook of Discourse Analysis 1, T.A. Van Dijk (ed.), 193–230. London: Academic Press. Duranti, Alessandro. 1992. EtnograWa del Parlare Quotidiano. Rome. La Nuova Italia ScientiWca. Duranti, Alessandro. 1997. Linguistic Anthropology. Cambridge: Cambridge University Press. Ensink, Titus. 1987. “Interpretative processes in discourse. The approach by John Gumperz”. Journal of Pragmatics 11:517–531. Giles, Howard, Scherer, Klaus and Taylor, D.M. 1979. “Speech markers in social interaction”. In Social Markers in Speech, K. Scherer and H. Giles (eds.), 343–381. Cambridge: Cambridge University Press. Goodwin, Charles and Goodwin, Marjorie H. 1992. “Context, activity and participation”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 77–99. Amsterdam and Philadelphia: John Benjamins. Grosjean, Michèle. 2001. “Verbal et non-verbal dans le langage au travail”. In Langage et Travail: Communication, Cognition, Action, A. Borzeix and B. Fraenkel (eds.), 143– 166. Paris : CNRS Editions. Grundy, Peter. 1995. Doing Pragmatics. London: Arnold. Gumperz, John J. 1972. “Introduction”. In Directions in Sociolinguistics: The Ethnography of Communication, J.J. Gumperz and D. Hymes (eds.), 1–25. New York: Holt, Rinehart and Winston. Gumperz, John J. 1974. The Sociolinguistics of Interpersonal Communication. Urbino, Italy: Centro Internazionale di Semiotica e di Linguistica. Working Papers and Prepublications 33, serie C. Gumperz, John J. 1977. “Sociocultural knowledge in conversational inference”. In Linguistics and Anthropology [Georgetown University Round Table on Languages and Linguistics 1977], M. Saville-Troike (ed.), 191–211. Washington, DC: Georgetown University Press.
Gumperz and the minims of interaction
Gumperz, John J. 1979. “The retrieval of sociocultural knowledge in conversation”. Poetics Today 1 (1–2):273–286. Gumperz, John J. 1980. “The sociolinguistic basis of speech act theory”. Versus 26/ 27:101–121. Gumperz, John J. 1982a. Discourse Strategies. Cambridge: Cambridge University Press. Gumperz, John J. 1982b. “The linguistic bases of communicative competence”. In Analyzing Discourse: Text and Talk [Georgetown University Round Table on Languages and Linguistics 1981], D. Tannen (ed.), 323–334. Washington, DC: Georgetown University Press. Gumperz, John J. (ed.). 1982. Language and Social Identity. Cambridge: Cambridge University Press. Gumperz, John J. 1984. “Communicative competence revisited”. In Meaning, Form, and Use in Context: Linguistic Applications, D. SchiVrin (ed.), 278–289. Washington, DC: Georgetown University Press. Gumperz, John J. 1989a. “Contextualization cues and metapragmatics: The retrieval of cultural knowledge”. CLS 25 (2):77–88. Gumperz, John J. 1989b. “Cadrer et comprendre une politique de la conversation”. In Le Parler Frais d’Erving GoVman, I. Joseph et al. (eds.), 123–154. Paris: Les Editions de Minuit. Gumperz, John J. 1990a. “Conversational cooperation in social perspective”. Parasession on the Legacy of Grice. Berkeley Linguistic Society 16:429–441. Gumperz, John J. 1990b. “Theory and method in pluriglossia: The interpretive analysis of language usage”. In Aspetti Metodologici e Teorici nello Studio del Plurilinguismo nei Territori dell’Alpe-Adria, L. Spinozzi Monai (ed.), 33–49. Tricesimo: Aviani Editore. Gumperz, John J. 1992. “Contextualization revisited”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 39–53. Amsterdam and Philadelphia: John Benjamins. Gumperz, John J. 1993. “Culture and conversational inference”. In The Role of Theory in Language Description, W.A. Foley (ed.), 193–214. Berlin and New York: Mouton de Gruyter. Gumperz, John J. 1995a. “Mutual inferencing in conversation”. In Mutualities in Dialogue, I. Markovà et al. (eds.), 101–123. Cambridge: Cambridge University Press. Gumperz, John J. 1995b. Personal communication to C.L. Prevignano, Bologna, 31 March. Gumperz, John J. 1996a. “Introduction to part IV”. In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 359–373. Cambridge: Cambridge University Press. Gumperz, John J. 1996b. “The linguistic and cultural relativity of conversational inference”. In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 374– 406. Cambridge: Cambridge University Press. Gumperz, John J. 1999. “On interactional sociolinguistic method”. In Talk, Work and Institutional Order, S. Sarangi and C. Roberts (eds.), 453–471. Berlin and New York: Mouton de Gruyter.
75
76
Language and interaction
Gumperz, John J. 2001a. “Contextualization and ideology in intercultural communication”. In Culture in Communication: Analyses of Intercultural Situations, A. di Luzio, S. Günthner and F. Orletti (eds.), 35–53. Amsterdam and Philadelphia: John Benjamins. Gumperz, John J. 2001b. “Interactional Sociolinguistics: A personal perspective”. In The Handbook of Discourse Analysis, D. SchiVrin et al. (eds.), 215–228. Oxford: Basil Blackwell. Gumperz, John J. and Berenz, Norine. 1993. “Transcribing conversational exchanges”. In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.D. Lampert (eds.), 91–121. Hillsdale, NJ: Lawrence Erlbaum. Gumperz, John J. and Cook-Gumperz, Jenny. 1982. “Introduction: Language and the communication of social identity”. In Language and Social Identity, J.J. Gumperz (ed.), 1–21. Cambridge: Cambridge University Press. Gumperz, John J. and Cook-Gumperz, Jenny. 1994. “The politics of a conversation: Conversational inference in discussion”. In What’s Going On Here? Complementary Studies of Professional Talk, A. Grimshaw et al., 373–394. Norwood, NJ: Ablex. Gumperz John J. and Cook-Gumperz, Jenny. 1996. “Treacherous words: Gender and power in academic assessment”. Folia Linguistica XXX (3–4):167–188. Gumperz, John J. and Herasimchuk, Eleanor. 1973. “The conversational analysis of social meaning: A study of classroom interaction”. In Sociolinguistics: Current Trends and Prospects [Georgetown University Round Table on Languages and Linguistics 1972], R.W. Shuy (ed.), 99–134. Washington, DC: Georgetown University Press. Gumperz, John J. and Hymes, Dell. 1972a. [“Note to H. GarWnkel’s ‘Remarks on ethnomethodology’”]. In Directions in Sociolinguistics: The Ethnography of Communication, J.J. Gumperz and D. Hymes (eds.), 301–309. New York: Holt, Rinehart and Winston. Gumperz, John J. and Hymes, Dell. 1972b. [“Note to H. Sacks’ ‘On the analyzability of stories by children’”]. In Directions in Sociolinguistics: The Ethnography of Communication, J.J. Gumperz and D. Hymes (eds.), 325–329. New York: Holt, Rinehart and Winston. Gumperz, John J. and Hymes, Dell. 1972c. [“Note to E.A. SchegloV’s ‘Sequencing in conversational openings’”]. In Directions in Sociolinguistics: The Ethnography of Communication, J.J. Gumperz and D. Hymes (eds.), 346–348. New York: Holt, Rinehart and Winston. Gumperz, John J. and Hymes, Dell (eds.). 1972. Directions in Sociolinguistics: The Ethnography of Communication. New York: Holt, Rinehart and Winston. Gumperz, John J. and Levinson, Stephen C. 1996. “Introduction: Linguistic relativity reexamined”. In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 1–8. Cambridge: Cambridge University Press. Gumperz, John J. and Levinson, Stephen C. (eds.). 1996. Rethinking Linguistic Relativity. Cambridge: Cambridge University Press. Kendon, Adam. 1990. “Some context for Context Analysis: A view of the origins of structural studies of face-to-face interaction.” In Conducting Interaction: Patterns of Behavior in Focused Encounters, 15–49. Cambridge: Cambridge University Press.
Gumperz and the minims of interaction
Kendon, Adam. 1996. “Cues of context” [Review article of The Contextualization of Language, P. Auer and A. di Luzio (eds.). Amsterdam and Philadelphia: John Benjamins, 1992]. Semiotica 109 (3/4):349–356. Kerbrat-Orecchioni, Catherine. 1997. “A multilevel approach in the study of talk-ininteraction”. Pragmatics 7 (1):1–20. Kerbrat-Orecchioni, Catherine. 1998. “La notion d’interaction en linguistique: Origines, apports, bilan”. Langue Française 117:51–67. Koerner, E.K.F. 1973. Ferdinand de Saussure: Origin and Development of his Linguistic Thought in Western Studies of Language. A Contribution to the History and Theory of Linguistics. Braunschweig: Vieweg. Levinson, Stephen C. 1979. “Activity types and language”. Linguistics 17:365–399. Levinson, Stephen C. 1995. “Interactional biases in human thinking”. In Social Intelligence and Interaction, E.N. Goody (ed.), 221–260. Cambridge: Cambridge University Press. Levinson, Stephen C. “Contextualizing ‘contextualization cues’”. This volume. Murray, Stephen O. 1994. Theory Groups and the Study of Language in North America: A Social History. Amsterdam and Philadelphia: John Benjamins. Orletti, Franca. 1994. “Sulla superWcie del conXitto”. In Fra Conversazione e Discorso: L’Analisi dell’Interazione Verbale, F. Orletti (ed.), 171–184. Rome: NIS. Orletti, Franca (ed.). 1994. Fra Conversazione e Discorso: L’Analisi dell’Interazione Verbale. Rome: NIS. Prevignano, Carlo. 1979. “Una tradizione scientiWca slava tra linguistica e culturologia”. In La Semiotica nei Paesi Slavi: Programmi, Problemi, Analisi, C. Prevignano (ed.), 23–99. Milan: Feltrinelli. Prevignano, Carlo. 1991. “Dopo la ‘conclusione deprimente’ di Brown e Yule, come studiare ponti pragmatici in dialoghi?”. In Dialoganalyse III, S. Stati et al. (eds.), 211– 225. Tübingen: Niemeyer. Prevignano, Carlo. 1993. “On reconstructions of dialogues as interplanning cases”. In Dialoganalyse IV, H. LöZer (ed.), 183–188. Tübingen: Niemeyer. Prevignano, Carlo. 1994. “From scripts to interscripting in dialogue analysis”. Paper presented at the 5th IADA Congress on Nouvelles Perspectives dans l’Analyse de l’Interaction Verbale. Paris, Université de la Sorbonne Nouvelle, 17–19.3.1994. Prevignano, Carlo. 1995. “Cross-framing in dialogue”. Paper presented at the 1995 IADA Round Table on Dialogue Analysis: Units, Relations and Strategies Beyond the Sentence. Bologna, Department of Linguistic and Oriental Studies, University of Bologna, 31.3–1.4.1995. Prevignano, Carlo L. 1997. “Gumperz and the minims of interaction”. In Discussing Communication Analysis 1: John J. Gumperz, S. Eerdmans, C. Prevignano, and P.J. Thibault (eds.), 48–56. Lausanne: Beta Press. Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This volume. Prevignano, Carlo L. and Thibault, Paul J. 2003 “Continuing the interview with Emanuel A. SchegloV”. In Discussing Conversation Analysis: The Work of Emanuel A. SchegloV. C.L. Prevignano and P.J. Thibault (eds.), 165–171 Amsterdam and Philadelphia: John Benjamins.
77
78
Language and interaction
Raskin, Victor. 1981. “Script-based lexicon”. Quaderni di Semantica II (1):25–34. Rawlins, William K. 1987. “Gregory Bateson and the composition of human communication”. Research on Language and Social Interaction 20:53–77. Sarangi, Srikant. 2000. “Activity types, discourse types and interactional hybridity: The case of genetic counselling”. In Discourse and Social Life, S. Sarangi and M. Coulthard (eds.), 1–27. Harlow: Pearson Education Limited. Scherer, Klaus and Giles, Howard. (eds.). 1979. Social Markers in Speech. Cambridge: Cambridge University Press. SchiVrin, Deborah. 1987. Discourse Markers. Cambridge: Cambridge University Press. SchiVrin, Deborah. 1994. Approaches to Discourse. Oxford: Blackwell. Selting, Margret and Couper-Kuhlen, Elizabeth. 2001. “Forschungsprogramm ‘Interaktionale Linguistik’”. Linguistische Berichte 187:257–287. Stati, Sorin. 1982. Il Dialogo: Considerazioni di Linguistica Pragmatica. Naples: Liguori. Stati, Sorin. 1990. Le Transphrastique. Paris: PUF. Svennevig, Jan. 1999. Getting Acquainted in Conversation: A Study of Initial Interactions. Amsterdam and Philadelphia: John Benjamins. Thibault, Paul J. “Contextualization and social meaning-making practices”. This volume. Thomas, Jenny. 1995. Meaning in Interaction: An Introduction to Pragmatics. London: Longman. Winkin, Yves. 1981/1995. “Présentation générale”. In La Nouvelle Communication, Y. Winkin (ed.), 11–109. Paris: Editions du Seuil. Winkin, Yves (ed.). 1981/1995. La Nouvelle Communication. Paris: Editions du Seuil.
Commentary on a discussion with John J. Gumperz
Chapter 6
A commentary on a discussion with John J. Gumperz1 Afzal Ballim
My contribution to this volume is that of a non-linguist. At the same time, it is that of a non-linguist who has worked for many years with computational linguists, and who works in the Weld of Natural Language Processing (NLP) — the treatment by computer of natural language. As such, I hope to be able to comment on the discussion with Gumperz from two perspectives: 1. In what way may Gumperz’s work in sociolinguistics and the ethnography of communication be of beneWt to those of us working in the computer treatment of language? 2. How may our work aid, or eVect, that of Gumperz? I have worked in the Weld broadly known as Natural Language Processing for over a decade, and during that time I have worked with both computational linguists and people who, like myself, have an interest in linguistics and linguistic theories, yet whose concern with computers using language has led them down diVerent paths from those explored by linguists. This has prompted me to ask myself why the work of linguists does not have a greater eVect on work within NLP. After all, we are both concerned with the same fundamental thing: language. Surely, with this common point of interest, there should be more interaction, in particular between NLP researchers and computational linguists who have computational models as an extra point of commonality? I believe the reason for the minimal interaction between the two is due to the fact that: a. NLP researchers believe that the work of computational linguists, rooted as it is in providing theoretical models of language, tends in many instances to be
79
80
Language and interaction
too far from the practical problems of using language for it to have much bearing on what they do; b. Computational linguists believe that the work of NLP researchers often lacks the rigorous theoretical models that they feel are necessary to explicate language. Of course, interaction does take place, and the work in computational linguistics on phonology, morphology, part-of-speech tagging, and syntactical analysis plays an important role in modern NLP. It is when one goes beyond these realms to semantics and pragmatics that the interaction breaks down. In part, what I wish to say in this commentary is that the work of Gumperz is an example of why the interaction should not fail. In order to situate my detailed comments on the discussion, it is necessary to give a brief description of my own work in NLP. I am concerned with all forms of computer-mediated communication; in particular, with humanmachine interaction (and in this I include language-based human-machine communication), as well as with human-human communication assisted by computer (such as in Computer-Supported Collaborative Work [CSCW]). The ultimate goal in this work is to create systems that communicate. As such, I am less encumbered by having to provide neat theoretical accounts of linguistic phenomena than are my colleagues in computational linguistics (although such neat theories are still desirable). Instead, I can take shortcuts around such philosophical problems as “symbol grounding” or even “what is semantics?”. Indeed, in recent years my work has concentrated on “robust interaction”, in which we admit that the computer interpretation and model of the interaction is incomplete, but we try to work with what we know the computer can do. This tendency towards robustness and practicality has been notable in work on human-computer interaction since the mid-90s. If one can describe the desired functionality of a system, and the system is able to communicate with a human to achieve that functionality, then the goal has been reached and it is not necessary to ask if the computer system has understood. That said, the sheer complexity of the problem of achieving communication between a computer system and a human is such that one cannot help but formulate theories that may well be applicable to humanhuman communication as well. For the same reasons, theories of humanhuman communication must be considered, as they are still important.
Commentary on a discussion with John J. Gumperz
The crux of my work (see Ballim and Wilks 1991; Ballim and Pallotta, forthcoming) is in interpreting utterances in the light of their context as perceived by the speakers and hearers involved. That is, the system formulates a model of the knowledge, beliefs, desires, intentions, and other attitudes of the interlocutors that are relevant in the context of the dialogue, and furthermore it formulates models of how the interlocutors model each other at that point. These models are then used as the basis for interpretation of the discourse. They are built using past and current experience with the individual interlocutors, as well as through generalizations about the background of various participants. In general terms, I feel that this approach is very compatible with Gumperz’s ethnographically-based sociolinguistic analysis of dialogue, since notions such as “networks of relationships”, “situated inference”, and “speech events” are all fundamental to the processing involved. In turn, the eVectiveness of this NLP method can be seen as providing indirect evidence of the validity of Gumperz’s approach. This can be seen more clearly by considering in more detail speciWc parts of the discussion.
Indirect reasoning processes This notion is fundamental to my approach to discourse, by which discourse must be interpreted with respect to the underlying reasoning processes that the participants in a dialogue mutually believe each other to use. Peers would normally believe each other to use similar reasoning, hence their communication would tend not to explicitly emphasize those processes. In this way, communication between children would be explained by reference to the types of reasoning that children would be expected to use. This, in turn, would allow for hypothesis formation about what aspects we would expect to see explicitly lexicalized, and what aspects we would expect to be implicit. Given hypotheses about the reasoning processes, my system could be used as a predictive model for peer communication, which could be used as veriWcation of these hypotheses.
81
82
Language and interaction
Speech events Gumperz insists that ethnographically-based sociolinguistic analysis must focus on speciWc speech events. In other words, it is not suYcient to consider discourse in terms of a uniform, disembodied community. There is no denying, however, that this notion of a community has got a role to play, but that role must be mediated by the personal experiences of the individuals involved. Thus an individual may display a discourse behavior consistent with a particular ethnic make-up when conversing with peers, yet display a diVerent behavior when involved in discourse with people from other backgrounds. This notion of interactively deWned behavior is also a key concept in my own approach to discourse, in which dialogue is conditioned by interlocutors’ perceptions of each other and of the context.
Context and communication The notion of the context of a communication has long been central in NLP. My own work with Yorick Wilks, as well as that of people such as Barbara Grosz, and Barwise and Perry’s (1983) “situation semantics”, are all examples of taking context as being of primary importance in discourse understanding. However, Gumperz’s notion of “contextualization cues” is not something accounted for or used by any of these theories. His aim to show the interaction of indexical signs with symbolic signs may, I believe, make a great contribution to NLP, where the problems of deWning the structure of discourse and demarcating shifts in context are far from resolved. Gumperz presents a schema for overcoming some of the ambiguity inherent in the inferential process associated with discourse. I believe that combining this schema with the methods commonly used in NLP, and those adopted in my own work (of Wnding hypotheses that are maximally consistent with the context and interlocutors’ background) may yield an interesting and novel approach to problems of ambiguity. The method that I mention in my work of Wnding maximally consistent hypotheses that explain discourse events has undercurrents of Gumperz’s “activity types”, in particular because they may be based on predictive reasoning on where the conversation is leading the participants. Such predictive structures are of course a notable aspect of much NLP work (in particular,
Commentary on a discussion with John J. Gumperz
Schank’s work on scripts and the derivatives from that work can be seen as aiming to embody form-context relationships).
Corpus pragmatics Gumperz is questioned about the use of corpora. His problem with such corpora is that these transcriptions lack many of the contextual cues that he believes to be fundamental to understanding conversation. A large part of the problem is the lack of standards for recording such cues. Given the growing and already widespread use of corpora in computational linguistics and NLP, a contribution in this domain which establishes standards for those cues would be of great beneWt to many people.
Conclusion In these few paragraphs, I have striven to give an appreciation of Professor Gumperz’s work from a diVerent perspective, that of someone working on methods to enable the use of language as a communicative medium between humans and computers. I hope that the comments have shown that, for myself at least, there is a lot to be learned from Gumperz’s research by people working in NLP, and that, reciprocally, work in NLP can be of beneWt to those working in areas such as sociolinguistics.2 Work on applied computational linguistics (sic) in recent years has shown that people are taking greater interest in using linguistic theories in the very practical process of making a computer interact intelligently with humans. It remains to be seen, however, whether we will be able to move away from using small parts of linguistic theories taken piece-meal, towards a greater linguistic foundation for our interactive systems of the future.
Notes 1. This paper was revised in February 2002. 2. See Coakes et al. 2000, Goddard 1998, Greif 1998, Horton and Keysar 1996, Lyons 1977, and Sproull & Kiesler 1988 for further relevant work in the areas of semantics, pragmatics, Computer-Supported Collaborative Work (CSCW), and NLP.
83
84
Language and interaction
References Ballim, Afzal and Pallotta, Vincenzo. Forthcoming. Dialogue through the ViewFinder. Ballim, Afzal and Wilks, Yorick. 1991. ArtiWcial Believers: The Ascription of Belief. Hillsdale, NJ: Lawrence Erlbaum. Barwise, Jon and Perry, John. 1983. Situations and Attitudes. Cambridge, Mass.: The MIT Press. Coakes, Elayne, Willis, Dianne and Lloyd-Jones, Raymond (eds.). 2000. The New Sociotech: GraYti on the Long Wall. London: Springer. Goddard, CliV. 1998. Semantic Analysis: A Practical Introduction. Oxford: Oxford University Press. Greif, Irene. 1998. Computer-Supported Cooperative Work: A Book of Readings. San Mateo, CA: Morgan Kaufmann. Horton, William S. and Keysar, Boaz. 1996. “When do speakers take into account common ground?” Cognition 59:91–117. Lyons, John. 1977. Semantics. Cambridge: Cambridge University Press. Sproull, Lee and Kiesler, Sara. 1988. “Reducing social context cues: Electronic mail in organizational communication”. In Computer-Supported Cooperative Work: A Book of Readings, I. Greif (ed.), 683–712. San Mateo, CA: Morgan Kaufmann.
John J. Gumperz: A review
Chapter 7
A review of John J. Gumperz’s current contributions to Interactional Sociolinguistics1 Susan L. Eerdmans
1. Introduction In this chapter, I would like to outline some of John Gumperz’s most important contributions to Interactional Sociolinguistics in the last few years. An attempt will be made to deWne in a nutshell Gumperz’s particular analytic approach, and to demonstrate, with examples taken from two of his case studies, how this is applied to real data. ‘Interactional Sociolinguistics’ is a general term used to describe a number of qualitative approaches to the analysis of interactive and dialogic processes in interpersonal communication. Its roots lie in the Welds of sociology, anthropology and linguistics, as reXected by interactional sociolinguists’ interest in the closely-knit interplay existing between language, society and culture (SchiVrin 1994, 1996), and in the “communicative import of diversity” (Gumperz 1999b:453). The focus of their attention is on naturally-occurring utterances in social context, the aim of their analysis being to discover and understand “how interpretation and interaction are based upon the interrelationship of social and linguistic meanings” (SchiVrin 1994:8).
2. Background As a linguistic anthropologist, John Gumperz has contributed to interactional approaches to discourse in a signiWcant way, “providing both the empirical evidence and the analytical framework for investigating the varied but system-
85
86
Language and interaction
atic ways in which language shift both reXects and deWnes social and cultural boundaries” (Gumperz 1992a:229). By focussing his sociolinguistic analysis on the qualitative aspects of verbal and non-verbal processes, together with those of shared cognition and understanding, Gumperz has demonstrated his commitment to a general theory. This is not only a theory of language, but of verbal communication, of communicative action, “which integrates what we know about grammar, culture and intercultural conventions into a single, overall framework of concepts and analytical procedures” (Gumperz 1982:4). The analytical methodology adopted by Gumperz for his in-depth investigation of discursive processes in what GarWnkel calls “naturally organized activities” is based principally on the inductive procedure of analysis adhered to by conversation analysts. 2 The aim of the conversation analytic, turn-by-turn study of sequential processes is to discover the naturally-occurring and recurring mechanisms and patterns of conversational management, seeking “to describe the underlying social organization[...]through which orderly and intelligible social interaction is possible” (Goodwin and Heritage 1990:283).3 In keeping with conversation analysts’ ethnomethodological background, which leads them to avoid any premature generalization or idealization, this research is carried out assuming very little about participants’ social and cultural background a priori; aspects of the context such as the social identity of participants and the setting are considered to be “a category of social life and conduct that is subject to locally situated interpretive activity” (SchiVrin 1994:235). Alongside elements of the conversation analytic method of investigating sequential processes, Gumperz’s interactional sociolinguistics also employs features of more general approaches to discourse analysis, including, for example, Gricean analysis of inferential processes and implicature. One of the key notions of Gumperz’s perspective on verbal communication is communicative intent, the interpretation made by the listener(s) of what the speaker is trying to communicate, roughly correspondent to the illocutionary force of what is conveyed (Gumperz 1992a:230).4 The assessment of participants’ communicative intentions is an essential part of conversational inference, another fundamental concept that Gumperz is concerned with, deWned as “the situated and presupposition-bound interpretive processes by which interlocutors assess what they perceive at any one point in a verbal encounter and on which they base their responses” (Gumperz 1996c:375). In the context of a “communicative ecology” (Gumperz 1990), conversational inference thus
John J. Gumperz: A review
corresponds to the interactionally-based process by which speakers and listeners cooperate to produce and evaluate speech. According to Gumperz, this process of conversational inferencing implies a three-part analysis of meaning and interpretation, in that at least three speaking turns are involved: move, counter-move and conWrmation or disconWrmation.5 One of the constituent moves of such a three-part exchange may also be non-linguistic, as illustrated by GoVman in the following example: A: [Enters wearing a new hat] B: No, I don’t like it. A: Now I know it’s right. (GoVman 1976:290, cited in Tsui 1989:547)
In his current research, Gumperz’s basic assumption about language owes a great deal to work in metapragmatics carried out mainly in the United States and particularly by Michael Silverstein and his students.6 This approach to the analysis of verbal communication is of a semiotic functional nature, its key concept being the Peircian notion of indexicality and the distinction that Peirce makes between symbolic and indexical signs. By symbolic signs, Peirce basically refers to arbitrary and conventional signs which denote their object on the basis of a general law or convention. For Gumperz, symbolic signs, which he describes as “denotational” (Gumperz and Toribio 1999), include the information that is conventionally conveyed via lexical meaning and grammatical processes. With indexical signs, on the other hand, the signaling value rests on the conventional associations between particular kinds of indexes or signs and context, this being a direct linkage, in no way mediated by any symbolic process. Symbolic signs therefore communicate by means of grammatical and lexical rules, whereas indexical signs communicate “by virtue of direct conventionalized associations between signs and context established or transmitted through previous communicative experience” (Prevignano and di Luzio, this volume: 10). We can further distinguish between two main types of indexical signs: (1) well-known, referential indexes (in Jakobson’s terms, “shifters”), such as deixis, anaphora and other discourse markers, which have at least some propositional content, some kind of semantic component. These indexical signs refer directly to context and can therefore be explained only in terms of
87
88
Language and interaction
how they are used in context, or, as Gumperz argues, in terms of how they enter into a particular interactive or dialogic process; (2) what Silverstein calls “pure indexes” whose signaling value is basically relational, indexical, in that they possess no propositional content. These indexical signs refer to, or rather, index a context; this signaling process does not depend on any kind of logical entailment, but on “conventionalised associations between co-occurring clusters of signs and contextual frames” (Gumperz 1995:118). One set of pure indexical signs Gumperz is particularly concerned with is that of contextualization cues, “verbal and non-verbal metalinguistic signs that serve to retrieve the context-bound presuppositions in terms of which component messages are interpreted” (Gumperz 1996c:379). These cues operate principally at the level of prosody, paralinguistic signs, code choice and choice of lexical expressions or formulaic sequences (Gumperz 1993), their meaning being established through interaction and not by the assignment of context-free semantic content. In this way, they convey limitations on interpretation, for example by cueing “the implicatures or lines of practical reasoning that enter into the interpretation of a particular word or expression” (Gumperz 1992b:50). The background knowledge necessary for conversational inference is made available by means of such contextualization cues, which “when processed in co-occurrence with symbolic grammatical and lexical signs [serve] to construct the contextual ground for situated interpretation, and thereby [aVect] how constituent messages are understood” (Gumperz 1999b:461). Gumperz’s approach to the analysis of discursive practices is very much focussed on how this interaction between indexically-linked signals, on the one hand, and symbolic signs, together with the sequential organization of speech exchanges, cultural and other background information, on the other, forms everyday social action, that is to say, goal-oriented communicative action.
3. Research method To illustrate the methodological procedures adopted by Gumperz when carrying out his research, examples will be taken from two particular case studies, one concerning a job center interview and the other a rape trial cross-examination.7 The advantage of using naturally-organized situations of this kind is
John J. Gumperz: A review
that they are in a sense socioculturally-deWned, in that we have a relatively clear idea of what cultural presuppositions they involve. In other words, we possess typiWed knowledge about such events, these typiWcations being culturally based: that is, they are built up and acquired over time through “historically speciWc interactive experience” (Gumperz 1992b:45). 8 By ‘culture’ here, Gumperz refers to cultural knowledge in the linguistic sense of the term, that is, the knowledge that we need, that we rely on, to communicate.9 The starting point of Gumperz’s interactional investigation is ethnographic research.10 Whereas the second case study discussed here, the rape trial cross-examination, is a re-analysis of a study carried out by Paul Drew (Drew 1992), in which Gumperz’s main aim is to demonstrate the relative inadequacy of sequential analysis in accounting for situated interpretation, the Wrst example, the job center interview, is a clear exempliWcation of Gumperz’s ethnographic approach. The data for the latter case study were collected during a period of about three years’ research on intercultural communication, mainly carried out in Great Britain. In this research study, Gumperz began his work as ethnographer with a period of interviewing in order to discover people’s opinions about their communication with people from diVerent ethnic backgrounds, in this case, Asians and Cypriots in London and the Midlands. Like most ethnographers, Gumperz started his investigation by asking a lot of seemingly random questions, searching for a hypothesis that would lead to a particular kind of query: “What is it about intercultural communication that people see as signiWcant?”. He was thus able to start to get a feel of the situation and to obtain a great deal of ethnographic information on “the history, social characteristics, everyday practices, and goals” of the community concerned (Gumperz and Berenz 1993:92). One of the most important elements of ethnographic Weldwork is participant observation in the community so that the researcher can begin to absorb the atmosphere and familiarize him/herself with the communicative ecology of the setting. Through contacts with the National Centre for Industrial Language Training, which uses ethnographic approaches to intercultural communication in order to diagnose, for example, problems on the shop Xoor, Gumperz was in fact able to participate as observer in various work situations, including a cafeteria at Heathrow Airport.11 In this way, he could observe for himself the occurrences of conversational misunderstanding and breakdown and obtain Wrst-hand information on the dynamics of the intercultural verbal
89
90
Language and interaction
and non-verbal interaction taking place. In interactional studies of this kind, audio and/or video recordings of varying lengths are made of the encounters most likely to prove useful for testing the ethnographer’s hypotheses, and this recorded material is then analyzed at the level of content and rhythmic organization in order to identify thematically coherent and sequentially-bounded units, that is, speech events. These can be deWned as “temporally organized sequences of verbal acts, separable from the surrounding talk by empirically identiWable beginnings and ends or outcomes, which provide evidence with which to validate the analysis of constituent interpretive processes” (Gumperz 1996c:380). The next step in the analytic procedure is to make detailed transcriptions of the isolated event sequences in order to prepare interactional texts and analyze them in greater depth. 12 The aim of this analysis is to investigate the interactive features of conversational management and to discover “the perceptual cues that speakers and audiences rely on in signaling what they intend to convey and in assessing what they hear” as part of conversational inference (Gumperz and Berenz 1993:119). These perceptual cues can be verbal or nonverbal and include paralinguistic, prosodic, formulaic and other contextualization cues such as style-shifting and code-switching.
3.1 Job center interview The Wrst case study to be examined here, the job center interview, reXects very clearly Gumperz’s ever-present interest in intercultural communication, and demonstrates the extent to which interactional sociolinguistic analysis can reveal the stated and unstated presuppositions that enter into interactions of this kind. As mentioned previously, the interview, aimed at obtaining information about the candidate’s past work experience, was recorded in the Midlands as part of a three-year project on intercultural communication. The interviewer was a native speaker of the English spoken in that area, and the interviewee a South Asian living locally.13 As one of a number of such situations, “compared systematically with a range of other encounters both within and outside the particular institutional context” (Gumperz and Berenz 1993:93), the analysis of this interview reveals the interviewer’s preoccupation with Wnding out exactly what the candidate does, the actual work he performs. With this knowledge, she is able to judge the kind of technical skills involved and
John J. Gumperz: A review
consequently obtain detailed information about the person as an individual. The South Asian candidate, however, consistently disappoints the interviewer by giving only general descriptions of his work program. Although he correctly interprets her responses (“yeah”, “hmm” etc.) as expressions of dissatisfaction and therefore as prompts eliciting more details, he does not have a clear idea of what she is looking for. Even when, more and more impatient, she uses the prosodic strategy of accenting to focus her questions on him as an individual, “what do you actually *do?”, “how involved are *you?”, he still gives a general answer, “he shows us what to do.”14 An obvious eVect of a sequence of nonspeciWc responses of this kind is that it gives the impression of noncooperation, and can therefore lead to negative attitudinal judgements by the person conducting the interview. Gumperz’s argument is that some of the responses the interviewer receives from the candidate are actually generated by the interaction itself. In fact, one of the most striking features of the dialogue is that at no point does the interviewer really make clear exactly what she is interested in. She is never very speciWc lexically and only indicates the extent of her dissatisfaction indirectly by modulating the prosody of her single word responses or by accenting key words. Not surprisingly, these indirect strategies used by the native speaker to elicit more detailed information are not transparent enough for the candidate; even though he realizes there is some problem in the communication, he seems “unable to process the interviewer’s use of stress as a guide to indicate what type of an answer is wanted” (Gumperz 1992c:323).15 The communicative diYculty demonstrated by this case does not seem to be due to lack of formal linguistic knowledge of English on the part of the interviewee, whose grammatical competence is fairly good. The problem instead appears to arise from the fact that his interactive experience of the language has been primarily with other immigrants from South Asia living in his neighborhood. These people have no problems at all when using English interactively amongst themselves, but they sometimes run into trouble when communicating with native speakers, due to their tendency to map the prosodic and rhetorical strategies of their own languages onto English.16 One of the characteristics of this mapping process is that South Asian immigrants like the interviewee do not use stress placement pragmatically when adopting English in dialogic interaction, relying on “phonological means for marking prosodic distinctions and relations” which diVer from those of native speakers of English (Gumperz 1982:122).17 As a consequence, the participants of inter-
91
92
Language and interaction
cultural encounters of this kind frequently experience serious interactional problems, which nearly always imply reciprocal misunderstandings and, in some cases, even breakdowns in communication. Together with these pragmatic causes for native-nonnative misunderstandings, also higher level causes exist, functioning at the level of interactive norms.18 By re-examining transcripts of interviews similar to the one described above, this time focussing his attention on the content of the encounter, Gumperz is in fact able to give a further explanation of what is happening in the interaction and to start to understand what type of “South Asian activity level presuppositions” are involved (Gumperz 1996c:400). One of his Wrst Wndings is that there is a systematic avoidance on the part of the immigrant interviewees to talk about themselves; this is clearly theme-speciWc, in that they tend to avoid revealing information which could be considered an expression of what they feel or believe about certain matters. Another important pattern to emerge from Gumperz’s analysis of topic is that the South Asian speakers of English participating in these interviews again systematically avoid presenting information that they assume their interviewers already know. For example, at the beginning of a face-to-face interview, when asked where they live (the interviewers simply checking the information they have on the form in front of them), the interviewees nearly always give very vague responses, demonstrating their unease by mumbling or stuttering. There thus seems to be a systematic diVerence in the ideology of intercultural relations that governs the conduct in such situations. All over South Asia, social encounters such as these job interviews are seen as “relatively formal hierarchical situations of pleading” (Gumperz 1996c:399, my emphasis), in which, for the context of that particular situation, the interviewee considers himself as hierarchically inferior and the interviewer is regarded as the expert. Volunteering personal information, especially information the interviewer is believed to already possess, is thought to be presumptuous; this explains the systematic stammering and general hesitation characterizing the responses of South Asian interviewees when forced to talk about themselves. Connected to the latter sociolinguistic behavior is the tendency of these immigrants to de-emphasize their qualiWcations, to “background their own accomplishments” (ibid.).19 By letting the interviewer discover this information from the records, rather than provide it themselves and risk appearing boastful, they show that what they are actually doing is pleading with their ‘superior’ to represent their case as an individual.
John J. Gumperz: A review
3.1.1 Concluding remarks In conclusion, it appears to be the case that in intercultural encounters of this kind, serious communicative problems can arise due to ideological diVerences, to diVerent ways of viewing or framing such situations on the part of South Asian speakers of English, the term frame being used by Gumperz to indicate “a class of related presuppositions that guide conduct in certain situations and that are directly related to choice of topic and verbal strategies” (Prevignano and di Luzio, this volume: 27). More speciWcally, it is the interaction taking place between inferences both at an activity and a sequential level, together with “the workings of contextualization conventions” and the general tendency of people to base interpretation on their own culture-bound presuppositions, that may lead to diYculties in communication and possible breakdowns in conversational cooperation (Gumperz 1996c:400). In situations of interethnic miscommunication, the interplay between the various potentially conXicting aspects indicated above tends to result in what Gumperz calls a “compounding” of interactional diYculties: misunderstandings occurring at one level may get repaired, but very often this act of repair is only partially successful or fails completely, thus compounding the communicative problems by adding yet another element of diYculty to an already complicated situation. Such a build-up in communicative tension, not only from the case at hand, but also from cumulative experience of similar situations in the past which may have favored the development of stereotypes, inXuences the interviewer’s “pejorative evaluation” of the candidate’s performance (Gumperz 1996c:400). Problems which in other circumstances would probably be interpreted as examples of lack of linguistic knowledge, in the hierarchical context of the job interview are considered to reXect “the speaker’s ability, truthfulness or trustworthiness” (Gumperz 1992c:326–327). On the other hand, also the interviewee is bound to feel dissatisWed with how the encounter is proceeding, and will probably think that the native speaker is being uncooperative in not giving him the help he needs.20 It would seem that the negative outcome frequently characterizing such “situations of diVerential power and interethnic stigmatization” (Gumperz 1992c:326) is basically due to participants’ diVerent communicative experience. Knowledge of the contextualization practices essential for the negotiation of shared understandings can only be acquired by actual participation over time in discourse communities, communities of learning and practice, particularly in situations where there is active involvement in “goal-oriented
93
94
Language and interaction
endeavours” (Gumperz 1996a).21 Where this communicative competence is in some way lacking, as for example in the case of the South Asian job candidate, the range of contextualization strategies in the individual’s control is limited, as is his skill in adapting these strategies to the situation at hand. In social situations where there are “institutional contacts with majority speakers” (Gumperz 1992c:327), these diVerences in communicative knowledge may become valorized and ideologized and in turn reinforce present or develop into future stereotypes. The stigmatization of South Asian speakers’ communicative practices thus becomes highly likely and cases of actual discrimination in decision-making situations will probably occur, in this way creating a real obstacle to “minority individuals’ success in the society at large” (ibid.).22
3.2 Rape trial cross-examination The second case study concerns a rape trial cross-examination. The purpose of cross-examining, at least in Anglo/American court proceedings, is to reveal, through questioning, inconsistencies in the witness’ testimony with regard to topics covered in the direct examination. In other words, the aim is to ‘deconstruct’ the direct examination, and this in fact is what the following extract deals with. The two main roles in this particular cross-examination are that of the defence counsel (C), who is intent on conveying (even though indirectly) the fact that the victim was actually a consenting participant in the rape attack, and that of the accuser (W), who claims that she was the unsuspecting victim of unprovoked aggression. The verbal interaction between these two opponents can only be fully understood, however, when analyzed in terms of the adversary situation seen as a whole. In fact, a cross-examination like this can be seen as a typical Bahktinian situation: the dialogic exchanges between the counselor and the victim are directed to moving the interaction forward in the traditional conversation analytic way of turn-by-turn sequences, but, at the same time, they have the function of persuading the judge and jury (in Bahktin’s terms, the “supersubject”) to consider their opposite points of view. According to Gumperz, if the transcript of this cross-examination testimony is studied purely in terms of turn-by-turn sequential analysis, “just concentrating on regularities of sequential organization across speech exchanges” (Gumperz 1992a:231), an incomplete picture of the interactive situ-
John J. Gumperz: A review
ation will be drawn. Even though sequentiality or positioning is obviously an essential part of any situated interpretation, it represents only one of the various indexical procedures that inXuence conversational inference. When considered on its own, and “only overtly lexicalised propositional content” is used as data (Gumperz 1999b:458), the study of indirect inferencing will be problematic; for these reasons, Gumperz believes sequential analysis to be insuYcient as an analytic tool. To illustrate this point, some examples have been taken from the exchanges in the cross-examination in which the defence counsel is questioning the victim. 23 When asked whether she had had a “fairly lengthy conversation with the defendant,” she replied, “we:ll we were *all talking;” when confronted with “well *you knew at that ti:me that the defendant was *in:terested in *you, didn’t you?”, she answered, “he: asked me how I’(d) *been;” to the question, “you went to a *ba:r?”, she said, “it’s a *clu:b.” The cross-examination continued along these lines, with the victim never directly assenting to or denying the counselor’s insinuations, but instead in some way modifying the way they had been worded. If analyzed in conversation analytic terms, the victim’s contributions here could be considered as instances of other repair (SchegloV, JeVerson and Sacks 1977), W’s substitution of C’s words with her own being seen as a kind of indirect or “oV-record” correction (Brown and Levinson 1978) of what he was suggesting, rather than a direct answer to his yes/no questions. From this analysis, it could be concluded that W was being rather uncooperative, “and in fact could be accused of interfering with (the defence counsel’s) eVorts to produce a coherent account” (Gumperz 1995:107). However, this possible act of repair on the part of the victim does not occur only once or twice, but, rather strangely for a potentially face-threatening act of this kind, recurs systematically throughout the whole encounter as a regular sequence. For Gumperz, it is the “cumulative eVect” (Gumperz 1995:108) of the victim’s replies that suggests that it is more than just a sequence of repairs, indicating that other information must be brought into the analysis in order for the verbal interaction to be fully understood. Even from a purely sequential analysis perspective, it is necessary to introduce also our knowledge of what constitutes a cross-examination, and, more speciWcally, of what constitutes a rape trial cross-examination, that is, the “historically and culturally based knowledge that can be brought to bear on the interpretation of constituent messages” (Gumperz 1992c:306).
95
96
Language and interaction
Returning to the testimony transcription, it can be seen how a series of discursive oppositions of mainly prosodic, but also syntactic nature, are set up by means of the actual interaction. For example, through the adjacency of “having a fairly lengthy conversation” and “we were all talking,” a syntactic opposition is created between the counsel’s suggestion of an intentional act, “you had a conversation,” and the victim’s description of a state of aVairs, “we were all talking,” thus conveying the idea of contradiction. If, however, we limit our analysis to looking at exchanges of this kind in isolation, the full import of the inferential process does not emerge; it is only through considering a whole series of such oppositions, in which W systematically selects certain key aspects of C’s questions and, echoing them, re-presents them in her own words, that we can begin to infer (using our knowledge of what a cross-examination involves) that the witness is “intentionally and consistently opposing her own interpretation to the counsel’s” (Gumperz 1995:109). She thus appears to be less uncooperative than if she were simply repairing the counselor’s individual utterances, her ‘narrowing’ or focussing of his questions in some way negating what he is suggesting, but, at the same time, through the interaction, conveying a coherent picture of victimhood as opposed to that of a consenting participant. This inferential process is further supported by another kind of discursive opposition, this time of a prosodic nature. In fact, the particular aspects of the attorney’s questions that the victim picks out to reformulate are also rhythmically related, being spaced in such a way that they are almost isochronous. By isochrony, Gumperz refers here to the “repetitive regular spacing of accented rhythmic beats across more than two turns of speaking” (Gumperz 1996c:386).24 In isochronous dialogue, the copying or echoing of a previous speaker’s prosodic contour across turns of speaking, at regular intervals, indicates agreement with what is being said (Selting 1991, cited in Gumperz 1995:109; cf. Selting 1995) and reinforces the point being made. By actually changing the rhythmic pattern of the interlocutor’s speech, on the other hand, the idea of disagreement is given. During the rape trial cross-examination, the victim and defence counsel’s diVerent lines of argumentation are reXected and contextualized in part through the individual rhythmic patterns of their speech. In addition, the accuser/victim adopts the indirect strategy of rhythmic isochrony to copy and highlight those parts of the attorney’s speech she wants to work on. For example, she contrasts “you had some fairly lengthy *conversation” with “we
John J. Gumperz: A review
were *all talking,” and, even more clearly, “you went to a *bar? in *Boston” with “it’s a *club.” In each of these cases, there is a falling prosodic contour with the accent being placed towards the end of the utterance. By instinctively copying this contour in such a way that she is isochronous with these and other key elements of the counsel’s speech, the victim is able to incorporate them unobtrusively into her own argument and metaphorically construct her own version of the story. In so doing, she seems to manage “to neutralize the counsel’s implicatures and to convey her own perspective on the facts without appearing overly uncooperative” (Gumperz 1995:109).
3.2.1 Summary To summarize, throughout this extract of a cross-examination testimony, the impression of victimhood is conveyed partly by linguistic means, but to a much greater extent indexically, by means of a process of indirect inferencing. It is only when this use of indexicality is considered within the context, in terms of the social factors that govern the conduct of this particular situation, that it takes on its real signiWcance; in fact, only in these conditions is it defensible as an essential part of the process of conversational inference.
4. Second language teaching and learning As a concluding note to this paper, I would like to refer to a speciWc research area which has been very much inXuenced, both directly and indirectly, by Gumperz’s work in interactional sociolinguistics: second language teaching and learning. It is now well-established that an individual’s knowledge of how to communicate, his/her ability to negotiate, to cooperate, is not simply a matter of producing and recognizing the appropriate grammar and lexis of the target language. The communicative, or rather, interactional competence which is “co-constructed” (Jacoby and Ochs 1995) by participants in any conversational interaction or interactive practice entails much more than this: to achieve mutual understanding and communicate eYciently in the particular speech event they are involved in, individuals must also have knowledge of the interactive and rhetorical strategies through which factual information is transmitted from speaker to hearer(s) and vice versa (Gumperz and Roberts 1991).25 The level of competence in the use and interpretation of these strate-
97
98
Language and interaction
gies, particularly with regard to “rhetorical eVectiveness” (Gumperz 1994), will depend a great deal on the extent to which interactants’ “‘brought along’ context made up of ideological and metapragmatic assumptions” (Roberts and Sarangi 1999:390) coincide. A crucial part of this “common ground” (Gumperz 2002) is culture-speciWc, acquired through an individual’s direct participation in discursive encounters with, for example, members of his/her family or peer group, as part of the experience of socialization in the community he/she is a member of (Gumperz 1997).26 The situation of today’s society, however, is that of a world of rapidly expanding networks of international contacts, interconnecting increasingly complex communities from the point of view of people’s cultural origins and experiences; all this has contributed to a proliferation in issues regarding intercultural communication. As illustrated in the Wrst case study discussed in this paper, “in culturally diverse communicative environments, hidden, normally unnoticed diVerences in perspectives may bring about radically conXicting interpretations of what is happening” (Gumperz 1994:xiv), often leading to serious intercultural misunderstandings and possible reinforcement of negative stereotypes. This is particularly the case in situations of ‘high stakes’, “where the interactions are likely to be stressful and protracted and where the outcomes are critical in people’s lives” (Roberts et al. 1992:89). Carrying out detailed sociolinguistic analyses of speech events in which there are frequent contacts between diVerent cultures and between diVerent ethnic groups is therefore becoming more and more important.27 Gumperz believes that the results of such research should be made available to all those involved in second language teaching, so that speciWc cultural content can be included in language curricula and actual teaching will not be “divorced from intrasocietal issues of linguistic diversity” (Gumperz 1996b:469). By highlighting the ways in which people of diVerent cultural backgrounds interpret what they see and hear, development of awareness of this kind can become a great asset for both teachers and learners of a second language. On the one hand, teachers will beneWt from research-based input on such communicative conventions in that they will be better equipped to “incorporat[e] that understanding into the goals, curriculum design, lessons, and everyday practices of their classrooms” (SchiVrin 1996:325). On the other, awareness of cultural diVerences in communication, together with knowledge of interactive strategies and other fundamental aspects of language such as grammar and lexis, will facilitate students’ interaction with others in everyday life and prepare
John J. Gumperz: A review
them to meet “the cognitive demands of multicultural environments” (Gumperz 1996b:480).
Notes 1. This paper was originally inspired by a talk that John Gumperz gave at the 1995 IADA Round Table on Dialogue Analysis held in Bologna, Italy. It has subsequently been revised to present an up-to-date picture of Gumperz’s recent work in Interactional Sociolinguistics. 2. “Naturally organized activities” are typically face-to-face encounters possessing some structure recognizable in everyday life (for example, casual chat, job interviews, court interrogations etc.). 3. Gumperz refers to these mechanisms and patterns of conversational management as “principles of conversing” (Gumperz 1992c:304). 4. Gumperz, Aulakh and Kaltman 1982, Gumperz, Kaltman and O’Connor 1984, cited in Gumperz 1992a:232–233. 5. Cf. Tsui’s work on three-part exchanges (Tsui 1989, 1994). 6. Descriptions of this approach can be found in: Silverstein 1992; Mertz and Parmentier 1985; Lucy 1993. 7. For a slightly diVerent and more detailed version of Gumperz’s analysis of these data, see Gumperz 1992c, 1995. 8. For the construct “typiWcation”, see Schutz 1971, cited in Gumperz 1982:22. 9. For a deWnition of cultural knowledge, see Clark 1992; for a discussion of the diVerent types of knowledge constituting cultural knowledge, see Gumperz 1995:105. 10. For fuller details on Gumperz’s interpretive level of analysis, see Gumperz and Berenz 1993:92–94. 11. For the description of a case study based on this particular participant observation experience, see Gumperz 1982. 12. Interactional texts are described by Silverstein as “transcripts that account for all the communicatively signiWcant verbal and non-verbal signs perceived” (Silverstein 1992, in Gumperz 2001). For a full description of the transcribing of conversational exchanges, see Gumperz and Berenz 1993. 13. For a more detailed analysis of this job interview extract, see Gumperz 1995:110–114. 14. Accenting or stress placement is thought to be “crucial in signalling coreference and alerting interlocutors to what is expected by way of an answer” (Gumperz 1992c:323). 15. Cf. Gumperz’s (2001) analysis of a job interview with an electrician; also in this case, the interviewers’ indirect accenting strategies are not fully understood by the South Asian candidate.
99
100 Language and interaction
16. For a detailed comparison of Indian and Western English prosody, see Gumperz 1982:119–129. 17. For the description of a telephone number test devised to ascertain the occurrence of these diVerences in stress patterns and the extent to which they constitute a natural characteristic of South Asian English, see Gumperz 1982:122–123. 18. In some cases, failure by non-native speakers to recognize and/or produce English stress patterns can also be grammatically motivated, as, for example, in speakers of Hindi, who convey the idea of emphasis in part by the discourse particle ‘hi’ (Gumperz, Aulakh and Kaltman 1982; see also discussion in Roberts et al. 1992:92). 19. Cf. Tyler’s (1995) study of miscommunication between a Korean tutor and a student who is a native speaker of American English. In this case, in order not to appear rude, it is the Korean tutor who plays down his knowledge of a particular skill. 20. It is important to note, however, that these ‘criticisms’ are not directly attributable to participants’ personal feelings or beliefs, but are instead “situated reactions” provoked by “indexical ties between verbal form and context that function metapragmatically to create interpretive eVects” (Gumperz 1996c:400–401). 21. In his talk on the concept of communicative competence in culturally complex societies, Gumperz proposed a deWnition of communicative competence as “the ability to create and maintain communicative involvement” (Gumperz 1996a). 22. Cf. Gumperz 1999a for the construction of a cultural defence in a court case involving a Northern California Indian. 23. Transcription conventions are taken from Drew 1992. Gumperz has added ‘*’ to indicate syllable stress. 24. Gumperz is following the deWnition of isochrony advocated in studies on prosody as an interactive phenomenon (Couper-Kuhlen and Auer 1991, Uhmann 1996). 25. For a comparison of the concepts of communicative and interactional competence, see He and Young 1998. “Interactive practices” are deWned as “structured moments of face-to-face interaction [...] whereby individuals come together to create, articulate, and manage their collective histories via the use of sociohistorically deWned and valued resources” (Hall 1995:207– 208). 26. For example, see Gumperz’s analysis of shared inferencing between two sisters in an informal coVee shop conversation (Gumperz 2000 and 2002). 27. For example, see the studies Gumperz has carried out with others in bilingual, “cooperative learning” classrooms to investigate the peer group inferential processes students rely on to establish collaborative interaction with each other (Gumperz and Field 1995; Gumperz, Cook-Gumperz and Szymanski 1999).
John J. Gumperz: A review 101
References Brown, Penelope and Levinson, Stephen C. 1978. “Universals in language usage: Politeness phenomena”. In Questions and Politeness: Strategies in Social Interaction, E.N. Goody (ed.), 56–311. Cambridge: Cambridge University Press. Clark, Herbert H. 1992. Arenas of Language Use. Chicago: University of Chicago Press. Couper-Kuhlen, Elizabeth and Auer, Peter. 1991. “On the contextualizing function of speech rhythm in conversation”. In Levels of Linguistic Adaptation, J. Verschueren (ed.), 1–18. Amsterdam and Philadelphia: John Benjamins. Drew, Paul. 1992. “Contested evidence in courtroom cross-examination: The case of a trial for rape”. In Talk at Work: Interaction in Institutional Settings, P. Drew and J. Heritage (eds.), 470–520. New York: Cambridge University Press. GoVman, Erving. 1976. “Replies and responses”. Language in Society 5:257–313. Goodwin, Charles and Heritage, John. 1990. “Conversation analysis”. Annual Review of Anthropology 19:283–307. Gumperz, John J. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Gumperz, John J. 1990. “Conversational cooperation in social perspective”. Parasession on the Legacy of Grice. Berkeley Linguistic Society 16:429–441. Gumperz, John J. 1992a. “Contextualization and understanding”. In Rethinking Context: Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 229–252. Cambridge: Cambridge University Press. Gumperz, John J. 1992b. “Contextualization revisited”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 39–53. Amsterdam and Philadelphia: John Benjamins. Gumperz, John J. 1992c. “Interviewing in intercultural situations”. In Talk at Work: Interaction in Institutional Settings, P. Drew and J. Heritage (eds.), 302–327. New York: Cambridge University Press. Gumperz, John J. 1993. “Culture and conversational inference”. In The Role of Theory in Language Description, W.A. Foley (ed.), 193–214. Berlin and New York: Mouton de Gruyter. Gumperz, John J. 1994. “Foreword”. In Crosstalk and Culture in Sino-American Communication, L.W.L. Young, xiii-xix. Cambridge: Cambridge University Press. Gumperz, John J. 1995. “Mutual inferencing in conversation”. In Mutualities in Dialogue, I. Marková et al. (eds.), 101–123. Cambridge: Cambridge University Press. Gumperz, John J. 1996a. “Communicative competence in culturally complex societies”. Paper presented at invited colloquium on Creating Communicative Competence at 1996 Conference on Discourse Communities, Chicago: American Association for Applied Linguistics. Gumperz, John J. 1996b. “On teaching language in its sociocultural context”. In Social Interaction, Social Context, and Language. Essays in Honor of Susan Ervin-Tripp, D.I. Slobin et al. (eds.), 469–480. Hillsdale, NJ: Lawrence Erlbaum. Gumperz, John J. 1996c. “The linguistic and cultural relativity of conversational inference”. In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 374– 406. Cambridge: Cambridge University Press.
102 Language and interaction
Gumperz, John J. 1997. “On the interactional bases of speech community membership”. In Towards a Social Science of Language. Papers in Honor of William Labov, Vol. II: Social Interaction and Discourse Studies, G.R. Guy et al. (eds.), 183–203. Amsterdam and Philadelphia: John Benjamins. Gumperz, John J. 1999a. “Culture in the cultural defense”. Proceedings of the Sixth Annual Symposium about Language and Society (1998), SALSA VI:115–132. Austin: Department of Linguistics, University of Texas. Gumperz, John J. 1999b. “On interactional sociolinguistic method”. In Talk, Work and Institutional Order, S. Sarangi and C. Roberts (eds.), 453–471. Berlin and New York: Mouton de Gruyter. Gumperz, John J. 2000. “Inference”. Journal of Linguistic Anthropology 9 (1–2):131–133. Gumperz, John J. 2001. “Interactional Sociolinguistics: A personal perspective”. In The Handbook of Discourse Analysis, D. SchiVrin, D. Tannen and H.E. Hamilton (eds.), 215–228. Oxford: Basil Blackwell. Gumperz, John J. 2002. “Sharing common ground”. In Soziale Welten und Kommunikative Stile: Festschrift für Werner Kallmeyer zum 60. Geburtstag, I. Keim and W. Schütte (eds.), 47–56. Tübingen: Gunter Narr. Gumperz, John J. and Berenz, Norine. 1993. “Transcribing conversational exchanges”. In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.D. Lampert (eds.), 91–121. Hillsdale, NJ: Lawrence Erlbaum. Gumperz, John J. and Field, Margaret. 1995. “Children’s discourse and inferential practices in cooperative learning”. Discourse Processes 19:133–147. Gumperz, John J. and Roberts, Celia. 1991. “Understanding in intercultural encounters”. In The Pragmatics of Intercultural and International Communication, J. Blommaert and J. Verschueren (eds.), 51–90. Amsterdam and Philadelphia: John Benjamins. Gumperz, John J. and Toribio, Almeida J. 1999. “Codeswitching”. In The MIT Encyclopedia of the Cognitive Sciences, R.A.Wilson and F.C. Keil (eds.), 118–120. Cambridge, MA: The MIT Press. Gumperz, John J., Aulakh, Gurinder and Kaltman, Hannah. 1982. “Thematic structure and progression in discourse”. In Language and Social Identity, J.J. Gumperz (ed.), 22–56. New York: Cambridge University Press. Gumperz, John J., Cook-Gumperz, Jenny and Szymanski, Margaret. 1999. Collaborative Practices in Bilingual Cooperative Learning Classrooms. CREDE Research Report 7. Santa Cruz, CA and Washington DC: Center for Research on Education, Diversity and Excellence. Gumperz, John J., Kaltman, Hannah and O’Connor, Mary C. 1984. “Cohesion in spoken and written discourse: Ethnic style and the transition to literacy”. In Coherence in Spoken and Written Discourse [Advances in Discourse Processes XII], D. Tannen (ed.), 3–19. Norwood, NJ: Ablex. Hall, Joan K. 1995. “(Re)creating our worlds with words: A sociohistorical perspective of face-to-face interaction”. Applied Linguistics 16:206–232. He, Agnes W. and Young, Richard. 1998. “Language proWciency interviews: A discourse approach”. In Talking and Testing: Discourse Approaches to the Assessment of Oral
John J. Gumperz: A review 103
ProWciency, R. Young and A.W. He (eds.), 1–24. Amsterdam and Philadelphia: John Benjamins. Jacoby, Sally and Ochs, Elinor. 1995. “Co-construction: An introduction”. Research on Language and Social Interaction 28 (3):171–183. Lucy, John A. (ed.). 1993. ReXexive Language: Reported Speech and Metapragmatics. Cambridge: Cambridge University Press. Mertz, Elizabeth and Parmentier, Richard J. 1985. Semiotic Mediation. Orlando: Academic Press. Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This volume. Roberts, Celia and Sarangi, Srikant. 1999. “Introduction: Revisiting diVerent analytic frameworks”. In Talk, Work and Institutional Order, S. Sarangi and C. Roberts (eds.), 389–400. Berlin and New York: Mouton de Gruyter. Roberts, Celia, Davies, Evelyn and Jupp, Tom. 1992. Language and Discrimination: A Study of Communication in Multi-Ethnic Workplaces. London: Longman. SchegloV, Emanuel A., JeVerson, Gail and Sacks, Harvey. 1977. “The preference for selfcorrection in the organization of repair in conversation”. Language 53:361–382. SchiVrin, Deborah. 1994. Approaches to Discourse. Oxford: Basil Blackwell. SchiVrin, Deborah. 1996. “Interactional sociolinguistics”. In Sociolinguistics and Language Teaching, S.L. McKay and N.H. Hornberger (eds.), 307–328. Cambridge: Cambridge University Press. Schutz, Alfred. 1971. Collected Papers II: Studies in Social Theory. 3rd printing. The Hague: NijhoV. Selting, Margret. 1991. Prosodie im Gespräch. Habilitationsschrift, Universität Oldenburg. Selting, Margret. 1995. Prosodie im Gespräch. Aspekte einer Interaktionalen Phonologie der Konversation. Tübingen: Niemeyer. Silverstein, Michael. 1992. “The indeterminacy of contextualization: When is enough enough?” In The Contextualization of Language, P. Auer and A. Di Luzio (eds.), 55– 76. Amsterdam and Philadelphia: John Benjamins. Tsui, Amy B.M. 1989. “Beyond the adjacency pair”. Language in Society 18 (4):545–564. Tsui, Amy B.M. 1994. English Conversation. Oxford: Oxford University Press. Tyler, Andrea. 1995. “The coconstruction of cross-cultural miscommunication. ConXicts in perception, negotiation, and enactment of participant role and status”. Studies in Second Language Acquisition 17(2):129–152. Uhmann, Susanne. 1996. “On rhythm in everyday German conversation: Beat clashes in assessment utterances”. In Prosody in Conversation: Interactional Studies, E. CouperKuhlen and M. Selting (eds.), 303–365. Cambridge: Cambridge University Press.
104 Language and interaction
Response essay 105
Chapter 8
Response essay John J. Gumperz
Let me begin with a little more background on the issues and Wndings that Wrst motivated my research on interaction. As Levinson (this volume: 31) puts it in discussing my approach to contextualization, “Readers will not understand his work if they view it just as the study of conversation” or, for that matter, “interaction”, I would add. When I replied to Prevignano’s initial question, stating that my interests are both theoretical and applied, I meant to suggest that I did not set out to study interaction in the abstract. What perhaps I did not make clear was how my current theoretical thinking about language and communication builds directly on earlier ethnographic Weld studies that led me to conclude that I needed to learn more about interaction in order to deal with the questions I had set out to resolve. I have always had a strong interest in linguistic and cultural diversity, and in particular in questions such as: How are sociocultural boundaries represented in language? How does this relationship aVect human understanding and especially people’s ability to convey information, persuade others and carry on their everyday aVairs? What is its import for our understanding of how social institutions work? Before responding to speciWc comments, I will brieXy describe, or, more appropriately perhaps, reconstruct, how such interests led me to concentrate on a theory of interaction. When I Wrst embarked on Weld research, the nineteenth century assumption that human populations come naturally divided into distinct, bounded groups was still widely accepted. It was assumed that each of such units can be treated as an island unto itself, with its own language or dialect and culture, which must be analyzed in their own terms. Communication across group boundaries was said to be problematic, and linguistic diversity was seen as potentially a major obstacle to societal consensus. Yet, apart from abstract discussions about translation as the process of
106 Language and interaction
converting the morpho-syntactic or semantic structures of one language into another, little in the way of research had been done to clarify what linguistic and cultural diversity involved at the level of everyday communication. My aim was to collect veriWable data on both linguistic and socio-cultural representations of diversity. I was familiar with the research of linguistic geographers in Europe and the United States who had convincingly demonstrated that current linguistic boundaries had arisen in the course of the preceding centuries as a function of the shifts in power relations, as well as of the related political, economic and religious forces which together led to the development of modern nation states (BloomWeld 1933). The new urban formations produced by these changes in eVect served to concentrate previously dispersed sources of inXuence into centralized, politically, economically and culturally powerful centers of innovation, which then radiated pressure towards uniformity into the surrounding hinterland. In this way, ecological conditions arose that led to the redirection of interpersonal relations inwards, towards the dominant center. The resulting communicative environments brought about a reduction of local diversity. Nationalist and religious ideologies, along with their associated new national language varieties, circulated within these zones of inXuence, while, at the same time, boundaries between separate zones were sharpened, further accelerating the homogenization process. In other words, present-day linguistic boundaries cannot be naturalized and accepted as preexisting facts. They basically reXect the human history of the past few centuries and the forces that led to the current situation, which, in turn, we must study. Reviewing these developments in his classic articles on the speech community, Leonard BloomWeld (1933) argued that the degree of linguistic diversity in any one region is a function of the patterning of interpersonal contacts over time. His formulation suggests an initial outline of a theory of diversity that rests on human interaction as an analytical prime, and does not rely on a priori assumptions about ethnic, class or group identity. Such a theory, among other things, provides explanations for how language distinctions like those between Dutch and German, French and Italian, Serbian and Croatian, and many others throughout the world, came about in the course of history, as a result of constraints on human contact imposed by changing power relationships. To be sure, the linguistic geographers’ research on which these arguments were based had been conWned to populations speaking histo-
Response essay 107
rically-related language varieties. But since the theory focusses on interpersonal relationships and their long-term historical eVects, not on how they are represented in speciWc languages, there seemed to be no reason why similar reasoning should not be extended to multilingual situations, provided appropriate social conditions existed. In my early Weldwork, I set out to explore the validity of the interactional approach to diversity, Wrst in a caste-stratiWed North Indian village and then in a small, culturally homogeneous, Northern Norwegian town with a strong egalitarian ideology (Gumperz 1971). In the North Indian village, for example, neither the divisions among named castes, nor the politically salient religious distinctions between Hindus and Muslims were directly reXected in linguistic signs. I did, however, discover signiWcant bundles of linguistic isoglosses. One set of features divided the majority of local touchable agriculturist and artisan castes who shared a single speech variety from the three untouchable castes. Each of the latter, untouchable groups in turn was set oV by additional sets of isoglosses from the other two. When I re-examined the linguistic data in the light of ethnographic information on interpersonal relations, it became apparent that touchable village residents formed a single communicative whole, where intercaste contacts were relatively frequent and children freely entered into mixed-caste neighborhood playgroups. Untouchables, on the other hand, lived in separate neighborhoods and did not maintain informal peer relations with other castes. Based on such Wndings, I hypothesized that interaction is most probably the determining factor in the construction and maintenance of the intra-village linguistic distinctions. As with the linguistic geographers’ isoglosses, the relevant signs were grammatically not very signiWcant. Nevertheless, they carried great social import. However, contrary to what my reading of the literature had led me to expect, frequency of interactive contacts alone could not explain the facts on the ground. Untouchable caste members tended to spend most of their days working in the households of their touchable employers and had only limited opportunities for socializing with peers. Yet, they did not adopt their more prestigious employers’ speaking practices, nor for that matter did they give up their own distinct manners and dress. To explain this, I argued that, apart from density of contact, ideologies of interpersonal relations and normative principles constraining what could be said to whom and how it should be said must also be at work. In other words, at issue is not just the number of times individuals speak to others. The quality of the interaction, that is to say, what
108 Language and interaction
is said, when, and by whom in the course of such encounters, seemed to count as an equally important factor. The Norwegian study yielded additional conWrmation for the interaction hypothesis (Blom and Gumperz 1972). The ideology of egalitarianism here was so strong that residents refused to acknowledge the existence of signiWcant class diVerences in the locality. And indeed it soon became evident that income and occupational status distinctions were not clearly reXected in talk. Nevertheless, it was possible to detect signiWcant diVerences, this time not at the level of phonology or syntax, but in patterns of language use and particularly in code-switching between the local dialect and the regionally dominant variety of Standard Norwegian. The diVerences in language use distinguished those who, by virtue of their occupation, maintained regular and close relationships with speakers in the surrounding small towns and in the more highly urbanized South of the country, from other residents whose signiWcant contacts were conWned largely to local ones. Once more, density and quality of interpersonal contacts turned out to be more signiWcant than social categorizations. Furthermore, as Wndings from research in previously little-known linguistic areas became better understood, it also became evident that interaction-based forces of linguistic convergence and diVerentiation do not necessarily respect language boundaries. Throughout the known world, both in large nation states, as well as in less urbanized regions, where small, separate, localized language/culture units live among majority populations, regular and frequent translocal, interpersonal contacts set oV pressure for change. Over time, these pressures tend to blur and ultimately eliminate established lexico-grammatical distinctions (Gumperz 1996; Gomez-Imbert 1996). A third ethnographic study of a multilingual village at the Indo-Aryan/ Dravidian language border largely supports this interpretation. In this Central Indian site, the dominant group of landowners spoke a dialect of Kannada, a Dravidian language, while Indo-Aryan Marathi was the oYcial administrative language in the surrounding region and the medium of schooling. A signiWcant minority population of artisan and untouchable castes, moreover, spoke dialects of Urdu (Indo-Aryan) or Telugu (Dravidian). We were struck by the frequent code-switching observed in everyday village talk. For example, just as the small town residents in Norway had alternately used Standard Norwegian and local dialect forms, people in this multilingual village readily switched between the historically-unrelated Marathi or Urdu and Kannada,
Response essay 109
often in the course of a single brief encounter. And they did so for similar communicative ends. When we compared transcripts of everyday talk, moreover, we found that texts in the local dialects of Urdu and Kannada could be described by a single underlying set of surface syntactic rules. What local peoples referred to as two separate languages were analytically distinct, for the most part, only at the level of word morphology and lexicon, so that translation from one local variety to the other was largely a matter of word-for-word substitution. In other words, for the speakers involved in these situations, code-switching was grammatically not more onerous than dialect-switching in the Norwegian case. Apparently, centuries of co-residence in the same locality had brought about large-scale structural convergence among local speech varieties. Yet, the changes had stopped short of complete creolization. If comprehensibility is not an issue, how is it that such structurally homologous varieties maintain themselves in this relatively small population in the face of long-term close contact? As the code-switching evidence shows, it is not because people cannot understand each other. We are dealing with a socially-motivated linguistic ideology of diVerentiation, a set of local norms similar to those found in the North Indian situation, which keep speakers from taking over the others’ modes of speaking. The picture of variability that emerges from these Weld studies is quite diVerent from the conventional one. Common practices that accept takenfor-granted notions of community, explaining language usage in relation to a presumably uniform, community-wide system of beliefs and values, that is, treating each as a self-contained island, are seriously in need of revision. Whereas truly isolated, homogenous communities have by now all but disappeared, as the pace of population migration continues to grow, multilingual or multidialectal communicative economies like those I have sketched above are rapidly becoming the rule rather than the exception. It would be more fruitful therefore to turn to a new base metaphor for sociolinguistic research, a metaphor that incorporates diversity into the research design. One might for example focus on a set of local formations marked by translocal economic, political, religious and other similar networks of relationships. All this is not to claim that communities do not have a basis in linguistic reality. I am arguing that communities must be seen as ideological constructs. Whereas communicative practices rely on common ground, that is, shared assumptions to achieve understanding, ideologies are by their very nature contested.
110 Language and interaction
Communities on one deWnition are treated as the oYcial units for much governmental and educational planning, but how these communities are deWned is always open to question, and cannot be taken for granted. Ultimately, communicative practices that are part of a daily reality may have an inXuence on governmental decisions. I am arguing that the relationships between ideological constructs and practices are the most useful point of entry for an investigation of how societal level forces can aVect and be aVected by everyday talk. In my earlier work, I used the term “linguistic repertoire” to refer to the speaking practices of social collectivities, arguing, as di Luzio (this volume: 2– 3) points out, that repertoires are describable in terms of sets of functionallydiversiWed speech varieties. I found the notion of repertoire useful because it served to call attention both to diVerences among constituent varieties and to commonalities due to convergence. To show, in other words, that we are dealing with what, for want of a better term, we can call “communicative economies”, that have their own local and translocal forms of social and linguistic organization. Recent, more detailed work on everyday communication, however, has led me to doubt that it is useful to assume that discourse level diversity can be explained in terms of alternation among sharplybounded speech varieties. I will return to this point later. The ethnographic insights that initially seemed most in need of explanation were the following. First, the surprising lack of correspondence between people’s perceptions of boundaries and the linguistic nature of the boundary markers. In none of the three Weld situations did we Wnd a direct one-to-one relationship between what I originally termed “socially”, but would now call “ideologically” grounded language/dialect distinctions, on the one hand, and units of language, deWned by grammatical distance alone, on the other. Both in multilingual and in multidialectal settings, what from a purely linguistic perspective may count as minor distinctions can often, for largely ideological reasons, attain great social import as badges of identity. If structure alone cannot explain such communicative facts, how do we look at verbal interaction to Wnd explanations? Secondly, people readily communicated across linguistic boundaries. What locals refer to as language or dialect distinctions do not by themselves prevent understanding, or as such limit interaction. In the early sites, as well as in later European and U.S. Weld work, local speech varieties formed part of a single communicative economy, where, depending on circumstances, switching among distinct languages, dialects or even styles
Response essay
of the same language can produce communicatively equivalent eVects. We are dealing with socially-motivated ideologies of diVerentiation, localized norms akin to those in the North Indian situation, that in some cases lead to convergence, and in other conditions keep speakers from taking over their interlocutors’ speaking habits. Linguistic and cultural boundaries are not just ‘naturally’ there, they are communicatively and, therefore, socially constructed. Thus, they cannot be essentialized and treated as self-contained islands in research on communicative practices. Apart from interaction as such, ideology, power and history are all central to the way diversity works; depending on how these factors interrelate in speciWc circumstances, interaction can serve either to accentuate or attenuate the eVects of diversity. I therefore decided that I needed to learn more about how and through what kind of linguistic knowledge people understand each other in everyday talk, as well as how this knowledge is acquired and how it circulates, before attempting to go on to further generalizations. Ethnographers of communication were the Wrst to turn away from community-based sociolinguistic analysis. They argued that, while grammatical analysis has advanced greatly in depth and sophistication since the nineteen sixties, we have little if any comparative information about everyday communication (Gumperz and Hymes 1972). The problem was not just that more of the world’s many languages and cultures were in need of detailed study. Rather, our sociolinguistic theories were unable to deal with many signiWcant facts of language use in today’s inherently diverse societies. It seemed important, therefore, to bypass established structural-functionalist research designs and Wnd ways of developing integrated semiotic approaches that would focus directly on speciWc events as the sites where language and culture meet. In my work on interaction, I built on my experience with ethnography of communication, relying on in-depth participant observation and interviewing to discover a range of naturally-occurring communicative situations or events. I concentrated on events that typically involved talk on issues of general interest, in settings where either my collaborators or I were suYciently familiar with the relevant cultural and historical facts. As I see it, ethnography is not simply a matter of investigators creating written accounts of their experiences; it is basically a reXexive hypothesis formation, hypothesis testing and reformulation process. Building on background knowledge acquired in this way, I embarked on comparative case study analyses, contrasting in-
111
112
Language and interaction
stances of “intercultural communication” with other situations where participants were able to rely on shared communicative experience to determine the relevant linguistic signs. The procedure is akin to those employed in linguistic Weld work, where grammatical and ungrammatical sentences are compared to gain understanding of grammatical processes. Initial insights into the semiotic/cognitive aspects of diversity came from code-switching. Both in the Norwegian and the Central Indian village, many speakers typically spoke both a local variety and a standard variety, alternating between the two in response to the communicative demands of the exchange in which they were engaged. In the Norwegian case, where we Wrst looked at the phenomenon in some detail, we discovered two distinct patterns. In the Wrst, situational switching, code alternation, co-occurs with or signals shifts in local, participant alignments or topic change (Blom and Gumperz 1972; Gumperz 1982). But there is a second pattern, where words or phrases of a second language are inserted into what at the level of discourse must count as the same conversational string. Using as a basis an analysis of preceding and following talk, I hypothesized that the juxtaposition here counts as a momentary allusion to, or evocation of, a diVerent set of contextual presuppositions, which has communicative import and aVects the listener’s interpretation. I initially used the term “metaphoric switching” to suggest that the communicative eVect of these switches is similar to metaphors in their rhetorical eVect. A code-switch of the latter kind can then be seen as a communicative strategy, an attempt to inXuence the listener’s interpretation of what the speaker intends to convey. My search for similar discursive phenomena in a range of situations, and attempts to explain their semiotic functioning, led to the approach to conversational inference I outlined in the discussion with Prevignano and di Luzio. The notion of contextualization cues was developed in the course of the work on intercultural communication as a way of generalizing insights gained through code-switching analysis (Gumperz 1982). Note that, as Levinson (this volume: 36) points out, contextualization cues constitute a natural class of signs that cuts across generally recognized channels of linguistic signaling. They function as co-occurring clusters of signs that aVect interpretation by evoking the contextual presuppositions in terms of which meaning is assessed. This suggests that we need to deal with understanding at a level of discourse organization which is distinct from more commonly studied lexico-semantic signaling processes. My use of the term “discursive practice” is akin to Hanks’ use of the term in his Language and Communicative
Response essay
Practices (1996), which presents detailed arguments to support the need for the above distinction. My approach to semiotic phenomena is broadly in line with Michael Silverstein’s Peircian semiotic functionalism. Silverstein argues, among other things, that contextualization cues are a class of indexicals (1992); that is to say, they function via vector-like cognitive processes radiating out in certain directions from a known origin, but have no deWnite terminus (Silverstein 1993). Yet, in spite of its inherent indeterminacy, indexicality, as most of today’s linguistic anthropologists will agree, is crucial for anyone concerned with discursive representations of culture. Among other things, indexicality suggests an explanation for what I referred to above as the lack of Wt between grammatical distance and ideologically-based language distinctions. What indexes do is act as Xags, cues or reminders to listeners to search their memory for possible alternative ways of explaining or framing what they hear or otherwise perceive or recall. That is to say, to arrive at explanations that make coherent sense of what is going on in the situation at hand. There is thus no generally agreed-upon, stable meaning relationship between the indexical sign and any speciWc explanation. The interpretations are always highly context-speciWc in that they depend on aspects of the communicative process that resist strict formalization. A key aspect of Peircian theory, not suYciently foregrounded in the currently available literature, is that signs have meaning only by virtue of being taken to stand for an object by some interpreter (Lucy 1993). This lays the groundwork for an integrated semiotic framework, capable of dealing with both the linguistic and socio-cultural aspects of everyday communication within a single analytical perspective. Saussurean denotational meanings must ultimately always be seen as what they are, abstracted from situated practice in accordance with procedures that set aside the very signaling processes on which our assessment of the situated interpretation of communicative intent rests. Useful as they are for comparative grammatical analyses and historical reconstruction, denotational meanings cannot by themselves explain everyday human action, partly because they do not account for discourse-level signaling processes. Once we look at verbal communication as relying on both symbolic and indexical signs, we can see that human interpreters become central to the interpretive process. That is to say, the interpreters’ communicative background, their ability to retrieve background knowledge while engaged in an interaction, the power relations aVecting access to relevant
113
114
Language and interaction
knowledge, and the ideological social and cultural constraints that aVect understanding, all become integral constituents of the communicative process. Although I have not explicitly said so before, I basically share this perspective in my attempt to develop ways of dealing with interaction that account for universal, as well as culturally and situationally-speciWc, aspects of discourse. Turning now to the speciWc comments, Levinson, Eerdmans and, to some extent, Prevignano, all expand on my responses in the discussion. Levinson explores the notion of contextualization from a more formal, that is, more linguistically-oriented, pragmatic perspective. His discussion of “context importation and context invocation” is useful and illuminating in that it cites a number of examples from the linguistic literature to suggest that the notion of contextualization cue is not as far removed as one might think from more familiar linguistic pragmatic processes. One might say that contextualization applies a pragmatic perspective to discourse analysis. But recall that, especially in my current work, I focus on events or situations of speaking where participants are treated as engaging in discursive practices, rather than speaking particular languages. Participants in an event engage in communicative practices, where verbal expressions, indexical signs like code-switching, prosody or non-verbal signs, and other signaling modes may all achieve similar communicative eVects. It is this notion of discursive or communicative practice, which, in accord with Hanks (1996), I take to be distinct from language as such, that is relevant here. It enables us to achieve the broader, more inclusive level of semiotic analysis, not tied to the speciWcs of particular linguistic signs, which we need to treat code-switching, nonverbal signs, dress and other signaling modes as contributing to the interpretation process. Along with Hanks, GoVman, conversational analysts and others, I think of discursive practice as a separate level of analysis and would argue that Levinson, illuminating as his points are, takes a more languageoriented perspective; while, from this perspective, the distinction between diVerent kinds of contextualization makes sense, it is less important at the level of communicative practice. Finally, as to what Levinson (this volume: 38) calls the “ineVability” of understanding, I agree that my answers to Prevignano’s “persistent” questions as to how to deal with the ambiguity that arises for the analyst as a result of the complexities of the interpretive process seemed somewhat evasive. Let me try to be more speciWc and rephrase Levinson’s second question (this volume: 38)
Response essay
as follows: What gives the analyst the right to say, “B perceives that A said this, but he thinks A intended to convey that”? As pointed out before, I do not claim that Interactional Sociolinguistic (hereafter, IS) analysis can provide deWnite answers to such questions. Since situated understanding ultimately always builds on indexical signs, we cannot avoid ambiguity. But then, as Prevignano (this volume: 17) points out, even properly socialized natives can understand diVerent things by the same utterances. In fact, we all know political discourse, advertising, committee negotiations and similar genres rely on ambiguities for rhetorical eVect (Cook-Gumperz and Gumperz 1994). Interpretation, moreover, is not a unitary assessment. Typically, inferences are interactively negotiated over several turns at speaking and such negotiations rarely yield uncontestable understandings. As I see it, the IS analyst’s task is to reconstruct the understanding process, relying on a varied array of formally distinct sources of knowledge both at the level of content and form. Let me turn once more to examples from my published writings to illustrate how I view the interpretive process: (i) in the cross-examination exchange described by Eerdmans (this volume:94–97), we understand what is intended by virtue of the fact that we can check our local inferences against widely-known background presuppositions about what goes on in a crossexamination and what we can expect the individual participants to have intended to achieve by means of their talk; (ii) take the case of the graduate student from Discourse Strategies which I referred to in the discussion (Prevignano and di Luzio, this volume:19–20), who responded, “Ahma git me a gig,” when I asked him to see me in my oYce. I was at Wrst puzzled by what seemed like an unexpectedly jarring code-switch to Black English, which did not Wt into my initial understanding of what had transpired. It was only when I compared my interpretation of the incident with comments of other black, as well as white, students, that I arrived at a reasonable interpretation. An interpretation, moreover, that Wts in quite well with what I later learned about African American communicative practices in my work with raciallymixed school populations; (iii) in a conversational exchange among American Asian graduate students (Gumperz 1992b:48–9), A asks B what his brother’s major in college is. When B answers “ah, bio*chem?”, A replies “ ...... *da::(ng), .. the *bio, huh?”. And when B conWrms, he goes on with “we Asians man! I’m telling you.” Although when I Wrst looked at the exchange, I did not quite understand what the talk was about, it was clear from the way the two participants responded to each other and jointly went about developing a shared
115
116
Language and interaction
theme that they understood each other. A few turns later in the exchange, it appeared that what the group was concerned with was the failure of Asian professionals to take leadership roles in public aVairs or in other technical Welds; (4) in the Wnal example (Gumperz 1996:388V), an older South Asian student (D) came to see a lecturer (L) in an adult education center to ask about enrolling in a new course in a nearby community college. When the lecturer told him that, although she had participated in planning the course, she was not involved with student admission, he directly contradicted her, arguing that she did have a role in the selection process. He kept on with this for several minutes, totally contradicting nearly all of her statements, and seemingly oblivious of the fact that she was becoming more and more annoyed. Taken at its face value, D’s behavior seemed at Wrst quite inexplicable both to me and to the two native South Asian assistants who had helped with the transcription. It was only when we stepped back from the details of the exchange, trying to imagine what someone like D might have been seeking to accomplish with his repeated denials, asking ourselves why he would want to oVend someone he was asking help from, that we came up with a likely solution. The assistants pointed out that D sounded as if he was pleading. And as they said this, I was reminded of an encounter with an older man in a North Indian village who asked me to help his son, a college student, obtain an assistantship in the United States. When I tried to tell him that, as a junior professor, I did not have the right connections, he replied “you can do everything,” using an expression remarkably similar, i.e. intertextually-related, to some of the statements D employed. What is analyzed here are events made up of a set of utterances or speech exchanges and the surrounding circumstances relevant to their interpretation. The aim is to arrive at interpretations that a) account for what transpires in the event as a whole and b) Wt into what we know independently about local ideologies of language and interpersonal relations. The discussion highlights the functioning of shared background knowledge or common ground in the interpretive process. Ethnographic observation is clearly important as the means by which the analyst acquires the knowledge of background presuppositions and of the indexical conventions that participants tend to rely on in deriving what they assume is intended, i.e. Levinson’s ‘that’, from what they perceive as ‘this’. But what I am referring to here is a type of ethnography of communication that focusses on what people do in speciWc types of situations, how they react to others and how they talk about it. Along with other
Response essay
interpretivists, I rely on triangulation, comparing my own with others’ interpretations or examining what happens at one point in the exchange to determine what coheres with or Wts into what went on before or after. Moreover, whenever possible, I look for intertextuality to suggest possible interpretations. In contrast to other interpretive traditions that focus only on content, however, IS analysis then seeks to reWne such possible interpretations in the light of what is known about the indexical value of the contextualization cues used, the utterance’s positioning within the stream of talk and the responses it receives. Interpretation, when seen in the IS perspective, integrates the socalled formal and content-level aspects of discourse. All such factors together act as a check against misunderstandings, so that in a large number of cases we can be fairly conWdent of our interpretations. I believe that interactional sociolinguistic analysis can provide valuable insights into the culturally-speciWc rhetorical traditions that underlie our reasoning, into how understandings come about and into what it is about interaction that can lead to misunderstandings. In her contribution, Eerdmans writes from the perspective of an applied linguist interested in analytical methods. As she points out, interactional sociolinguistic analysis relies on sequentially-organized conversational exchanges like those of conversational analysis, as well as on Gricean notions like implicature and inference. I should add that this combination of methodological approaches does not just constitute eclecticism on my part. My notion of inference informally builds on Grice in that I assume that what is at issue is “what a speaker intends the listener to do with the message,” so that communication is always goal-oriented. Furthermore, I believe that while listeners’ interpretations, as they emerge in discourse, are often quite far removed from denotational meanings, they can always be derived from what was literally said. Basically, I regard myself as a linguistic anthropologist for whom the study of interaction is integral to his more comprehensive ethnographic investigations of the often taken-for-granted ways in which local populations deal with issues they encounter in going about their aVairs. That said, it is true I share the conversation analysts’ position that talk is constituted by sequentially-organized conversational exchanges, and that conversation in a way creates its own communicative ecology. But consider my deWnition of conversational inference: “the situated or context-bound process of interpretation, by means of which participants in an exchange assess [each] others’ intentions
117
118
Language and interaction
[at any one point in an interaction], and on which they base their responses” (Gumperz 1982:153). Given this deWnition, we cannot take understanding for granted and must assume that the relation between adjacent turns is always mediated by an evaluative assessment as a precondition for the production of a succeeding turn. How constituent moves are assessed by participants is crucial to the production of conversational exchanges. The object of IS analysis is thus not a ready-made sequence of turns at speaking. From an interactional sociolinguistic perspective, exchanges are best represented as sequences such as the following: ‘move … evaluative assessment … countermove … evaluative assessment … further move’, and so on. That is to say, any exchange depends on participants’ interpretive acts at any one of the many transition points in the exchange. Apart from the usual semantic processes, moreover, ideology, which in turn closely reXects societal power relationships, also enters into the assessments. In all these matters, IS analysis clearly diVers from conversation analysis. Eerdmans’ two examples (this volume: 90V) illustrate these matters. In the second case, where background assumptions are shared, each of the two principals can rely on shared knowledge of indexical conventions to indirectly convey conXicting accounts of what transpired. In the Wrst example, the job center interview, the two participants diVer in background knowledge and are unsuccessful in establishing a productive exchange, with the result that the interviewer is unable to make a fair assessment of the applicant’s abilities. Because of the prevailing ideology of minorization, and what are in fact diVerences in background knowledge, the applicant’s abilities are disvalued and his career prospects damaged. This one exchange clearly reveals how macro-societal stereotypes can aVect our assessments of what transpires in such encounters. Ballim’s brief comments are helpful because they point out speciWc areas of application where the interactional sociolinguistic approach may be useful. The comments highlight areas of interest where IS and Natural Language Processing converge. Among other things, the two share an interest in situated interpretation as it aVects participants’ perception of context. But as an anthropologist, I am inclined to place more emphasis on shared traditions and how they aVect speaking practices. For instance, building on Wndings from work on intercultural communication, I have tried to isolate shared interpretive principles or maxims demonstrably grounded in the typiWcations that arise from participation in speciWc networks of relationships. Such maxims
Response essay
must of course ultimately be shown to be veriWed in the conduct of interpersonal relations. With respect to indirect reasoning, IS analysis could serve as an initial empirical heuristic that could test a priori assumptions about the kinds of reasoning processes children would assume to be shared and what they would assume to be explicitly lexicalized. This is a distinction that has been useful in my own work on classroom talk, where I often had to deal with diYculties in teacher-student communication. As to the other topics Ballim brings up, inference, as I use the term, clearly involves indirect reasoning and the distinction between lexicalized and implicitly conveyed matters is basic. I often use Harold GarWnkel’s term “takenfor-granted knowledge” to refer to such matters. In the studies of classroom teaching processes, brieXy referred to at the beginning of the discussion, English teachers, as well as their ten-year-old students, all talked about grammatical issues that came up in reading comprehension. But whereas the teachers relied on explicitly lexicalized grammatical terminology to make their point, students, when talking, dealt with grammar almost entirely in indirect terms. They had diYculty with the teachers’ grammatical terminology (Gumperz and Field 1995; Gumperz, Cook-Gumperz and Szymanski 1999). I am grateful to Ballim for the way he characterizes my position on community. My main point is, as he puts it, that “it is not suYcient to consider discourse in terms of a uniform, disembodied community.” What we must account for is individual action as reXected in practice, but I Wnd the term “rhetorical tradition” more useful in accounting for the fact that whatever we know or believe is acquired through participation in some collectivity. I have used the term “event” in two senses. First, historically, in describing the ethnographers of communication’s shift away from community-wide ethnography. Secondly, as a heuristic in IS Weldwork on communicative practice. With respect to context, psychologists, cognitive scientists, and many linguists who pay attention to context tend to deWne it almost entirely in extracommunicative terms. I argue that, while these factors are, of course, signiWcant, contextual information is imported into the interpretive process primarily via indexical contextualization cues, in the form of presuppositions of what the activity is and what is communicatively intended. More comparative work on inferential processes needs to be done to test out what are at the moment very tentative generalizations. The little comparative work I have done so far indicates that the indexical and inferential processes readily circu-
119
120 Language and interaction
late across language boundaries, which is another argument for distinguishing between language and communicative practice. But how they circulate is still far from clear. Finally, on the matter of what Prevignano in our discussion (this volume: 21V), and later Ballim in his commentary (this volume: 83) have called “corpus pragmatics”: one of the obstacles to valid corpus-based comparative research is the lack of agreement on what is analytically signiWcant data. Particularly in pragmatics, we Wnd wide disagreements on what constitutes valid data and how it needs to be transcribed and formatted for further analysis. In my recent work, I have used Michael Silverstein’s notion of “interactional text” (1993), which my co-author and I deWne as a transcription that sets down on paper “all those perceptual cues that past research and ongoing analyses show participants rely on in their online processing of conversational management signs” (Gumperz and Berenz 1993:92).1 The current discussions of “entextualization” as the process by which mental processes are translated into words should also be relevant here (Silverstein and Urban 1996). Thibault, in his contribution, critically examines the IS approach to communicative practices, comparing it to an alternative which, like the IS approach, assumes multimodality of signaling processes, but where communicative practice is analyzed with reference to a “community’s resource systems.” Seen from my perspective, locating shared linguistic resources in “a community” raises serious problems. To speak of a community’s resource systems implies a deWnition of what is meant by the term. Currently, uses vary from the traditional treatments of community as an abstract, disembodied, uniform whole, to notions like “communities of practice” (Lave and Wenger 1991), that is, sets of individuals regularly cooperating in the accomplishment of some set of tasks. For example, cognitive psychologist Herbert Clark (1996) deWnes community as any set of individuals who share some form of expertise: residential groups, professional groups, voluntary organizations and the like. It is unclear, apart from lexical labels, how these deWnitions are reXected in language. Traditional deWnitions of community, on the other hand, are, as we all know, becoming more and more controversial in today’s increasingly diverse communicative environments. In sum, I am far from clear about how the concept of community can be integrated into an analysis of communicative practices. My position is that more comparative research on the distribu-
Response essay
tion of shared linguistic and social knowledge is needed before we can resolve this issue (see also Gumperz 1997). Towards the beginning of his discussion of context and context-independent meanings, Thibault refers to some of my statements that he interprets as contrasting “pure indexicals with context-independent or propositional meaning.” Perhaps my use of the term “propositional” and “context-free” is confusing, but that is not what I intended to say. I agree with Thibault that all naturally-occurring semiotic forms are indexical. To the extent that we can talk about propositional content and context-independent meanings, we refer to forms or interpretive assessments that have been extracted from some context by analysts for comparative or other analytical purposes. I should have made that clearer. But other sections of Thibault’s comments reXect fundamental diVerences in basic assumptions with respect to those on which IS rests. For example, he argues that communicative practices have to be studied in relation to a community’s resource systems. These are described in terms of arrays of semiotic modalities, including grammar, prosody, non-verbal signs etc., in fact pretty much the same types of signs that IS treats as entering into the inferential process. But whereas for IS the shared inferencing of human actors is the locus of interpretive assessment and ultimately understanding (or “meaning-making”, in Thibault’s terms), Thibault argues that meaning construal is achieved by a conjunction of mutually contextualizing modalities. In other words, meaning, on this view, resides in language, whereas IS follows the linguistic anthropological tradition in arguing that it is humans who construe meaning (Lucy 1993). Both approaches have overcome the limitations of traditional approaches that see only denotation as properly linguistic. But beyond that, I have problems with Thibault’s position. How does he deWne a community in today’s conditions of ever-increasing diversity? How does he account for the human factor in interpretation, particularly its ideological character and limited access to knowledge? Through IS analysis it is possible, for example, to gain insights into how communicative barriers arise; what it is that brings about their disappearance; how diversity can be studied via communicative practices in ways that do not rely on a priori assumptions about group or community membership; how interpretation reXects past communicative experience. I believe that I have reason to claim that IS has, at the least, made signiWcant steps towards the solutions of the problems I set myself and I am not sure how Thibault would deal with such issues. But then
121
122 Language and interaction
the two approaches reXect diVerent interests and cannot be judged by their inability to deal with matters they were not designed to deal with. Elsewhere in this volume, I have already partially answered Prevignano’s questions about what makes interaction possible and about the validity of inference-based interpretive assessments. Let me brieXy recap: I see interaction as a turn-by-turn process, where the production of each turn, including the opening move, is preceded by an assessment that is simultaneously pragmatic and meta-pragmatic. Interaction always presupposes conversational involvement, so that the maintenance of involvement depends, among other things, on the extent to which knowledge of linguistic, that is, symbolic and indexical signs, and underlying background knowledge are shared. Both interactants and analysts are thus always faced with ambiguities. But the resulting interpretive diYculties seem less problematic if, instead of looking at interpretation as a one-shot assessment, we think of it as an ongoing interactive process where inferences are constantly revised and attuned as circumstances change. As to the question of minims and maxims, Prevignano (this volume: 63), if I interpret him correctly, is comparing two scholarly positions, the Gricean tradition, which postulates universal conversational principles or maxims at the level of content, and a second one that he associates with Gumperz. The latter argues that interpretive assessment also depends on context-speciWc aspects of surface discursive form that do not ordinarily enter into other types of interpretive analysis. I Wnd Prevignano’s discussion interesting because he seems to be attempting to deal with issues of appropriate speech that many pragmaticists and sociolinguists tend to account for by means of the much criticized normative rules of appropriateness, of the form: “In situation A, do or use Y”, where the term “situation” refers to extra-communicatively available sets of circumstances. The two traditions he alludes to both take a basically semiotic position on what makes for rhetorically eVective talk, attempting to capture regularities of discursive practice without reifying them in terms of a priori rules. The two diVer in that whereas the Wrst tradition concentrates on content, the second argues that interpretation rests on participants’ ability to relate form to content both to decode propositional meaning and to derive contextually-grounded hypotheses about presuppositions and possible entailments that aVect understanding. My insights into the communicative import of linguistic form derive from work in dialectology and intercultural communication, where what from a structural perspective
Response essay 123
seemed like insigniWcant phonetic and prosodic processes, can have great communicative import. I believe that to account for the facts of discourselevel interpretive assessments, we need to treat discursive or communicative practice as a separate level of analysis. Prevignano brings up the issue of metacommunication in discussing the validity of analysts’ models or descriptions. Let me point out, however, that one of the main reasons linguistic anthropologists, building on Voloshinov’s discussion in Marxism and the Philosophy of Language (1973), have begun to concentrate on meta-communicative phenomena is to avoid their predecessors’ reliance on analytically unjustiWed, dichotomous distinctions like ‘language and thought’, ‘language and culture’, and the like. Linguistic anthropologists have turned to metacommunication as a strategy for avoiding these dichotomies. In so doing, they open up new possibilities as well as empirical methods for dealing with these issues. The reason for these new approaches is the realization that the way people talk about talk provides at least an indirect indication of what people might think about what they hear. In other words, metacommunicative phenomena provide us with benchmarks for evaluating analysts’ interpretations. The distinctions between explicitly metacommunicative signs and indirect indexicals such as contextualization cues is, analytically, of great importance. The fact that the former are readily subject to conscious manipulation, whereas the latter are automatically produced below the level of consciousness, opens up interesting possibilities for interpretive analysis, as several recent articles suggest (Gumperz 1999; Cook-Gumperz and Gumperz 1996). Finally, as to my current interests, at the moment I am working on a book to be entitled Language in Social Theory, which reexamines Wndings from my Weldwork over the years in the light of recent theoretical developments, expanding, among other things, on the ideas expressed here. In addition, I am about to Wnish a case study, Communicative Competence as Communicative Practice, which reviews the notion of communicative competence in the light of what we have learned about discourse since the seventies. I am also cooperating with Marco Jacquemet on an edited volume of ethnographically-based articles that will pick up where Gumperz and Hymes’ Ethnography of Communication (1972) left oV.
124 Language and interaction
Note 1. See also the article by Du Bois et al. in the same volume (1993), and the more recent, more comprehensive, “Gesprächsanalytisches Transkriptionssystem (GAT)” (Selting et al., 1998).
References Ballim, Afzal. “A commentary on a discussion with John J. Gumperz”. This volume. Blom, Jan-Petter and Gumperz, John J. 1972. “Social meaning in linguistic structure: Code-switching in Norway”. In Directions in Sociolinguistics: The Ethnography of Communication, J. J. Gumperz and D. Hymes (eds.), 407–434. New York: Holt, Reinhart and Winston. BloomWeld, Leonard. 1933. Language. New York: Holt, Rinehart and Winston. Clark, Herbert H. 1996. “Communities, commonalities and communication”. In Rethinking Linguistic Relativity [Studies in the Social and Cultural Foundations of Language 17], J.J. Gumperz and S.C. Levinson (eds.), 324–355. Cambridge: Cambridge University Press. Cook-Gumperz, Jenny and Gumperz, John J. 1994. “The politics of a conversation: Conversational inference in discussion”. In What’s Going on Here? Complementary Studies of Professional Talk, A.D. Grimshaw et al. (eds.), 373–397. Norwood, NJ: Ablex. Cook-Gumperz, Jenny and Gumperz, John J. 1996. “Treacherous words: Gender and power in academic assessment”. Folia Linguistica XXX (3–4):167–188. di Luzio, Aldo. “Presenting John J. Gumperz”. This volume. Du Bois, John W. et al. 1993. “Outline of discourse transcription”. In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.D. Lampert (eds.), 45–89. Hillsdale, NJ: Lawrence Erlbaum. Edwards, Jane A. and Lampert, Martin D. (eds.). 1993. Talking Data: Transcription and Coding in Discourse Research. Hillsdale, NJ: Lawrence Erlbaum. Eerdmans, Susan L. “A review of John J. Gumperz’s current contributions to Interactional Sociolinguistics”. This volume. Gomez-Imbert, Elsa. 1996. “When animals become “rounded” and “feminine”: Conceptual categories and linguistic classiWcation in a multilingual setting”. In Rethinking Linguistic Relativity [Studies in the Social and Cultural Foundations of Language 17], J.J. Gumperz and S.C. Levinson (eds.), 438–469. Cambridge: Cambridge University Press. Gumperz, John J. 1971. Language in Social Groups. Stanford: Stanford University Press. Gumperz, John J. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Gumperz, John J. 1992a. “Contextualization and understanding”. In Rethinking Context: Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 229–252. Cambridge: Cambridge University Press.
Response essay
Gumperz, John J. 1992b. “Contextualization revisited”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 39–53. Amsterdam and Philadelphia: Benjamins. Gumperz, John J. 1996. “The linguistic and cultural relativity of conversational inference”. In Rethinking Linguistic Relativity [Studies in the Social and Cultural Foundations of Language 17], J.J. Gumperz and S.C. Levinson (eds.), 374–406. Cambridge: Cambridge University Press. Gumperz, John J. 1997. “On the interactional bases of speech community membership”. In Towards a Social Science of Language: Papers in Honor of William Labov, vol. II: Social Interaction and Discourse Studies, G.R. Guy et al. (eds.), 183–203. Amsterdam and Philadelphia: Benjamins. Gumperz, John J. 1999. “Culture in the cultural defense”. Proceedings of the Sixth Annual Symposium about Language and Society (1998), SALSA VI:115–132. Austin: Department of Linguistics, University of Texas. Gumperz, John J. and Berenz, Norine. 1993. “Transcribing conversational exchanges”. In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.D. Lampert (eds.), 91–121. Hillsdale, NJ: Lawrence Erlbaum. Gumperz, John J., Cook-Gumperz, Jenny and Szymanski, Margaret. 1999. Collaborative Practices in Bilingual Cooperative Learning Classrooms. CREDE Research Report 7. Santa Cruz, CA and Washington DC: Center for Research on Education, Diversity and Excellence. Gumperz, John J. and Field, Margaret. 1995. “Children’s discourse and inferential practices in cooperative learning”. Discourse Processes 19:133–147. Gumperz, John J. and Hymes, Dell (eds.). 1972. Directions in Sociolinguistics: The Ethnography of Communication. New York: Holt, Reinhart and Winston. Gumperz, John J. and Levinson, Stephen C. (eds.). 1996. Rethinking Linguistic Relativity [Studies in the Social and Cultural Foundations of Language 17]. Cambridge: Cambridge University Press. Hanks, William F. 1996. Language and Communicative Practices. Boulder, CO: Westview Press. Lave, Jean and Wenger, Etienne. 1991. Situated Learning: Legitimate Peripheral Participation [Learning in Doing: Social, Cognitive, and Computational Perspectives]. Cambridge: Cambridge University Press. Levinson, Stephen C. “Contextualizing ‘contextualization cues’”. This volume. Lucy, John A. 1993. “ReXexive language and the human disciplines”. In ReXexive Language: Reported Speech and Metapragmatics, J.A. Lucy (ed.), 9–32. Cambridge: Cambridge University Press. Prevignano, Carlo L. “On Gumperz and the minims of interaction”. This volume. Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This volume. Selting, Margret et al. 1998. “Gesprächsanalytisches Transkriptionssystem (GAT)”. Linguistische Berichte 173:91–122. Silverstein, Michael. 1992. “The indeterminacy of contextualization: When is enough
125
126 Language and interaction
enough?” In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 55– 76. Amsterdam and Philadelphia: Benjamins. Silverstein, Michael. 1993. “Metapragmatic discourse and metapragmatic function”. In ReXexive Language: Reported Speech and Metapragmatics, J.A. Lucy (ed.), 33–58. Cambridge: Cambridge University Press. Silverstein, Michael and Urban, Greg. 1996. Natural Histories of Discourse. Chicago and London: University of Chicago Press. Thibault, Paul J. “Contextualization and social meaning-making practices”. This volume.
Body dynamics, social meaning-making, and scale heterogeneity 127
Chapter 9
Body dynamics, social meaning-making, and scale heterogeneity Re-considering contextualization cues and language as mixed-mode semiosis1 Paul J. Thibault
1. Language as mixed-mode semiosis: Topological and typological dimensions Gumperz makes the distinction between “contextualization cues”, which are a class of pure indexicals, and “symbolic signs”, including symbolic lexical and grammatical signs. According to Gumperz, contextualization cues “have no propositional content” (Prevignano and di Luzio, this volume: 9). Unlike symbolic lexical and grammatical signs, contextualization cues qua pure indexicals “signal only relationally and cannot be assigned context-free lexical meanings” (Prevignano and di Luzio, this volume: 9). Levinson makes a similar distinction between “conscious ‘verbal’ coding and less conscious, less fully-coded prosodic and paralinguistic channels” (Levinson, this volume: 36). The former are explicit and fully foregrounded; the latter tend to be vague, less explicit, non-propositional and backgrounded (see Levinson, this volume: 36). “Contextualization cues” provide the interpretive background for foregrounded explicit propositional meaning. In this chapter, I shall explore some further implications of distinctions proposed by Gumperz and Levinson in order to explore how we can develop a more integrated account of the relations between implicit and explicit modes of semiosis. Both implicit and explicit modes are intrinsic to the internal organization of language. Both are necessary for explaining how language functions in contexts which implicate semiotic processes on a diversity of space-time scales that go beyond the here-now scale of the given occasion of discourse.
128 Language and interaction
It is interesting to note, in the light of the observations made in the preceding paragraph, how the distinction between formally segmentable propositional meanings and the holistic and non-propositional properties of contextualization clues reXects the entrenched separation in much of the language sciences of semantic-propositional meanings as explicitly coded in lexicogrammar, and other meanings of the kind grouped under the term “contextualization cue” as being extrinsic to language form, i.e. as impinging upon or otherwise aVecting the latter from outside (Thibault and Van Leeuwen 1996). The problem with this account is that it preserves a view of language in which segmentable propositional forms are de-contextual and unvarying from one context to another, standing independently from the subjective perspectives of language users and the richly textured and situated nature of language-in-use. In this view, the core of linguistic organization stands apart from the eVects of time-bound contextualization as abstract and unchanging formal features. The solution to this problem lies in a framework which recognizes that semiotic systems, including language, are dynamic open systems which are cross-coupled to both the ecosocial and bodily environments of their users. Language, along with its ecosocial and bodily environments, are continually interacting with each other in the real-time of discourse events. In the process, language entrains bodily and ecosocial dynamics to its own dynamics. The distinctions between, for example, explicit propositional and implicit nonpropositional meanings suggest that language can be divided into indexical and symbolic components and functions. However, in discourse, such distinctions only emerge and co-evolve as the overall trajectory of the discursive event unfolds in time. In such a view, context is not something which is external to and independent of the intrinsic properties and dynamics of semiotic systems. But this is exactly the view which the separation between explicit, formally segmentable proposition forms and contextualization cues has not entirely succeeded in shaking oV. This is so, I argue, because the former notion preserves the idea that some properties of language exist independently of context, whereas others are contingent contextualization cues which act on the former without being essential to the explanation of the internal organization of language form. However, it is the fact that both formal and functional properties emerge as a result of the embeddedness, so to speak, of language in context which should be taken as a central fact about language. This requires an
Body dynamics, social meaning-making, and scale heterogeneity 129
alternative view, in which the distinction between context-independent and context-dependent meanings and forms gives way to one in which both proposition-bearing forms and contextualization cues, to stay with the Gumperzian parlance for the moment, are seen as depending on the overall ecosocial context in which the various components interact with and are integrated with each other. In order to accomplish this task, it is necessary to investigate language as a mixed semiotic mode, displaying both typological-categorial and topologicalcontinuous modes of organization. In this regard, it is useful to reXect on the following quotation from Lemke (1999), who distinguishes two main classes of semiosis along these lines, as follows: …(a) those cases in which the features of representamen that are criterial for some SI [system of interpretance, PJT] to interpret it as a sign of some X may vary continuously, so that quantitative diVerences of degree in a feature of R normally lead to diVerences of degree or kind in the interpretant, vs. (b) those in which all representamina are classiWed by the SI into a discrete spectrum of types, and each R-type is interpreted as a distinct X. (Lemke 1999: 9)
The Wrst type is topological-continuous variation; it is quantitative diVerences of degree that makes a diVerence in this mode. The second type is founded on discrete typological-categorial distinctions. The Wrst type is a generalization of Bateson’s notion of “analogic communication”; the second generalizes his notion of “digital communication” (Bateson 1973 [1972]:342–344). The lexicogrammar and semantics of natural language are, for the most part, based on typological-categorial diVerence. Lemke further points out that while it is possible to “map continuous variation in X onto continuous variation in R (and vice versa),” (topological) “discrete variants of X onto discrete variants of R (and vice versa),” (typological) there may also be “mixed modes of semiosis in which the continuous is mapped onto the discrete and vice versa.” These mixed modes are fundamental to the understanding of the dynamics of multiscale dynamical systems such as language. Language is often made to appear to be less of a mixed semiotic mode than I believe it actually is. This is so because of the focus on explicit propositional meaning qua symbolic representational mode, along with the formally segmentable lexicogrammatical forms which realize this, seen as the deWning characteristic of language. This focus emphasizes the predominantly typological-categorial nature of lexicogrammatical forms and the semantic categories
130 Language and interaction
which these forms realize. This means that both the semantic categories and the lexicogrammatical forms which realize them are discrete. That is, the categories and the forms which realize them are deWned by contrasts with other discrete forms and categories in relation to the paradigmatic systems and the structures whereby language is theorized as a system of contrasting terms in which their respective values are deWned by the place of each term in the overall system. For example, consider the clause complex, When tides are very large, Wsh in the estuaries for golden snapper, trevally, salmon, cod and the elusive barramundi. The second clause in this complex is an instance of the grammatical category ‘imperative mood’. Imperative mood is usually seen as a discrete type-category which corresponds with other mood categories such as declarative, interrogative, exclamative moods in the mood system of English grammar. The noun tides also illustrates the same point. This noun is plural rather than singular, as indicated by the morphemic suYx -s. The noun is either plural or singular; the grammar does not recognize continuously varying degrees of plurality or singularity between these two categorial distinctions. This is a discrete typological-categorial distinction which the lexicogrammar of English realizes through, for example, the presence or absence of morphemes such as the one mentioned here. English grammar has no resources for specifying topological-continuous variation between the semantic distinctions [singular] and [plural]. Furthermore, the nouns tides, estuaries, snapper, trevally, salmon, cod and barramundi specify discrete type-categories of the schematic category Thing, which is prototypically realized by nouns in English. Thus, a given Wsh which the fortunate angler catches in the estuaries around Darwin Harbour is likely to be categorized as an instance of the semantic type-category represented by the nouns snapper, trevally, salmon, cod and barramundi. These words refer to discrete categories of Wsh, just as the lexicogrammatical forms qua exponents of the word class ‘noun’ are themselves discrete: the noun snapper names the semantic category [snapper] rather than the categories [salmon] or [barramundi]. Nor does this noun name continuously varying degrees of ‘snapperness’ or ‘salmonness’ or ‘barramundiness’ which lie somewhere between these discrete distinctions. The typological mode of semiosis is certainly a prominent characteristic of language. Nevertheless, language is not exclusively organized according to typological-categorial principles. As the observations made by Gumperz and Levinson on contextualization cues show, language also exhibits topological-continuous
Body dynamics, social meaning-making, and scale heterogeneity
modes of semiosis which are integrated with the typological-categorial principles outlined above. In other words, language has properties of mixedmode semiosis and this requires that both its internal organization and contextual functioning be explained accordingly. The observations made in the preceding two paragraphs may also be extended to the phonology of the language. In the Indo-European tradition of linguistics, the longstanding focus on the phoneme has privileged segmental forms of phonological organization, along with the function of phonemes and syllables, in symbolizing categorial distinctions such as morpheme and word units and boundaries on the lexicogrammatical level. Yet, the contextualization cues that Gumperz focusses on have to do with how “the phonological Xow manages the Xow of discourse” (Halliday 2000a:108). The distinctions which phonemes and syllables make between this and that morpheme or word are typological-categorial ones. Furthermore, they reXect that aspect of the phonological system which arbitrarily realizes lexicogrammatical distinctions. Firth’s critique of such a monosystemic approach to phonology had long ago envisaged a polysystemic science of the sounds of language as involving “a plurality of systems of interrelated phonematic and prosodic categories” (1957 [1948]:137). The role of the phonological Xow in the Xow of discourse has to do with nonarbitrary relations between phonology, on the one hand, and lexicogrammar and discourse, on the other. In English, for example, rhythm and intonation are central here. However, both rhythm and intonation are intrinsic to the internal organization of English phonology, such that phonology, on analogy with lexicogrammar, is seen as comprising a number of domains of systemic organization deWned as a rank scale of hierarchically organized units. In English, these phonological units are a tone unit, a rhythm unit (the foot), and an articulatory unit (the syllable) (Halliday 1992, 2000a). In English, nonsegmental modes of interpersonal meaning (e.g. mood and modality) and textual meaning (e.g. the organization of discourse as Given and New information) in lexicogrammar are often integrated with nonsegmental features of rhythm and intonation in discourse. Therefore, such “contextualization cues” in English phonology are not external to language design; rather, they are nonsegmental forms of phonological organization which are constitutively inseparable from the segmental or particulate forms of organization which are also found in both phonology and lexicogrammar.
131
132 Language and interaction
The tripartite perspective on language as “particle”, “wave”, and “uncoded Weld”, Wrst proposed by Pike (1967) and subsequently taken up and adapted by Halliday (1979), shifts the emphasis away from the distinction between fully-coded propositional content and contextualization cues. Instead of this distinction, Halliday’s adaptation of Pike’s tripartite perspective proposes that a number of diverse functional regions, or metafunctions, are simultaneously conWgured in lexicogrammatical form. Halliday’s (e.g. 1979) metafunctional account claims that the intrinsic functional organization of lexicogrammatical form is organized in terms of four general semantic regions, each having their characteristic modes of formal organization in the lexicogrammar of natural language. According to Halliday, lexicogrammar simultaneously: (i) categorially construes experience, including naming and referring; (ii) enacts interpersonal relationships; (iii) provides the means whereby language coheres textually as discourse which is operational in context; and (iv) connects linguistic units to each other in relations of logical (causal, temporal, etc.) dependency. These four distinct functional regions are referred to in systemic-functional linguistic theory as the “experiential”, “interpersonal”, “textual”, and “logical” metafunctions, respectively. Each of the four metafunctions favors a diVerent mode of formal (lexicogrammatical) realization. Thus, experiential meanings are realized by particulate or segmental forms of realization, textual meanings by wave-like or periodic ones based on the ways in which waves of discursive activity Xow into and merge with each other, and interpersonal meanings favor prosodic or scopal forms of realization such that an interpersonal Weld is the domain surrounding a given interpersonal selection — e.g. mood or modality in the clause. In this perspective, the given interpersonal selection (e.g. mood or modality in the clause) syntagmatically extends over and in some way modiWes or deforms (inXuences) some other feature within the given syntagmatic Weld which it holds in its scope. In so doing, the interpersonal feature inXuences and shapes the given syntagmatic Weld for a particular interactive purpose. Pike’s original proposal thus allows us to see that language is internally organized according to a number of diVerent, yet complementary, perspectives, such that no single perspective such as the particulate or segmental one which has dominated linguistics can satisfactorily explain all aspects of the formal organization of language. Experiential, textual, and interpersonal meanings are nonarbitrarily-related to their diVerent modes of
Body dynamics, social meaning-making, and scale heterogeneity
lexicogrammatical realization such that both typological-categorial and topological-continuous modes of semiosis are seen to operate within lexicogrammatical form, simultaneously. Parallel principles of organization also operate on the expression stratum of phonology, as Firth (1957 [1948]) showed, such that the nonsegmental dimensions of linguistic organization are fully integrated with the segmental ones. In such a perspective, the distinction between segmental propositional forms and nonsegmental contextualization cues can be assimilated to one in which the nonarbitrary or natural relation between phonological processes, the human body, lexicogrammatical and discourse processes is revealed as fundamental to the nature and functioning of language. This requires a perspective which sees language as organized on diverse space-time scales. In the next section, I shall explore this question.
2. The semiotic Principle of Alternation and mixed-mode semiosis Lemke has proposed the Principle of Alternation as a way of theorizing, in terms of the three-level scalar hierarchy, how semiotic functions are mapped onto dynamical scale levels, and the re-organization of continuous variation into discrete variants. Lemke deWnes this Principle as follows: Each new, emergent intermediate level N in a complex, hierarchical, self-organizing system functions semiotically to re-organize the continuous quantitative (topological) variety of units and interactions at level (N–1) as discrete, categorial (typological) meaning for level (N+1), and/or to re-organize the discrete, categorial (typological) variety of level (N–1) as continuously variable (topological) meaning for level (N+1). (Lemke 1999: 9; italics in original)
Gumperz’s discussion (Prevignano and di Luzio, this volume: 9–10) of the interaction between symbolic signs with propositional content and indexical signs functioning as contextualization cues suggests the mixed semiotic modes mentioned above. This brings into focus two complementary perspectives on semiosis. The Wrst, topological mode is grounded in the materiality of our embodied being-in-the world interactively. It is closer to our primary bodily being-in-the-world before language. Its scale is, in the Wrst instance, the organismic one of the body-brain in its immediate physical-material environment. The second mode represents the system’s expansion into typological-
133
134 Language and interaction
categorial phase space whereby semiosis cascades and ramiWes across spacetime scales not tied to the here-now scale of the organism’s material interactivity with its immediate environment (Lemke 2000:194; Thibault 2000). Whereas the topological mode is more tightly cross-coupled to the material processes of the body, the typological mode predisposes the linking of a particular meaning-making event to space-time scales beyond the herenow scale of the organism’s material interactivity with its immediate environment. This does not mean that the typological mode transcends the topological mode. The point is that language-in-use, rather than language as abstract system, operates on the basis of “mixed-mode” semiosis such that it combines both topological and typological principles as central to its contextual functioning. The mapping of semiotic functions onto diverse organizational scales has dynamical implications for the emergence of meaning, viz. the emergence of a new level in the scalar hierarchy of semiotic functions can only occur, both phylogenetically and ontogenetically, if and only if a new level in the hierarchy of semiotic interpretance emerges. Each higher level in the hierarchy has multiple realizability at lower levels. For example, it is the many-to-one mapping of topological-continuous variety at the N–1 level of the continuous topological Xux of the phenomena of experience onto the N level of lexicogrammar that allows for the Wltering of N–1 non-criterial Xuctuations at the N level. Furthermore, discrete lexicogrammatical forms at this level can be re-construed at the N+1 level as having an increased variety of meanings, depending on the ways in which constraints on the N level of lexicogrammar are taken up and responded to on the basis of their connections to a culture’s intertextual networks on the more global, longer term N+1 level of the SI (see Lemke 1999; Thibault, forthcoming). Following Lemke’s Principle of Alternation, we see that the semiotic transformation of continuous topological variation in the Xux of perceptual and other phenomena of experience to discrete lexicogrammatical forms and of these to continuous semantic meanings, in the progression from level N–1 to level N+1, constitutes what Lemke deWnes as “a semiotic transformation of the information content of lower levels as signs for higher levels, allowing manyto-one classiWcations and one-to-many context-dependent reinterpretations” (1999: 9). It is the dynamics of the semiotic transformations mapped onto the scalar hierarchy that allows for the emergence of the symbolic possibilities of
Body dynamics, social meaning-making, and scale heterogeneity
language, and hence its capacity to form contextual links with many diverse space-time scales in a given act of meaning-making.
3. Material and semiotic dimensions of meaning-making Meaning-making is both a physical-material process and a semiotic-discursive one. Gumperz emphasizes the social dimension of meaning-making without, however, giving equal importance to the material dimension. The material and social dimensions are equally necessary to the development of an adequate theory of how meanings are made in a given ecosocial system. Social practices are also material processes. Gumperz emphasizes the social practices through which meaningful relations are jointly constructed in and through the social activities whereby agents interact with each other. The material dimension of meaning-making brings it into contact with other physicalmaterial processes and Xows, including biological ones, in the ecosocial system. In meaning-making, semiotic and material processes are cross-coupled as meaning-making activity. It is such networks of activities which have spatial and temporal extension beyond the local here-now that support and sustain particular cultural practices. There is no question of a localized, homogeneous community deWned on the basis of one’s membership of a given language, social dialect, and so on. Networks of human activity are not necessarily localized in this sense. The point is that symbolic meaning-making resources in particular enable human agents, material objects, and so on to interact with each other across very many diVerent space-time scales qua social-cultural network. They do so on the basis of their participation in the meaning-making practices in and through which a given network is deWned. I want to make clear that the deWnition of network oVered here diVers in several ways from Gumperz’s use of this term. According to Gumperz, “community membership is of course directly linked to participation in such networks of relationships, but in our postindustrial worlds, it is less and less possible to take this for granted” (Prevignano and di Luzio, this volume: 18). Gumperz directly equates “community membership” with participation in “institutionalized networks of relationship”, such that networks are seen as localized to speciWc communities in space and time, presumably according to (unspeciWed) assumptions made about the pre-industrial world. In the deWnition I am operating with here, a
135
136 Language and interaction
network is the organization of material and social processes across potentially diverse space-time scales. In this way, the problem of the localization of a community and its reduction to linguistic membership gives way to that of which cross-couplings of material and social processes sustain the socialcultural practices of the network. Such a network cannot exist independently of such cross-couplings. Ecosocial networks have a “1-dimensional reticular topology spread through a three-dimensional ecosystem” (Lemke 2000:200). Networks do not have deWnite edges or boundaries. For this reason, they cannot, by deWnition, be reduced to speciWc point locations. In the view outlined above, a language system’s internal organization consists in the speciWc terms and the diVerential relations among these. However, a system also has higher-scalar environmental constraints or boundary conditions that interact with the system at any given moment in time such that the former can aVect and shape the latter. In such a view, contextinvocative, aVectual, rhetorical, subjective, and so on, properties of languagein-context are neither epiphenomenal nor secondary. Moreover, the notion of a “language system” does not refer to some reiWed entity ‘out there’ which exists independently of the social meaning-making practices of a given social group and which has an independent causal status. Instead, the notion of the language system refers to the semiotically salient diVerences which the members of a given social group recognize as potentially meaningful. These diVerences are immanent in the meaning-making practices of the members of some social group as well as being immanent in the symbolic neural space of biological individuals. The causal status of such a system of diVerences is not explainable in externalist terms as if the language system, so deWned, causally acted upon and hence determined the meaning-making behavior — the linguistic choices, and so on — of social agents. It is not a question of a mechanistic cause-and-eVect view of grammar as in, say, Chomsky’s (1965) transformational-generative model. In such a view, grammar qua competence and individual performance stand in a cause-and-eVect relation within the same scalar level of the individual organism. Rather, it is a system of constraints or boundary conditions whereby higher scalar levels informationally constrain lower levels. The semiotic activities of the social group is not a mere aggregate of the behaviors of the individual organisms. This applies to all forms of semiotic activity such as, for example, language, gesture, depiction, and so on. These cannot be reduced to or uniquely explained in terms of the neurophysiology of indi-
Body dynamics, social meaning-making, and scale heterogeneity
vidual organisms or the locally perceived goals and intentions of social actors. Rather, the behaviors of the individual organisms involved are entrained to higher-scalar ecosocial patterns. Furthermore, the organization of the higher-scalar level cannot be described in terms of the organization of entities and processes existing on lower scalar levels. For example, Cowley (1998) has shown how the interpersonal attunement of individuals to each other in conversational events can depend on physical properties of utterances such as the rate at which participants vibrate their vocal chords as they tune in to and synchronize their voice with variables such as Xuctuations in fundamental pitch of the other’s voice (1998:549). Such sensori-motor events take place on a temporal scale that is much faster than the rising and falling tones that are recognized as meaningful contrasts in intonation. Intonational contrasts are perceivable and are salient on a slower temporal scale than that of the centiseconds characteristic of the micro-temporal scale of vocal chord vibration. Sonograph records show, as Cowley (1998:550–551) documents, that participants in conversation tune into and act on each other at this physical acoustic level of, for example, pitch Xuctuation. The diVerent temporal scales of the entities and processes on the lower scalar sensori-motor level as compared to that of the higher-scalar level where contrasts in intonation are salient shows how, in speaking, participants both act on and respond to each other on the basis of voice dynamics existing on potentially many diVerent temporal scales which implicate physical-material (bodily) processes not normally considered to be intrinsic to linguistic form, but which are acted on by participants as meaningful components of the act of conversing with others. The process whereby participants converge on a particular fundamental pitch frequency is an example of the embodied ways in which interactants tune into each other in ways recognized as interpersonally salient. In other words, the very fast, lower scalar vocal dynamics of individual participants are entrained to higher-scalar rhythmic and harmonic patterns at the level of the conversational event. In this way, the sensori-motor behavior of the individual participants spontaneously coheres to higher scalar patterns which have to do with, for example, relational dynamics existing on quite diVerent temporal scales. Sensori-motor dynamics such as vibrating vocal chords are biological initiating conditions that self-organize into patterned behavior at the level of the particular conversational event. In turn, the spontaneous emergence of such patterns is entrained to still higher level patterns
137
138 Language and interaction
such as, for example, the historical-biographical unfolding of the social relations among the participants to any given conversational event. The relational dynamics of this level provide boundary conditions which regulate and entrain the dynamics of the lower levels to its patterns. In the examples analyzed by Cowley (1998), the higher scalar level of the social group to which individual dynamics are entrained is, in the Wrst instance, that of the family, comprising the two parents and their daughter, which is a social unit having its own dynamical patterns characteristic of that level and their history. In postulating such patterns and their history on their particular temporal scale — months and years as compared to centiseconds! — the analyst is, in eVect, postulating an entity — a traditional one in this case — which has the possibility of becoming part of some theoretical discourse whereby diVerent scalar levels come clearly into view. In this case, the focal level is that of the conversational event. The scalar hierarchy of relevant levels with reference to this focal level can be described as a three-level system, each with its own processes and entities unique to that level. Thus: N+1: relational dynamics of family as social group and the history of its evolving patterns N: the conversational event comprising a number of individual participants whose behaviors cohere to the higher-scalar patterns of the overall event N–1: the sensori-motor dynamics of individual organisms as biological initiating conditions The hierarchy of levels featured here illustrates how language-in-use cannot be explained in terms of a single, scale-homogeneous level of semiotic or other relations. Instead, it is a cross-scalar, semiotic phenomenon which is able to link potentially very many scales, ranging from the organismic to a diversity of space-time scales extending ‘up’ to the history of an entire culture (Lemke 2000). This aspect of language thus draws attention to the principle of the scalar heterogeneity of language which I alluded to in the title of this chapter. For this reason, it is perfectly legitimate contra Gumperz (This volume: 121) to claim that language and other semiotic modalities construe experience (see also Halliday 2000b:223). This is so in the sense that a language exists on space-time scales that go beyond the “shared inferencing of human actors”,
Body dynamics, social meaning-making, and scale heterogeneity 139
seen as “the locus of interpretive assessment” (Gumperz, this volume: 121). Meaning-making does not occur solely on the here-now scale in which social actors’ interpretive assessments take place; it is also a constitutive part of processes on much larger spatial and temporal scales right up to long-term historical processes of the evolution of a meaning system. Meaning-making implicates complex cross-scalar semiotic and material relations operating simultaneously over very many scales rather than being limited to the locallyinterpreted intentions and inferences of social actors interacting with other social actors. Scalar heterogeneity is therefore intrinsic to the nature and functioning of language (Lemke 2000; Thibault 2000). For this reason, scalar heterogeneity must be an essential component in the theorization of language.
4. Contextualization cues as indexical signs Gumperz, as we have seen above, makes a distinction between context-free symbolic signs and context-dependent indexical signs of which contextualization cues are a class (Prevignano and di Luzio, this volume: 10). The former have propositional content; the latter do not. Contextualization cues, Gumperz argues, “serve to retrieve the contextual presuppositions conversationalists rely on in making sense of what they see and hear in interactive encounters” (Prevignano and di Luzio, this volume: 9). These function on the basis of “direct conventional associations between signs and context, established or transmitted through previous communicative experience” (Prevignano and di Luzio, this volume: 10). Such conventional indexical associations arise as a consequence of the ways in which initially random associations between indexical sign and what is seen or heard in the perceptual purview of the interactants emerge from the micro-level dynamics of a given here-now association so as to take on more global conventional associations on the macro-level of the social group. Prosodies and rhythm are grounded in the bio-physical characteristics and potentialities of the human body. The characteristics of the body allow the individual to perform a certain range of bodily behaviors. That is, the anatomical and physiological structures and processes of the body qua biological organism impose their own lower-level constraints on what the body can do relative to its environment. However, the conventional indexical associations referred to by Gumperz refer to collective patterns which emerge as a result of
140 Language and interaction
the ways in which the bodily potential of the organism is entrained to largerscale patterns of social interaction. In other words, a large number of degrees of freedom of our bodily dynamics on level N–1 of the three-level hierarchy described in Section 3 is entrained to a more restricted set of possibilities in virtue of the conventional indexical associations on level N through our attempts to reciprocally adapt to the behavior of others in our interactive encounters with them on level N+1. In this way, we see how the conventional indexical associations discussed by Gumperz result from our ongoing eVorts to adapt our own bodily dynamics to the bodily dynamics of other individuals with whom we engage in interactive, goal-directed or purposeful activity. Such mutually modulated body dynamics constitute emergent collective patterns which, in turn, exert their own eVects and constraints on the activities of individuals. Gumperz describes such indexical associations as “context-dependent” because their contextual eVects arise in immediate situations in the here-now of a given interaction. This fact is evidenced by the ways in which the given (indexical) sign is dependent on speciWc sensori-motor cues for its interpretation. Indexicals of the sort discussed by Gumperz under the rubric of “contextualization cues” allow the individual to interact with other individuals on, roughly speaking, the space-time scale of their shared perceptual purview. In this case, the principle of semiotic mediation at work here means that the here-now body dynamics — e.g. the particular vocal or manual-brachial gesture — which take place on the scale of seconds or fractions of seconds in the real-time of our perceptual-motor activity, indexes contextual associations which are commensurate with the scale on which the gesture qua indexical sign itself is articulated. This is the scale on which perceptual-motor activity and its environmental eVects is both enacted and observed through the perceptual systems of the body in interaction with its ecosocial environment. In other words, semiotic mediation by pure indexicals already entails scales beyond the organismic one per se at the same time that it is constrained by higher-scalar constraints in the form of the conventional associations that have emerged between indexical signs and their respective contextual values. Figure 1 models the complex, non-linear, inter-level constraints that operate here across a number of diVerent scalar levels, ranging from the organismic scale of the human body to the higher-scalar, socio-cultural conventions involved. Now, the fact that contextualization cues are conventional associations
Body dynamics, social meaning-making, and scale heterogeneity
emergence of conventional associations entraining of bodily dynamics to large-scale or content patterns of human interaction
situation-specific value Index
bio-physical dynamics of human body
expression
constraints on bodily dynamics
Figure 1. The entraining of body dynamics qua conventional indexical association to higher-scalar ecosocial dynamics.
already implies that they have a history and that this history has both individual and social dimensions. That is, a given use of a conventional indexical association implies scales of dynamical complexity in the given ecosocial system which go beyond the momentary use of a particular contextualization cue in some context. How did the user come to use the particular contextualization cue? In relation to which social practices in the ecosocial system? How did the given cue come to be part of the particular user’s repertoire of meaning-making resources? In relation to which social networks and their characteristic bodily dispositions (cf. habitus)? Contextualization cues such as rhythm and prosodies have their basis in the neurophysiological substrate and perceptual-motor activity of the organism, yet their eVects extend beyond the organismic scale of the individual and into the ecosocial environment of other social agents for whom they are potentially meaningful. The questions posed here point to the fact that contextualization cues qua indexical signs are embedded in at least two interpenetrating scales which have a history. First, there is the personal-biographical history of the individual’s trajectory through speciWc social networks. Secondly, there is the history of the ecosocial system within which individual trajectories are in turn embedded. Contextualization cues can only have their meaning in relation to both embodied individual memory and the accumulated history of the meaningmaking practices of a given culture or some part of this. Indexical signs mediate bodily processes so as to amplify their eVects on the ecosocial environment, including other individuals, beyond the scale of the micro-level processes which constitute their bio-physical substrate on the organismic
141
142 Language and interaction
level. Indexicals integrate phenomena on the basis of the perceptual-motor access of interactants such that the latter perceive the index as being contiguous with a given contextual object or value. In this way, the use of a particular prosody or rhythm is construed as being contiguous with an inferred communicative intent in the sense that the situated perception of the cue provides access to the inferred intention (see Prevignano and di Luzio, this volume: 12– 14 for examples). The indexical inferring of such communicative intentions is necessarily situated and on-line, so to speak. In other words, the relevant space-time scales imply scales of dynamical complexity which are roughly commensurate with the space-time of the unfolding meaning-making activity. While intentions cannot be directly perceivable, they are, however, indexically invoked by on-line contextualization cues, based on perceptualmotor access to the relevant domain. This means that the inferred intention is an indexically-invoked contextual value — not a material object — which is integrated with a perceivable cue through processes of semiotic mediation. The fact that contextualization cues “channel the interpretative process” (Prevignano and di Luzio, this volume: 10) so as to constrain the possible meanings of an utterance demonstrates this last point. That is, the given contextualization cue is selectively co-contextualized in relation to some other features of the wider situation which the speciWc co-contextualization helps to create or enact. In this sense, situated inferences as to what a speaker ‘intends to mean’ in a given context are not external to the situation, but are a constitutive part of the meanings which constitute and deWne that situation. Thus, the use of, say, a particular prosody, rhythm or voice quality, in order to perform the constraining function mentioned above, implies a number of space-time scales, all of which are simultaneously and seamlessly interwoven in the realtime unfolding of the communicative event. There is, for example, the scale of the activity which is being performed. Furthermore, there is the scale of the history of the personal relationship between the participants, which may vary from the contingent and Xeeting to one spanning an entire life-time. There is also the scale of the social networks in which particular prosodies and so on typically redound with other semiotic features in the formation of a given social situation-type. This last point brings me to a Wnal consideration of the relations between semiotic forms and the discursive practices in which the former are embedded.
Body dynamics, social meaning-making, and scale heterogeneity 143
5. Linguistic practice and discursive practice According to Gumperz: Sociolinguistic explanation, if it is to be relevant to today’s concerns, cannot implicitly accept traditional categories of language, culture and society. I believe that interaction at the level of discursive or communicative practices […] must be seen as separate from either linguistic or socio-cultural processes. It is constituted by the interplay of linguistic, social and cultural/ideological forces and governed or constrained by partly universal and partly local-speciWc organizational principles. (Prevignano and di Luzio, this volume: 9–10)
This formulation of discursive practices, on the one hand, and linguistic and socio-cultural practices, on the other, suggests that socio-cultural groups and their associated practices have a prior existence with respect to the ways in which they are constituted, construed and evaluated through language and other semiotic activities, material processes, and related objects and artifacts. In my view, any such separation is an analytical, rather than a constitutive, one (Thibault 1991:229–240). Social formations are neither logically nor ontologically prior to the social semiotic and material activities in and through which meanings are made. Instead, they are constituted by these activities. Take Gumperz’s example of the Afro-American speaker’s use of the utterance Ahma git me a gig (Prevignano and di Luzio, this volume: 19–20). As Gumperz’s analysis shows, the lexicogrammatical and phonological features which co-pattern in this utterance instantiate semantic patterns which are typical of register-speciWc semantic relations that occur in the speech of the Afro-American speakers referred to. Such patterns are characteristic of the discourse community to which these speakers belong. Moreover, it is not simply a question of the speaker talking about a particular topic in a way which diVers from the ‘academic norm’ of standard American English referred to by Gumperz. Rather, the speaker activates a small fragment of a much wider intertextual thematic formation which is recognizable and interpretable by members of the discourse community in question. It is a thematic formation which is also typically associated with a particular evaluative orientation or social viewpoint insofar as it implicates an evaluative stance both towards his interlocutors as well as towards the thematic content of his utterance. For this reason, the utterance can be seen as the activation of what Bakhtin (1981) called a particular “social voice”, i.e. a particular social view-
144 Language and interaction
point on some topic. For Gumperz, the particular co-patterning of lexicogrammatical and phonological selections in the utterance functions indexically to convey the speaker’s message by the use of “metapragmatic strategies” (Prevignano and di Luzio, this volume: 19) which index the speaker’s identiWcation with Afro-American social groups who talk in this way. Thus, the utterance is an index of such sociolinguistic group membership or alignment with the members of this group. Nevertheless, it is clear that the choice of this particular utterance does more than index group aYliation. It also realizes a social voice, in Bakhtin’s sense, within a wider system of potentially alternative, conXicting, aligned, and so on, evaluations and viewpoints concerning various topics. My point is that the index of group membership, the theme or topic of the discourse, and the evaluative stance which is invoked are not neatly separable categories: they all come together and are variously realized by the linguistic choices made by the speaker in the context in which he spoke. For this reason, the linguistic processes analyzed by Gumperz, presumably along with other bodily features associated by the given social habitus of the speaker, are said to be constitutive of the given social formation on that particular occasion of talk. That is why, in my Wrst contribution to this volume (Thibault, this volume: 50), I argued for the three-way view of meaning-making practices articulated there as a way of showing how particular associations of the evaluative, thematic, and other meanings realized in linguistic and other semiotic forms are created in discourse at the same time that they directly implicate the higher-scalar social formations relevant to their interpretation. It is the use of small-scale lexicogrammatical patterns in potentially very many local contexts that leads to the emergence of higher-scalar intertextual formations and their associated practices. A given usage in some local context can therefore invoke the entire formation of which it is a mere fragment. By the same token, the higher-scalar intertextual formation, which exists on the space-time scale of an entire community and its history, semiotically constrains speciWc, fragmentary instances of the whole formation. Without some such system of the complex, multilevel and non-linear constraints operating across all scalar levels from the most global to the most local, it is diYcult to see how social meaning-making practices shape social and individual life without having recourse to outmoded models based on mechanistic cause-and-eVect relations. This further suggests that higher scalar levels such as the intertextual formations mentioned here have their own ontological reality. Consequently, they cannot be
Body dynamics, social meaning-making, and scale heterogeneity 145
reduced to or theorized in terms of the lower level constituent units and processes per se from which they emerge and which they in turn constrain. That is one reason why I proposed in Thibault (this volume) the three interrelated perspectives on our meaning-making practices as a way of showing how higher-scalar relations and processes enact their own constraints on the selections made in speciWc texts. Gumperz’s claim that contextualization cues allow interactants to infer intended meanings tends, moreover, to see meaning as internal cognitive processes which originate from within the individual organism. Thus, contextualization cues such as prosody and rhythm are indexes which provide access, by means of inferential processes, to the internal cognitive states of individuals in interaction. This suggests that contextualization cues can be ‘read oV’ individual behavior as indexes which (indirectly) point to or are contiguous with the postulated internal cognitive states of individuals. Furthermore, there are macro-social interpretative conventions which allow contrasts in prosody and rhythm and so on to be appropriately ‘assessed’ and ‘interpreted’ as to their intended meaning in a given context (see Prevignano and di Luzio, this volume: 14). Gumperz further suggests that there are context-free signs which he equates with symbolic, rather than indexical, modes of sign-making. In my view, symbolic signs, as seen from the instantial perspective of their deployment in discourse, are no less context-dependent than are indexical signs. Symbolic signs construe phenomena of experience — both physical-material and imaginary — as instances of the semiotic category realized by the sign. Furthermore, a symbolic system of possibilities such as language is formatted in regular lexicogrammatical patterns which can allow for oV-line reXection on the symbolic signs themselves as objects of knowledge in their own right. Lexicogrammar embodies both indexical and symbolic functions. The diVerence between the indexical and symbolic modes lies in the way in which symbolic modes of semiosis make possible the maximal intersection of diVerent scales such that a particular use of a symbolic sign in some here-now context may implicate many diverse and heterogeneous spatial and temporal scales, all of which seamlessly interact in a given use of a particular symbolic sign in some context. Rather than being “context-free”, symbolic signs function to integrate potentially many diVerent contexts on varying space-time scales (Thibault 2000), such that the texts and discourses which we make out of our symbolic meaning-making resources are relatively small-scale units
146 Language and interaction
which, however, directly implicate larger-scale social networks, larger community-scale systems of intertextual meaning relations and their histories, larger-scale spatial extension beyond the physical-material reality of the textual artifact or the speciWc occasion of discourse in the here-now. Symbolic semiosis does not necessarily entail that interactants have perceptual-motor (indexical) access to the phenomena of experience which are construed through the experiential resources of lexicogrammar. Symbolic meaning-making resources allow for the integration of many diVerent spacetime scales which are not directly accessible to the interactants, but which can be opened up and made accessible so that our relations to other scales beyond the here-now can be both modeled and made relevant to the here-now scale of our moment-by-moment interactions with our ecosocial environment.
Note 1. I wish to thank Carlo Prevignano for his valuable comments on an earlier version of this chapter.
References Bakhtin, Mikhail. 1981. “Discourse in the novel”. In The Dialogic Imagination: Four Essays by M.M. Bakhtin, C. Emerson and M. Holquist (trans.) M. Holquist (ed.), 259–422. Austin and London: University of Texas Press. Bateson, Gregory. 1973 [1972]. “Problems in cetacean and other mammalian communication”. In Steps to an Ecology of Mind, 334–348. London and New York: Granada. Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, Mass.: MIT Press. Cowley, Stephen J. 1998. “Of timing, turn-taking, and conversations”. Journal of Psycholinguistic Research 27 (5):541–571. Firth, J.R. 1957 [1948]. “Sounds and prosodies”. In Papers in Linguistics 1934–1951, 121– 138. London and New York: Oxford University Press. Gumperz, John J. “Response essay”. This volume. Halliday, M.A.K. 1979. “Modes of meaning and modes of expression: Types of grammatical structure, and their determination by diVerent semantic functions”. In Function and Context in Linguistic Analysis: A Festschrift for William Haas, D.J. Allerton, E. Carney and D. Holdcroft (eds.), 57–79. Cambridge and London: Cambridge University Press. Halliday, M.A.K. 1992. “A systemic interpretation of Peking syllable Wnals”. In Studies in Systemic Phonology, P. Tench (ed.), 98–121. London and New York: Pinter.
Body dynamics, social meaning-making, and scale heterogeneity 147
Halliday, M.A.K. 2000a. “Phonology past and present: A personal retrospect”. Folia Linguistica XXXIV (1–2):101–111. Halliday, M.A.K. 2000b. “Grammar and daily life: Concurrence and complementarity”. In Functional Approaches to Language, Culture and Cognition, D.G. Lockwood, P.H. Fries, and J.E. Copeland (eds.), 221–237. Amsterdam and Philadelphia: John Benjamins. Lemke, Jay L. 1999. “Opening up closure: Semiotics across scales”. Paper presented at the conference, Closure: Emergent Organizations and their Dynamics, University of Ghent, Belgium, May 1999. Lemke, Jay L. 2000. “Material sign processes and emergent ecosocial organization”. In Downward Causation: Minds, Bodies and Matter, P. Bøgh Andersen, C. Emmeche, N. Ole Finnemann, P. Voetmann Christiansen (eds.), 181–213. Aarhus: Aarhus University Press. Levinson, Stephen C. “Contextualizing ‘contextualization cues’”. This volume Pike, Kenneth L. 1967. Language in Relation to a UniWed Theory of the Structure of Human Behavior. Second, revised edition. The Hague and Paris: Mouton. Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This volume. Thibault, Paul J. 1991. Social Semiotics as Praxis. Text, Social Meaning Making, and Nabokov’s Ada. Minneapolis and Oxford: University of Minnesota Press. Thibault, Paul J. 2000. “The dialogical integration of the brain in social semiosis: Edelman and the case for downward causation”. Mind, Culture, and Activity 7 (4):291–311. Thibault, Paul J. Forthcoming. The Signifying Body: An Ecosocial Semiotic Theory of Situated Meaning-Making, Embodiment, and Consciousness. Thibault, Paul J. and Van Leeuwen, Theo. 1996. “Grammar, society, and the speech act: Renewing the connections”. Journal of Pragmatics 25:561–585.
148 Language and interaction
Continuing the discussion with John J. Gumperz 149
Chapter 10
Continuing the discussion with John J. Gumperz Carlo L. Prevignano and Paul J. Thibault
Carlo L. Prevignano: In the conclusion to your Response essay, you mention a case study of yours, Communicative Competence as Communicative Practice. For Gumperz in 2002, what are communicative and interactional competences, and what are their relationships with practice? What solution do you propose today to the attempt to capture “regularities of discourse practice without reifying them in terms of a priori rules” (Gumperz, this volume: 122)? John J. Gumperz: The case study, Communicative Competence as Communicative Practice, is still unWnished. Its title alludes to the theoretical notion Wrst introduced by Dell Hymes in the nineteen seventies in connection with his argument against Chomsky’s narrowly-deWned “linguistic competence”. Hymes’ claim was that there are additional forms of systematic knowledge apart from grammatical rules, and that these also enter into the production and comprehension of talk. He concludes that we need a broader, more inclusive notion to account for situated language use along with linguistic structure. In the paper, I analyze an informal, student coVee shop conversation to argue that communicative competence cannot be described in terms of “sociolinguistic rules” abstracted from everyday talk and patterned on those used in grammatical analysis, as is currently done in educational psychologists’ tests of English as second language learners’ speaking abilities. Communicative competence is more directly reXected in the way participants use language to respond to others in negotiating shared understandings in naturally-organized discursive encounters. Based on sample passages from the transcript, my analysis seeks to illustrate the sort of things participants have to know to engage in and maintain discursive coherence in interactive encoun-
150 Language and interaction
ters. I argue that educators constructing tests of communicative competence might concentrate on Wnding ways of assessing learners’ situated interpretive ability. The resulting tests would not just ‘measure competence’, they might also yield normally unavailable insights into how such communicative diYculties can aVect the course of an interaction that might improve teaching programs. For those who are interested, a second paper, Sharing common ground, dealing with the same conversational exchange, makes similar arguments. This paper is published in a Festschrift for Werner Kallmeyer (Gumperz 2002). I don’t remember where I used the term “interactional competence”. Most probably I intended to refer to knowledge of the kind alluded to in the two above-mentioned papers. But let me clarify my basic assumptions a bit more. I believe that if we want to determine how speakers understand each other, we must account for all signs that demonstrably aVect the course of an encounter. I use “communicative practice” (CP) as a general term to cover the relevant verbal and non-verbal signaling processes. Strictly speaking, “discursive practice” is a subcategory referring to the discourse-level elements of CP. But the two terms are sometimes used interchangeably. Along with Silverstein (1992, 1993) and Hanks (1996), whose terminology I build on here, I see communicative practice as constituted by three elements. First, language, roughly describable as a Saussurian system of abstract (symbolic) categories and rules for their combinations. Second, contextualized communicative activity including: (a) symbolic or lexicogrammatical signs; (b) indexical signs such as prosody, rhythm, code- or style-switching, formulaic expressions and other cues; (c) nonverbal signs such as gesture, posture, facial expression, clothing etc.. Third: (a) ideology, in as much as it motivates actors’ meta-communicative evaluations, based, among other things, on cultural beliefs and values; (b) background knowledge relevant to what transpires in an encounter. Like Thibault, I claim that all three elements are always simultaneously involved in the production and interpretation of talk. That is, at the level of CP, talk is grammatical, contextualized and ideologized. This tripartite division has some similarity to Pike’s three dimensions discussed by Thibault in his second paper in this volume. But as I will point out below, my view of the semiotic processes involved is signiWcantly distinct from his. The three elements can of course be, and have for a long time regularly been, abstracted from talk for purposes of in-depth analysis and thus treated as ‘analytically distinct’. But
Continuing the discussion with John J. Gumperz
this means we must see them as abstractions and cannot claim that any one element or constituent thereof can by itself account for the phenomenological facts of communicative practice. Turning to the quote at the end of your question, in using that expression, I meant to argue against scholars who resort to extra-communicatively deWned, largely lay categories, like class, gender, ethnicity, professional identity, community or group membership and the like, to explain or naturalize what happens in situated interactive exchanges, without demonstrating how the relationship works. As the preceding paragraph implies, some forms of social/cultural/ideological knowledge always organize our actions and demonstrably aVect CP interpretive assessments. But this is not equivalent to claiming that participants’ membership in one or another human collectivity as such is directly reXected in talk. What we are talking about here is how interpretive presuppositions, ideologically associated with such a collectivity, aVect interpretation. The nature and workings of such presuppositions must be investigated through the analysis of context-bound verbal exchanges. As to “regularities” of communication practice, I believe that these should ultimately be derived from or related to in-depth analyses of situated encounters in a variety of settings. Because of the way interaction works, the usual questionnaire studies are not satisfactory here. I made some informal comparisons between South Asian and English interviewing practices in a paper of mine (Gumperz 1996), but these are observations that might inform ethnographic studies. I don’t know of any systematic research along these lines. C.L.P.: At the end of your Response essay, you indicate a “distinction [analytically of great importance, I agree] between explicitly metacommunicative [and consciously manipulable] signs and indirect indexicals [which are produced below the level of consciousness] such as contextualization cues”. Do you see today any other diVerence between the former and Gumperzian cues? J.J.G.: On second thought, it might be more fruitful here to speak of a distinction between overtly-lexicalized indexicals, on the one hand, and nonlexicalized or indirect indexicals (including, among others, prosody, rhythm, tempo, certain phonetic or grammatical variables, along with code- and styleswitching), on the other. That would avoid value-laden terms. Linguistic anthropologists, following Silverstein (1992, 1993), sometimes use the term “pure indexicals” to refer to the latter category. That is to say, they are pure in the sense that they basically function relationally or inter-discursively. Unlike
151
152
Language and interaction
lexicalized indexicals, they have no phonetic substance of their own and hence, are not meaningful when isolated from the stream of talk. In other words, we cannot, for example, treat stress, rhythm, intonation contours as having meanings of their own. In a way, indirect indexicals are a bit like Benjamin Lee Whorf’s covert categories. They readily lend themselves to ideological re-interpretation or displacement, such that individual speakers become invested with the attitudes, abilities (or presumed lack thereof), stereotypically attributed to the group or occupational category indexed by their ways of speaking. Such ideological displacements abound in everyday life, particularly in today’s increasingly diverse urban societies. Empirical studies on intercultural and interethnic communication in classroom teaching, performance or ability assessments, counselling sessions, courtroom testimonies and so on, document the eVect they can have on individuals’ life chances (see, for example, Gumperz 1982; Jacquemet 1996; Young 1994). From a more general theoretical perspective, treating prosody, rhythm and style shifts as information-bearing indexicals calls attention to the pervasive importance of indexicality in everyday communication and suggests new, replicable ways of recovering the unverbalized assumptions that underlie what is actually said. Last but not least. Since basic indirect indexical practices are acquired very early in a child’s life, they are diYcult to unlearn. Even speakers who speak a second language well often show traces that remind us of possible historical origins of these practices (Gumperz 1996:390). Relatively little work has been done exploring the import of this fact. Additional comparative studies might show that more systematic comparative studies of contextualization-based inference lend a historical dimension to Interactional Sociolinguistics (IS) that other related traditions of discourse analysis lack. C.L.P: For those who could not attend the IPrA Congress in Budapest, and/or the Special Event Panel dedicated to you (July 10, 2000), could you sum up your Wnal ReXections? J.J.G.: Many of the papers dealt with the authors’ own work. After the panel presentations, there was time for only a few minutes of oral comments. I have no record of my comments. Paul J.Thibault: In the Discussion, you state your belief that “interaction at the level of discursive or communicative practices […] must be seen as separate
Continuing the discussion with John J. Gumperz
from either linguistic or socio-cultural processes” (Prevignano and di Luzio, this volume: 9). Could you clarify the nature of the relationship, as you envisage it, between discursive practices, on the one hand, and linguistic and socio-cultural practices, on the other? Do socio-cultural groups and their associated practices have a prior existence with respect to the ways in which they are constituted, construed and evaluated through language and other semiotic activities, material processes, and related objects and artifacts? In your view, is the distinction an analytical or a constitutive one? J.J.G.: As it stands, the statement you refer to is somewhat confusing. What I meant to suggest is that Wndings based on CP analysis are distinct from Wndings based on either linguistic or socio-cultural analysis, since these forms of analysis build on diVerent kinds of source data and rely on diVerent modes of validation. I therefore suggested that the three dimensions are not comparable. As I implied in the answer to Prevignano’s second question, I see linguistic and socio-cultural knowledge as analytically distinct, yet constitutive elements of communicative practice. As for the second part of your question, let me repeat: what is of primary importance from an IS perspective, is what we believe about such groups and how these beliefs aVect interpretation, not whether or not the groups have a “prior existence”. P.J.T.: In your Response essay, you object to the claim (that I also make in my second contribution to this volume (Thibault, this volume: 138)) that language and other semiotic modalities construe experience. In your view, the agency is located at the level of the individual social actors who engage in “shared inferencing”. But given that language and other semiotic modalities exist on space-time scales that go beyond the “shared inferencing of human actors”, seen as “the locus of interpretive assessment” (Gumperz, this volume: 121), don’t you think that larger space-time scales are also implicated and that agency is distributed across these, rather than being located in social actors per se? How do you see the relationship between the space-time scale on which actors engage in shared inferencing and the larger spatial and temporal scales involved? J.J.G.: What I intended to convey in the Response essay was that from an IS perspective, participants construe experience via interpretive assessments. This does not mean that I am taking a radical constructionist or postmodernist position. I argue that what you call knowledge [of linguistic struc-
153
154 Language and interaction
tures] existing on “larger spatial and temporal scales” has its origin in conventions that arise in interaction and over time have become grammaticalized. These may then be brought into interaction by participants as part of the interpretation process. Like other grammatical knowledge, these conventions are subject to preservation, reaYrmation or change through what happens in later communicative practice. I do of course agree that we need to make a distinction between what I would call “interactional time” or, as you put it, “the space-time scale” of the interactional exchange and “larger spatial and temporal scales”. But what I have to say does not really contradict that. Your remarks concerning my objection to the claim that “language and other semiotic modalities construe experience” go to the heart of the diVerences between your and the IS approach. Your systemic perspective leads you to privilege formal semiotic structure as a source of insight into the nature of linguistic processes. IS, on the other hand, argues that we have no access to experiential reality, except as it is Wltered through our own or some other human interpreter’s mind, so that we must look to communicative practice. This is why I emphasize shared inferencing. The very fact that interactants maintain conversational involvement over a string of movecountermove-rejoinder sequences constitutes empirical proof that they share the presupposed grammatical and indexical knowledge, since the inferential process constitutes the cognitive space where grammatical and indexical knowledge meet. But this does not mean that “language and other semiotic modalities” existing “on space-time scales” that go beyond the “shared inferencing of human actors” are not relevant for IS. What IS suggests is that we should not assume there is a one-to-one relationship between the linguist’s bodies of abstract rules and regularities and what goes into shared inferencing. The relationship between the two needs to be investigated in specially-designed studies. Until this is done, the two approaches will remain largely incommensurable. C.L.P.: Many years ago, you proposed the interaction-oriented notions of “linguistic repertoire”, “linguistic community”, and later, “social network”. How do you see these today? J.J.G.: The notion of “speech community” or “linguistic community” (the two terms were often used interchangeably, but see my Wnal remarks), has always been rather loosely employed to refer to the social space in which language is
Continuing the discussion with John J. Gumperz
spoken and with respect to which it is evaluated. In what follows, I will concentrate on my own use of the concept. I Wrst became aware of the deWnitional problems that can arise in the course of my dissertation research on the Swabian dialect of Washtenaw County, Michigan. A number of results emerged: (a) a survey of local second and third generation dialect users showed that this relatively small population, normally seen by outsiders as a single community, was divisible into three groups or subgroups, each with its own distinct speech variety; (b) a comparison with the dialect situation in the country of origin showed that individual speakers’ language use did not directly reXect the ancestral dialect; (c) local speech varieties had undergone signiWcant change in the two or three generations of residence in the United States; and (d) the main factor determining current language use was membership in one or another of the three local Lutheran church groups which also served as the main centers of local social life. My postdoctoral research in India followed the then generally-accepted practice of taking the village community as the initial universe of description and analysis. But when I looked at the communicative bases of community membership, it soon became evident that the dialect isoglosses I discovered did not follow caste and community boundaries. Finally, in the course of my stay in India, I became aware of the multilingual communities in South and South East Asia, such as those studied in Edmund Leach’s Hpalang in Highland Burma (in Leach 1954), where four grammatically distinct and mutually unintelligible languages are spoken in what, on socio-political grounds, must count as a single community. Based on such data, I argued that while a speech community, if it is to function, must constitute a communicative whole, we cannot tie the deWnition to one or more particular languages or varieties thereof. The 1968 deWnition of the speech community (“any human aggregate characterized by regular and frequent interaction by means of a shared body of verbal signs and set oV from similar aggregates by signiWcant diVerences in language usage” (Gumperz 1968:381)), which was widely discussed at the time, reXects this position. I adopted the term “linguistic repertoire” to account for intra-community diversity. A repertoire was said to consist of a set of speech varieties which constitute the totality of the community’s linguistic resources. The assumption was that members of a community choose among the varieties of the
155
156 Language and interaction
repertoire, depending on what is appropriate in a particular context of speaking. When speakers regularly alternate among varieties of a repertoire, they set in motion processes of linguistic convergence which gives the local group a linguistic identity of its own that sets it oV from the surroundings. The Kupwar study referred to in the Response essay describes an extreme case of change where local varieties of genetically-unrelated languages come to converge through regular and frequent contact over several hundred years. In hindsight, the above considerations can be seen as initial steps towards a socially-constituted sociolinguistics that accounts for shared systems of verbal signs without, however, privileging speciWc languages or varieties thereof. The exclusive focus on ongoing communication in the discussion so far is also limiting in important respects. I became aware of this Wrst in the work with J.-P. Blom in Northern Norway, where we noticed a linguistically signiWcant distinction between those who saw themselves as members of the “local team” of friends and neighbors, and those who, for professional reasons, also maintained regular relations with outside members of the northern Norwegian commercial elite. What we were talking about in this case were networks of relationships, although at the time we did not overtly state this. The network concept becomes central in the research in the Slovenian/German border area (Discourse Strategies, Chapter 3), where I argued that Slovenian is disappearing because of restructuring of social networks caused by the economic changes in the region. Had I focussed just on intra-community relations, I could not have explained the changes that concerned me, since the notion “speech community” denotes a more or less bounded group, or one of a series of such entities. A network, by contrast, can and often does denote relationships not necessarily limited to any one locality. The term “network” is by now widely used in sociolinguistic research. But whereas most sociolinguists tend to treat networks as associations of individuals, my own work, as illustrated in the Introduction to section 4 of Rethinking Linguistic Relativity (Gumperz and Levinson (eds.) 1996), sees networks as constituted by social relationships established over time, which are linguistically marked and indexically associated with speciWc communicative goals which enter into the evaluation of what is said. A Wnal, serious limitation of my earlier work is its lack of attention to linguistic ideology. In a way, this notion is implicit to the argument in the introduction to Ferguson and Gumperz’s Linguistic Diversity in South Asia (1960), where we drew a distinction between languages as “socially-deWned”
Continuing the discussion with John J. Gumperz
units — as we put it at the time — and speech varieties as “linguisticallydeWned” units. We cited a number of cases, where what I would now call “linguistic ideology” led individuals to identify themselves as speakers of one language, whereas grammatically their speech counts as a variety of another. The workings of ideology are clearly brought out in Michael Silverstein’s Monoglot ‘Standard’ in America: Standardization and Metaphors of Linguistic Hegemony (1996). Silverstein also makes an important distinction between speech communities “sharing a set of norms and regularities for interaction by means of languages” and linguistic communities deWned as “a group of people who, in their implicit sense of the regularities of linguistic usage, are united in adherence to the idea that there exists a functionally diVerentiated norm for using their ‘language’ [...], the inclusive range of which the best language users are believed to have mastered in the appropriate way” (1996:285). These deWnitions are the most useful I have seen to date, since they account for the pervasive normativeness and linguistic hegemony characteristic of our current nation states. But there is no space to discuss this here. C.L.P.: A question concerning Gumperz as an analyst of interaction, with a preamble. The idea of inference that is at the center of your interpretive approach to human “interaction order” connects culturally speciWc “practical reasoning” and presuppositions with interaction surface and cue processing. Cross-inferencing and misinteractions are possible just because conversations are not simply exchanges of cues corresponding to behaviouristic patterns of cue/response and response/cue. Several studies seem, on the contrary, to persist in adopting an implicit behaviouristic approach to cue- and turnprocessing, while you have been proclaiming (at least since 1982) the relevance of interactional and conversational “strategies” in your analyses, with a focus on members’ or conversationalists’ interpretive, inferential procedures. Starting from some Gumperzian cues, which I have reconsidered in relation to human interaction protocols and ascribed to the class of interaction cues, I have tried to identify as “minims” (Prevignano, this volume) some default rules governing the human use of activity-type cues in both output (cue emission) and input (cue reception and interpretation). In my view of cue processing, minims correspond to minimal tactical requirements for Wnding economical human solutions to everyday interaction problems; these solutions would be much more diYcult and time-consuming, not only without cues, but also without a minim-governed use of them. What is at stake
157
158 Language and interaction
here, as in the study of other human interactional tasks and abilities, is an idea of interactional intelligence (representable as a cultural set of strategies, tactics, heuristics, default rules, meta-rules, and so on), as has been anticipated and practiced by your IS research over a long period of time. The construct of interactional intelligence is necessary, in my opinion, in order to deal with the study of human, interactionally intelligent conversationalists. After this long premise, I come to a short question: could you, as a de facto analyst of interactional intelligence in and between diVerent cultures, give some suggestions for future research on it? J.J.G.: Let me restate parts of the question’s preamble in my own terms, expanding on what I understand the questioner to have said and on arguments I made elsewhere in this volume. A. As to interaction analysis. Along with other Linguistic Anthropologists, I assume that we, that is, specialists as well as lay persons, have no direct access to physical, social or psychic reality, except as it is processed by some interpreter’s mind. It is this processing, as reXected in the inferences participants draw on, on the basis of what they believe they see, hear, recall, or otherwise perceive in speciWc sets of circumstance, that lies at the center of IS analysis. So that the focus is no longer on word-to-world relationships, but on the ‘how’ of interpretation, i.e. the procedures we and others rely on to assess experience. B. For analytical purposes, it is useful to treat the inferential process as made up of several stages. In the Wrst of these stages — the most basic one in the sense that it underlies further, higher-level inferencing — perceptions are (among other things) chunked, categorized into event types and phrases and assigned phonetic, prosodic, rhythmic values. This yields what, using a modiWed version of Silverstein’s (1993:36–8) term, I call an “interactional text”. As I use the term, the notion of interactional text makes explicit the assumptions of lay persons and analysts about what they perceive, and at the same time provides the indexical ground for further interpretations. Unlike Saussurian analysis where etic constituents of talk are seen as part of the physical world, an IS approach treats what the preamble calls “interaction surface and cues” as components of interactional texts and thus as interactively constructed. C. As to “cross-inferencing and misinteractions” (I prefer the term “interaction breakdown” for the latter). My research shows that these are frequently
Continuing the discussion with John J. Gumperz 159
due to interpretive diVerences at the level of interactional text. The example “Ahma git me a gig”, cited by Prevignano (this volume: 19–20), is a case in point. When I Wrst looked at the example, I simply did not hear the singsong intonation which later proved to be crucial for the interpretation. It was only through interaction and direct questioning of my initial perceptions that I became sensitized to the contouring and its communicative import. After that I gradually became aware of its general semiotic import by transcribing and analyzing other similar materials. D. Turning now to the actual question about interactional intelligence (personally, I prefer a more modest term like “interactional skills” or perhaps “interactional competence”). You are right, my work raises the question of what we have to know to signal active participation in interaction, and I believe that in today’s increasingly diverse societies, where cultural boundaries are frequently not readily perceptible, this type of work is becoming increasingly important. What I believe my work does is to suggest some initial insights into the relevant issues and to propose theoretically-grounded empirical methods of attack. But so far the issue has not been systematically topicalized as a main object of research. Perhaps your term “interactional intelligence” might help as an attention-raiser. In any case, we need more comparative research contrasting how people signal participation in interactions where communicative background is shared, with interactions where sharing cannot be taken for granted, in societies around the world. P.J.T.: How do you anticipate future developments of IS in the light of current trends in the Weld? How do you see IS in relationship to other schools with similar interests? J.J.G.: Let me begin with the second part of the question. I have already referred to the diVerences between IS and Conversational Analysis in earlier parts of this volume. There are a number of schools that share the IS interest in a linguistic analysis that can illuminate public issues in today’s societies. One that has received a great deal of attention recently is Critical Discourse Analysis (CDA), as reXected, for example, in the work of Norman Fairclough (1989), Carmen Caldas-Coulthard and Malcolm Coulthard (1996), and Ruth Wodak (1997). Like IS, CDA is grounded in talk and pays attention to sequential positioning of moves. But beyond that, CDA seeks to account for societal factors like ideology, power and inequality, whether or not these are reXected in talk. In a separate development, Linguistic Anthropologists in the U.S. have
160 Language and interaction
also increasingly begun to turn to discourse analysis to document the role of culture and social organization in human communication. An excellent example here is Judith Irvine’s (1993) study of insulting practices among the Wolof, which shows how locals use their knowledge of social structures to manipulate roles in order to (among other things) dissociate a speaker from responsibility for an insult. What distinguishes IS from these traditions is its attempt to develop a general analytical frame capable of dealing with communicative situations of all kinds, which takes diversity of language and culture as a given. As to part one of the question, I have already pointed out the need for more general treatments of conversational inference and its import for understanding our current, inherently diverse environments. There is also a need for more systematic, interactively-based studies of the workings of linguistic ideology in human interpretation. I am currently cooperating with Jenny Cook-Gumperz on a paper on the ideology of standardization in modern nation states and its eVect on educational policy.
References Caldas-Coulthard, Carmen R. and Coulthard, Malcolm. 1996. Texts and Practices: Readings in Critical Discourse Analysis. London and New York: Routledge. Fairclough, Norman. 1989. Language and Power. [Language in social life series]. London: Longman. Ferguson, Charles and Gumperz, John J. (eds.). 1960. Linguistic Diversity in South Asia. International Journal of American Linguistics 26 (3), special issue. Gumperz, John J. 1968. “The speech community”. In International Encyclopedia of the Social Sciences, J. Sills (ed.), 9:381–386. New York: Crowell Collier and Macmillan. Gumperz, John J. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Gumperz, John J. 1992. “Contextualization and understanding”. In Rethinking Context: Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 229–252. Cambridge: Cambridge University Press. Gumperz, John J. 1996. “The linguistic and cultural relativity of conversational inference”. In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 374–406. Cambridge: Cambridge University Press. Gumperz, John J. 2002. “Sharing common ground”. In Soziale Welten und Kommunikative Style: Festschrift für Werner Kallmeyer zum 60. Geburtstag, I. Keim and W. Schütte (eds.), 47–56. Tübingen: Gunter Narr. Gumperz, John J. “Response essay”. This volume.
Continuing the discussion with John J. Gumperz
Gumperz, John J. and Levinson, Stephen C. (eds.), 1996. Rethinking Linguistic Relativity. Cambridge: Cambridge University Press. Hanks, William F. 1996. Language and Communicative Practices [Critical Essays in Anthropology Series]. Boulder, CO: Westview Press. Irvine, Judith T. 1993. “Insult and responsibility:Verbal abuse in a Wolof village”. In Responsibility and Evidence in Oral Discourse, J. Hill and J.T. Irvine (eds.), 105–134. Cambridge: Cambridge University Press. Jacquemet, Marco. 1996. Credibility in Court: Communicative Practices in the Camorra Trials. Cambridge: Cambridge University Press. Leach, Edmund R. 1954. Political Systems of Highland Burma. London: G. Bell and Sons. Prevignano, Carlo L. “On Gumperz and the minims of interaction”. This volume. Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This volume. Silverstein, Michael. 1992. “The indeterminacy of contextualization: When is enough enough?” In The Contextualization of Language, P. Auer and A. di Luzio, 55–76. Amsterdam and Philadelphia: John Benjamins. Silverstein, Michael. 1993. “Metapragmatic discourse and metapragmatic function”. In ReXexive Language: Reported speech and Metapragmatics, J.A. Lucy (ed.), 33–58. Cambridge: Cambridge University Press. Silverstein, Michael. 1996. “Monoglot ‘standard’ in America: Standardization and metaphors of linguistic hegemony”. In The Matrix of Language: Contemporary Linguistic Anthropology, D. Brenneis and R.K.S. Macaulay (eds.), 284–306. Boulder, CO: Westview Press. Thibault, Paul J, “Body dynamics, social meaning-making and scale heterogeneity: Reconsidering contextualization cues and language as mixed-mode semiosis”. This volume. Wodak, Ruth. 1997. Disorders of Discourse [Real Language Series]. London: Longman. Young, Linda W.L. 1994. Cross-talk and Culture in Sino-American Communication. Cambridge: Cambridge University Press.
161
162 Language and interaction
Bio-bibliographical note 163
Bio-bibliographical note John J. Gumperz was born in 1922 in Germany, and emigrated to the United States in 1939. Currently Professor Emeritus of Anthropology at the University of California, Berkeley, in 1954 he was awarded his Ph.D. in German Linguistics by the University of Michigan, after Wrst completing a B.A. in Chemistry at the University of Ohio, Cincinnati in 1947. His doctoral thesis, which was a study of the Swabian dialect of third-generation farmers in Washtenaw County, Michigan, marked the beginning of his interest in sociolinguistics. In this study, he formulated his early thinking on the relations between social group membership and sociolinguistic processes. This was the Wrst of a long series of research projects and Weld studies. On completion of his doctorate, he undertook two years of Weldwork in India. He subsequently founded the Hindi-Urdu program and was later Chairman of the Center for South and South East Asia Studies (1968–71), both at Berkeley. In 1956, Gumperz began collaborating with Charles A. Ferguson when both were visiting faculty at Deccan College. The ensuing exchange of ideas on linguistic diversity and language development led to their organizing a symposium at the 1958 annual meeting of the American Anthropological Association. The jointly published volume, Linguistic Diversity in South Asia (1960), represents a landmark in the Weld. Since 1966, Gumperz has been an active member of the Committee on Sociolinguistics of the Social Science Research Council. This work led to the publication of A Field Manual for Cross-Cultural Study of the Acquisition of Communicative Competence (1967), which has played an important role in the development of cross-cultural research on sociolinguistic processes. He also collaborated with Dell Hymes, Susan Ervin-Tripp, Edward T. Hall, Erving GoVman, Charles O. Frake, and William Labov in the organization of symposia at the spring meeting of the Kroeber Anthropological Society and the annual meeting of the American Anthropological Association. The seminal volume, The Ethnography of Communication (1964), edited by Hymes and Gumperz, was the product of these meetings. Gumperz Wrst took up the position of Professor in Anthropology at Berkeley in 1965. In 1992, he was awarded a Dr. Hon. Causa by the University of Konstanz in Germany. He was made Fellow of the American Academy of Arts and Sciences (1992), the Max Planck Project Group in Cognitive Anthropology in Berlin (1991), Churchill College, University of Cambridge (1987–88), as well as being a Senior Post-doctoral Fellow at the National Science Foundation (1961–62). He was LSA Professor at the Linguistic Institute of America, Georgetown University (1985). He has held Visiting Professorships at the Universities of Bergen (1982) and Konstanz (1985). Professor Gumperz’s major research interests lie in the areas of interactional sociolinguistics, linguistic anthropology, social anthropology, ethnography of communication, cross-cultural and interethnic communication, and bilingualism. His major publications include Rethinking Linguistic Relativity, co-edited with Stephen Levinson (Cambridge 1996); Engager la Conversation (Paris 1989); Sociolinguistique
164 Bio-bibliographical note
Interactionnelle: Une Approche Interprétative (Paris 1989); Discourse Strategies (Cambridge 1982); Language and Social Identity, edited (Cambridge 1982); Directions in Sociolinguistics, co-edited with Dell Hymes (New York 1972); Language in Social Groups (Stanford 1971); Conversational Hindi-Urdu, with June Rumery (Berkeley 1963). He is also Editor of the series, Interactional Sociolinguistics, at Cambridge University Press. His work in progress includes two book-length studies, Intercultural Encounters and Language in Social Theory. A bibliography of John J. Gumperz can be found at http://sunsite.berkeley.edu// Anthro/gumperz/gumppub.html
Bio-bibliographical note 165
Subject Index accenting 91, 99 n. 14, n. 15 act/action/activity typologies/tokenologies 64-65 activity 8, 13, 14, 15, 86, 93, 99, n. 2 activity type 13, 14, 15, 82 – analysis 72 – application 63, 65 – approach 72 – cues 63 – minims 63 – theory 72 Afro-American English 20 ambiguity 17, 115 analogic communication 129 background information 50, 53, 55 background knowledge 4 Bahktinian situation 94 Black English 115 bodily dynamics 140 boundary conditions 138 central/peripheral information 36 central/peripheral processing 68 clause grammar 52 co-construction 97 code-switching 3, 31, 90, 108-109, 112, 114 cognitive saliency 35, 36 common ground 98 communication analysis vii, 66 communication (or communicative) ecology vii, 86, 89 communication failure ix, 5, 89, 92-93 communicative action 86, 88 communicative competence 93-94, 97, 100, n. 21, n. 25 communicative experience 87, 89, 93, 98 communicative intent 11, 86, 142 communicative (or conversational)
involvement 8, 9, 14, 100 n. 21, 122, 154 communicative (or discursive) practices 8, 9, 11, 18, 41, 42, 94, 98, 112, 114, 120, 121, 143, 150, 151 community 41, 54, 82, 109-110, 119, 120, 121, 135-136, 143, 151, 154, 155 —see also speech communities competence 3, 136 compounding 93, 95 computational linguistics ix, 79-80 – applied computational linguistics 83 Computer-Supported Collaborative Work (CSCW) 80, 83 n. 2 context 3, 4, 6, 22, 42, 49, 86, 88, 90, 93, 98 contextual redundancy 49 contextualization 4, 57 contextualization conventions 93 contextualization cues vii, viii, 4, 5, 9, 10, 16, 32, 33, 34, 36, 38, 45, 58, 82, 88, 112, 113, 114, 127-129, 131, 133, 139141, 145 – as pure indexicals 9, 14 – see also cues, Gumperzian contextualization strategies 93-94 contrast sets 56, 130 convention 15-16, 26, 86-87, 93, 98, 154 conventional implicatures 34 conversation analysis 8, 11, 14, 86, 94-95, 117-118 conversational assessment 11, 19, 21, 118, 121 conversational inferences, inferencing 10, 11, 12, 19, 86-88, 90, 93, 95-97, 100 n. 26, n. 27, 112, 117-118 conversational management 86, 90, 99 n. 3 conversational repair 17, 93, 95-96 cooperation 93, 97, 100 n. 27 corpus linguistics 21 corpus pragmatics 21, 83, 120
166 Subject index
correlational methods 7 Croatian 106 cross-examination 88-89, 94-97, 99 n. 2 cue analysis 68ff. cue processing 68, 157 cueing 68 cues, Gumperzian 35, 67, 69 culture 85, 86, 89 cultural knowledge 89, 99 n. 9 deictic categories 48 dialectal variability 1 dialectology 1 dialogue 19, 81 – and reasoning 7, 19ff, 81 digital communication 129 diversity x, 1, 85, 98 Dutch 106 ecosocial environment x English 25, 48, 130, 131 entextualization 120 envisionments 13 ethnography 3, 89-90 ethnomethodology 86 face-threatening act 95 frames 4, 10, 11, 13, 27, 37, 59 n. 6, 93 framing 14, 27 French 106 genre 3, 15, 28, 43, 44, 54, 56, 59 n. 6 German 106 German Swabian dialect, see Swabian Gricean analysis 86 Gumperz level (of interaction analysis) 6768 habitus 55, 141, 144 Head noun 48 heteroglossia, social 54 heteroglossic diversity 43 Hindi 4, 100 n. 18 historical method 11
illocutionary force 86 implicatures 35, 86, 88 indexes, referential 87 indexical practices, indirect 152 indexicality 57, 87, 97, 113, 152 indexicals (or indexical signs) 9, 10, 11, 4546, 49, 87-88, 95, 100 n. 20, 113, 121, 123, 139-142, 145, 150, 151 indexicals, pure or indirect 88, 151, 152 inference 4, 86, 93, 142 interaction analysis vii interaction, human ix, x, 63, 69, 88, 94, 122, 157 – as a multitrack space 69 – an inferential model of 72 interaction, robust 80 interaction order 8, 157 interaction principle, semiotic 67-68 interactional competence 97, 100 n. 25 interactional intelligence 72, 158, 159 interactional linguistics 70 interactional sociolinguistics (IS) ix, x, 3, 6, 31, 85-103, 118 interactional sociolinguistic analysis (IS analysis) 86, 88, 90, 98-99, 117, 118-119 —procedures of 11-12, 86, 88-97 interactional texts 12, 90, 99 n. 12, 120, 158 interactive norms 92 interactive practices 97, 100 n. 25 interactive strategies 97-98 intercultural communication ix, 5, 89-90, 92-93, 98 intertextual formation 144 intertextuality 53 intonation 4, 47, 58-59 n. 5, 131, 137 isochrony 96-97, 100 n. 24 Italian 15, 106 job interview 88-94, 99 n. 2, n. 13, n. 14 Kannada 2, 108, 109 language as particle, wave, and field 132
Subject index 167
language shift 1, 86 language system 136 language variation 1 language variety 2 langue 3, 43 lexicogrammar 43, 44, 53, 131-133, 145 linguistic ideology 156-157, 160 linguistic repertoire 3, 110, 155-156 mapping 91 Marathi 2, 108 meaning – context-dependent 45 – context-independent 45, 47, 51 – referential 47 meaning-making practices – indexical viii, 50, 51-53, 58 – intertextual viii, 50, 53-57, 58 – metadiscursive viii, 50, 56-58 meaning potential 48, 52 metacommunication 123 metafunction 132 metaphoric switching 112 metapragmatics 59 n. 6, 87, 99 n. 6 metapragmatic assessment 15 metapragmatic function 100 n. 20 metapragmatic strategies 19, 20 metasystem 56-57 minimists/maximists viii, 63 minims vii, 63, 64, 122, 157 – input minims 63-64 – interaction minims viii, ix, 63-64 – output minims 63-64 – participation minims 64 minims/maxims 122 misunderstanding ix, 66, 89, 91-93, 98, 100 n. 19 mixed-mode semiosis x, 129, 133-134 move 12, 13, 14, 64, 65, 87 multimodal 58 multimodal semiotic performance 44 native interpreter 24 Natural Language Processing (NLP) 79-80, 83
nominal group 48, 130 noncooperation 91, 93, 95, 97 nonlinear constraint 144 norms, social 2, 3 Norwegian, Standard 108 noun (see nominal group) parole 3 participant observation 89, 99 n. 11 participation analysis 72 performance 3, 136 phonology 131 Piedmontese 25 pleading 92 postdictions 71 postmodern condition 66 postmodernist conclusion 38, 65 power asymmetry 5 power structure 5 pragmatic creativity 21, 25-27 preference organization 14 Principle of Alternation 133-134 prosody 44, 88, 90-91, 96-97, 100 n. 16, n. 18, n. 24, 114, 131, 139, 141, 142, 145 Punjabi 2 reconstruction, ethnographic ix recording, audio/video 90 repair, see conversational repair rhetorical effectiveness 98 rhetorical strategies 97 rhythm 4, 44, 131, 139, 141, 145 rhythmic organization 90, 96-97 Romansch 25 scales, organizational 134, 137 scalar heterogeneity 139 scalar hierarchy 138 scenario 13 scene 13 script 66, 72, 83 scripticals 66 scriptionary 72 second language learning 97-99
168 Subject index
second language teaching 97-99 semiotic action formation 43, 44 semiotic acts — content-specific 51 — type-specific 52 semiotic functional analysis 87 semiotic resource systems 9, 42, 43, 51 sequential analysis 11, 86, 89, 94 sequential ordering (of an interaction) 9, 88, 93, 95 Serbian 106 shifters 87 social identity 5 social network 18, 37, 135, 156 social voice 143-144 socio-cognitivism, Gumperzian 72 sociolinguistic co-occurrence rules 2, 3 space-time scales 142 speech act strategy 54 speech activities 3, 4 speech communities 2, 93 speech events 82, 90, 97-98 stereotypes 93-94 stigmatization 93-94
structuralism 8, 43 style, rhetorical 4 supersubject 94 symbolic signs 45-46, 87, 88, 145 Swabian vii, 1, 155 Telugu 108 three-part exchange 87, 99 n. 5 topic 92-94 topological-continuous variation 46, 129 transcription 22, 90, 96, 99 n. 12, 100 n. 23 type-category 48, 51, 130 typification 15-16, 89, 99 n. 8 typological-categorial (digital) distinctions 45-46, 129 Ungarinyin 58 n 4 Urdu 2, 108, 109 value 43 variables, linguistic 2 word-object correspondence relation 45
Subject index 169
Author Index Akinnaso, Niyi 17-18 Auer, Peter 26, 67, 70, 71, 100 n. 24 Aulakh, Gurinder 99 n. 4, 100 n. 18 Bakhtin, Mikhail 43, 54, 55, 56, 59 n. 6, 94, 143, 144 Ballim, Afzal ix, 66, 81, 118, 119, 120 Barwise, Jon 82 Bateson, Gregory 9, 27, 36, 46, 55, 67, 69, 129 Bazzanella, Carla 69 Berenz, Norine 72, 89, 90, 99 n. 10, 99 n. 12, 120 Berge, Kjell L. 65 Berger, Charles R. 67, 72 Blom, Jan-Petter 31, 66, 72, 108, 112, 156 Bloomfield, Leonard 8, 106 Blum-Kulka, Shoshana 66 Bolinger, Dwight L. 58 n. 5 Bourdieu, Pierre 5, 8 Bradac, James J. 67 Brown, Penelope 38 n. 1, 68, 69 Cacioppo, John 68 Caldas-Coulthard, Carmen 159 Cassell, Justine 72 Castelfranchi, Cristiano 64, 68 Chomsky, Noam 136, 149 Clark, Herbert H. 72, 99 n. 9, 120 Coakes, Elayne 83 n. 2 Cook-Gumperz, Jenny 70, 100 n. 27, 115, 119, 123, 160 Coulthard, Malcolm 159 Couper-Kuhlen, Elizabeth 70, 73 n. 3, 100 n. 24 Cowley, Stephen J. 137, 138 Dascalu Jinga, Laurentia 69 Di Luzio, Aldo vii, viii, 14, 15, 26, 28, 31,
41, 42, 44, 50, 53, 64, 67, 70, 71, 93, 110, 112, 115, 127, 133, 135, 139, 142, 143, 145, 153 Drew, Paul 27, 89, 100 n. 23 Du Bois, John W. 124 n. 1 Duranti, Alessandro 33, 63, 72 Eerdmans, Susan L. ix, 29 n. 1, n. 3, 114, 115, 117, 118 Ensink, Titus 70 Erickson, Frederick 67, 68 Fairclough, Norman 159 Ferguson, Charles 156 Field, Margaret 100 n. 27, 119 Fillmore, Charles 13, 32 Firth, John R. 131, 133 Foppa, Klaus 27 Fraser, Colin 68 Gal, Susan 38 n. 1 Galilei, Galileo 66 Garfinkel, Harold 8, 10, 11, 19, 59 n. 6, 67, 68, 72, 86, 119 Giles, Howard 67, 68 Goddard, Cliff 83 n. 2 Goffman, Erving 8, 10, 13, 14, 19, 27, 56, 59 n. 6, 67, 72, 87, 114 Gomez-Imbert, Elsa 108 Goodwin, Charles 72, 86 Goodwin, Marjorie H. 72 Greif, Irene 83 n. 2 Grice, H. Paul viii, 26, 32, 35, 63, 72, 86, 117 Grosjean, Michèle 64 Grosz, Barbara 82 Grundy, Peter 72 Gumperz, John J. vii, viii, ix, x, xi, 1-6, 7-10, 10-14, 15, 17-18, 19-21, 22-23, 24-25, 26-
170 Author index
27, 27-28, 29 n. 1, 29 n. 2, 29 n. 3, 31, 32, 33, 35, 37, 38, 41, 42, 43, 44, 45, 46, 49, 50, 53, 54, 55, 56, 57, 58, 59 n. 6, 63-73, 79, 81, 82, 85-100, 107-111, 112, 115, 116, 118, 121, 123, 127, 130, 131, 133, 135, 138, 139, 140, 143, 144, 145, 149, 150, 151, 152, 153, 154, 155, 156 Hall, Joan K. 100 n. 25 Halliday, M.A.K. 43, 47, 52, 58 n. 5, 131, 132, 138 Hanks, William F. 9, 28, 112, 114, 150 He, Agnes W. 100 n. 25 Herasimchuk, Eleanor 71 Heritage, John 86 Horton, William S. 83 n. 2 Hymes, Dell 3, 72, 111, 123, 149 Ionescu-Ruxandoiu, Liliana 15, 29 n. 2 Irvine, Judith 160 Jacquemet, Marco 23, 123, 152 Jacoby, Sally 97 Jakobson, Roman O. 87 Jefferson, Gail 22, 95 Kaal, Bertie xi Kallmeyer, Werner 150 Kaltman, Hannah 99 n. 4, 100 n. 18 Kendon, Adam 67, 68 Kerbrat-Orecchioni Catherine 70 Keysar, Boaz 83 n. 2 Kiesler, Sara 83 n. 2 Koerner, E.K.F. 66 Labov, William 37 Lakoff, Robin 32 Langacker, Ronald W. 48, 51 Lave, Jean 120 Leach, Edmund 155 Leech, Geoffrey viii, 63 Lemke, Jay 41, 43, 44, 49, 51, 52, 53, 54, 55, 129, 133, 134, 136, 138, 139 Levinson, Stephen C. viii, 31, 52, 55, 64, 65,
67, 69, 70, 72, 105, 112, 114, 116, 127, 130, 156 Lucy, John A. 113, 121 Lyons, John 83 n. 2 Martin, James R. 43 Mertz, Elizabeth 99 n. 6 Müller, Frank 70 Murray, Stephen O. 73 n. 2 Nöth, Winfried 49 Ochs, Elinor 97 O’Connor, Mary C. 99 n. 4 Orletti, Franca 68 Pallotta, Vincenzo 81 Parmentier, Richard J. 99 n. 6 Parret, Herman vii Peirce, Charles S. 10, 16, 49, 59 n. 6, 69, 87, 113 Perry, John 82 Petty, Richard 68 Pike, Kenneth L. 58 n. 2, 132, 150 Prevignano, Carlo L. vii, viii, x, 7, 10, 15, 16, 17, 18, 19, 21, 22, 23, 24, 25, 26, 27, 28, 41, 42, 44, 50, 53, 64, 66, 69, 71, 72, 73 n. 2, 93, 105, 112, 114, 115, 120, 122, 127, 133, 135, 139, 142, 143, 145, 146 n. 1, 149, 151, 152, 153, 154, 157-158, 159 Raskin, Victor 72 Rawlins, William K. 69 Roberts, Celia 97, 98, 100 n. 18 Rumsey, Alan 45, 58 n. 4 Sacks, Harvey 32, 67, 72, 95 Sapir, Edward 8, 43 Sarangi, Srikant 72, 98 Saussure, Ferdinand de 8, 10, 43, 158 Schank, Roger C. 83 Schegloff, Emanuel A. 72, 95 Schiffrin, Deborah 23, 68, 72, 85, 86, 98 Schutz, Alfred 15, 99 n. 8
Author index
Searle, John R. 32 Selting, Margret 70, 96, 124 n. 1 Silverstein, Michael 36, 45, 47, 56, 59 n. 6, 87-88, 99 n. 6, 99 n. 12, 113, 120, 150, 151, 157, 158 Sinclair, John 21 Sproull, Lee 83 n. 2 Stati, Sorin 28 n. 1, 68 Svennevig, Jan 67 Szymanski, Margaret 100 n. 27, 119 Tannen, Deborah 38 n. 1 Thibault, Paul J. viii, x, 41, 43, 44, 49, 50, 51, 52, 53, 54, 55, 69, 73 n. 2, 120, 121, 128, 134, 139, 143, 144, 145, 150, 152153, 159 Thomas, Jenny 72 Toribio, Almeida 87 Tsui, Amy 87, 99 n. 5 Turing, Alan M. 72
Tyler, Andrea 100 n. 19 Uhmann, Susanne 100 n. 24 Urban, Greg 120 Van Leeuwen, Theo 44, 51, 52, 128 Vanelli, Laura 69 Vico, Giambattista 6 Voloshinov, Valentin N. 123 Weigand, Edda 17 Weizman, Elda 66 Wenger, Etienne 120 Whorf, Benjamin L. 43, 152 Wilks, Yorick 66, 81, 82 Wittgenstein, Ludwig 58 n. 1 Wodak, Ruth 159 Young, Linda 23 Young, Richard 100 n. 25
171
