No title

Applied Linguistics No.3,

Vol. I,

Autumn 1980

THEMATIC ISSUE APPLIED DISCOURSE ANALYSIS Guest Editor: J. McH. Sinclair

CONTENTS Introduction. JOHN McH. SINCLAIR You Did Say Oral Interactive Discourse? HENRY HOLEC When Communication Breaks Down: Levels of Coherence in Discourse:

Page 185 189

PHILIP RILEY

201

CHRISTIAN BACHMANN

217 224

Le Social Pese Lourd sur le Discours: un cas D'inegalite Interactionnelle. Interactional Markers in Dialogue. EDDY ROULET Conceptual and Communicative Functions in Written Discourse. H. G. WIDDOWSON

Analyse de Discours et Pragmatique de la Parole dans Quelques Usages d'une Didactique des Langues. DANIEL COSTE Some Implications of Discourse Analysis for ESP Methodology. JOHN Me H. SINCLAIR

Discourse as Self-Expression on the Reduced Personality of the Second Language Learner. PETER HARDER Some Problems Concerning the Evaluation of Foreign Language Classroom Discourse. WILLIS J. EDMONDSON

234 244 253 262 271

ACKNOWLEDGEMENTS This volume gathers together a set of papers which originated in a colloquium held in Berne from 30 May to 2 June 1979. The colloquium was sponsored by the Swiss National Foundation for Scientific Research and the University of Berne and organized by Rene Richterich of that university. Several of the papers were originally written in French, and two remain in that language, though English versions have been made. The translations were done with great care and promptitude by Philip Riley, who also helped in tracking down an elusive reference in French. The Editors and Publisher would like to thank the sponsors, the organiser and the translator, and all the contributors for making their work available and revising it for publication. The guest Editor would like to add thanks to Henry Widdowson for his helpful liaison and final reading of the manuscript.

NOTE It was the intention to include in this collection a paper by Teun. A. Van Dijk entitled 'Discourse Studies and Education' which was prepared for the Berne Colloquium but could not be presented there. Unfortunately shortage of space has obliged the Editors to defer publication of this important contribution to the next issue of Applied Linguistics.

REFERENCES All references are listed at the end of this issue, with the exception of those belonging to the article by Christian Bachmann, which appear on p. 221.

New Nelson can now provide you with

SKILLS FOR LEARNING Foundation Foundation gives students, even those with a limited knowledge of English, training and practice in techniques which will enable them to -extract important information from written texts. -profit from their reading. -enjoy success in reading even the most complicated academic text.

Foundation is designed for all students preparing for higher education. Publication dates: Foundation, plus Teacher's Book-August 1980 Development, plus Teacher's Book-January 1981 Further books in this series- 1981 For a free sample unit from Skills for Learning, Foundation please write to TheELT Dept., Thomas Nelson & Sons Ltd., Nelson House, Mayfield Road, Walton-on-Thames, Surrey KT12 5PL.

THEMATIC ISSUE APPLIED DISCOURSE ANALYSIS Edited by J. McH. SINCLAIR

CONTENTS

Introduction JOHN McH. SINCLAIR

You Did Say Oral Interactive Discourse? HENRI HOLEC When Communication Breaks Down: Levels of Coherence in Discourse PHILIP RILEY Le Social Pese Lourd sur le Discours: un cas D'inegalite Interactionnelle CHRISTIAN BACHMANN Interactional Markers in Dialogue EDDY ROULET Conceptual and Communicative Functions in Written Discourse H. G. WIDDOWSON

Analyse de Discours et Pragmatique de Ia Parole dans Quelques Usages '

d'une Didactique des Langues DANIEL COSTE

Some Implications of Discourse Analysis for ESP Methodology JOHN McH. SINCLAIR

Discourse as Self-Expression on the Reduced Personality of the Second Language Leamer PETER HARDER Some Problems Concerning the Evaluation of Foreign Language Oassroom Discourse WILLIS J. EDMONDSON

Aims

The aim of this journal is to Typescripts Contributions should promote a principled approach to be typed in double-spacing on one language education and other lan­ side of the paper. The presentation related concerns by en­ should conform as closely as possible guage couraging inquiry into the relation­ to the printed style of the journal. ship between theoretical and practical References are to be incorporated in studies. The journal is less interested the text according to the following in the ad hoc solution of particular convention: problems and more interested in the Krashen (1978) considers this handling of problems in a principled phenomenon. way, by reference to theoretical This phenomenon seem$ to support studies. the distinction between acquisition Viewing applied linguistics as a and learning (Krashen 1978). relation between theory and practice, Works referred to should be listed at the editors give priority to papers which develop specific links between the end of the article. Footnotes theoretical linguistic studies, educa­ should appear not at the foot of the relevant page but at the end of the tional research, and the planning and implementation of practical pro­ text. grammes. Within this framework, the Proofs Galley proofs will be sent to journal welcomes contributions in the author for correction and should such areas of current inquiry as first be returned to the London editor and second language learning and within ten days of receipt. Page teaching, bilingualism and bilingual proofs will be corrected by the editor. education, discourse analysis, trans­ Copyright The editors will assume lation, language testing, language that an article submitted for their teaching methodology, language consideration has not previously been planning, the study of interlanguages, published, and is not being con­ stylistics and lexicography. One of sidered for publication elsewhere, the features to be developed is a either in the submitted form or in a section of notes and discussion which modified version. Acceptance of gives an opportunity for readers to author's copyright material is on the make short and relatively informal understanding that it has been comments on the articles and reviews assigned to the Oxford University published in the journal and on other Press subject to the following con­ ditions: The author is free to use the matters of current interest in the field. Contributions Four copies of arti­ article in subsequent publications cles should be sent in typescript to written or edited by himself provided one of the three editors. If that editor that acknowledgement is made of judges that an article is to be con­ Applied Linguistics as the place of sidered for publication he will send original publication. Except for brief copies to his fellow editors and an ex­ extracts the Oxford University Press ternal reader for assessment and will not give permission to a third comment. Authors should retain one party to reproduce material from an copy to refer to when they receive article unless two months have editorial comment and for the cor­ elapsed without response from the rection of proofs. Articles should authors after the relevant application normally be in English although con­ has been made to them. tributions in other languages will be Offprints Thirty offprints of arti­ considered if they are provided with cles and reviews will be provided to authors free of charge. a substantial summary in English.

APPLIED DISCOURSE ANALYSIS: AN INTRODUCTION JOHN McH. SINCLAIR

University ofBirmingham

Applied Linguistics, Vol. I, No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

This collection of papers shows something of the wide range of activity that is pushing itself forward in the description of what I hardly dare to call discourse; and its inipact on the profession of language teaching. It shows a set of ideas on the move, starting out from origins in· sociology, linguistics, philosophy and psychology, and jostling for the attention of scholars and teachers. There is nothing fixed or static here, but there is plenty to encourage new thinking, stimulate better descriptions, and suggest innovations in classroom practice and materials. All the papers were composed around the same theme, and so the sequence I have chosen reflects emphases rather than clearly distinct topics. The earlier papers propose and discuss categories of description, with a growing attention to textual examples as we go along. The later papers pay more detailed at­ tention to the implications for_practical teaching that arise from description, although they also attend to theoretical matters. If the reader has not yet acquired the convenient disease of terminological blindness, he should feel the symptoms before very long. The different disciplines that feed discourse description, the linguistic and national con­ ventions, and the tentative nature of many of the proposals for description, all combine in a minefield of terminology. I shall tiptoe through it, but in order to say anything at all I am bound to set a few mines off, while at the same time repeating what has already been said somewhere in the papers which follow. This is European work-Western European, in fact. Although the bibliography contains many references outside the area, particularly to American sources, there were (exceptionally for this annual gathering) no transatlantic representatives. But even though no-one had to travel a thousand miles to Berne, and all knew each other's work, fluidity of the topic area led to discussions which were rich in attempts to get alignment of views. In the process of editing, I became confident that there was much more common ground than a first study might suggest. Scholars will quite rightly defend distinction A or category X until they are utterly convinced that it is either identical to someone else's, or very like it, but carrying less justification. One major feature of this kind of work is the concern to reveal the organization of language above the sentence. Although there is a case to be made for the sentence as the highest unit of grammar, there are repetitive patterns, and restrictions on occurrence, and meaningful sequences to be observed in longer stretches of language. Whether such organisation is linguistic, or how much of it is, is discussed in Harder's paper; but awareness of it certainly contributes to our understanding of the task of learning a language.

1 86

A P PLI E D D I S C O U R S E A N A LY S I S

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Speech and writing are held to b e manifestations of the same general type, in that they are said to be 'in the same language'. But in many ways they can be obviously different, and many of the papers make proposals which maintain the distinction while attempting to relate them to each other by elaborating the descriptive framework. A key development in this kind of analysis is the recognition that language is addressed from one person to another, and the instances show at least traces of this fact. In conversation there is so much attention paid to the mechanism of interaction that it can be used as the basis for description. Written and spoken texis of much greater length than the sentence are required in order to realise this important aspect of description. It could be argued from the evidence in this volume that the task has just begun; that the large-scale categories are but thinly described, and that the preponderance of descriptive work is onlyjust hovering above the sentence. Caution, however, is appropriate because the shift of focus to more ex­ tended instances of language reveals that other forces are at work, partly determining the selections made in the texts. A sentence, perhaps, can be chipped out of its co-text and displayed on a blackboard for elucidation of its construction-in fact some of the received definitions of a sentence seem to do little more than paraphrase this property. But a longer text displays its place in social intercourse, and matters concerning the people involved, the reasons for their acts of communication, the social forces that play upon them, the wider interpretations of their behaviour, cannot be excluded. Substantial problems in research methodology are raised, and while there is little evidence yet of generally accepted solutions, the authors are notably willing to recognise the consequences of broadening their perspectives of description. They are struggling to make what they say correlate, at least, with the mass of cross­ cutting issues which are currently under consideration. Only through this process, and the painstaking establishment of links between actual instances and descriptive categories, will a satisfactory comprehensive model be achieved. No particular prominence is accorded to the spoken language in this volume, but it is probably true to say that much of the work has been stimulated by the demands made on description by recent attention to dif­ ferent types of spoken verbal interaction. Spoken language is more firmly rooted in social intercourse, more difficult to isolate and observe, more problematic in its structure-at least in the present state of our knowledge. Changes in people's relationships, shared knowledge, awareness of their environment can be studied as they happen. The results of such study can then be compared with the written language, which we can credit with having similar intentions to effect changes but suffering delays in execution. An important point is emerging from the study of conversation, shown most clearly in Riley's paper, though related to the meticulous observations of Sacks, Schegloff, Jefferson and their circle. It is that the actual execution of conversations needs to be separately accounted for, and distinguished from the underlying structures that they are held to realise. The distinction between these two may not yet be quite clear. However, the intricate process by which speakers co-operate to produce conversation seems to have at least two facets. On the one hand, each utterance can be described in terms of the way in which

J O HN M c H. S INC L A I R

187

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

the speaker is trying to guide the discourse. The description is made with reference to a structural model which attempts to show the underlying generalities, representing the conversation as a series of bids to control. Since in any conversation there is more than one participant, the control moves may conflict, cancel out, converge or relate to each other in many other ways. But the conversation remains usually coherent (despite the appearance of most transcripts) because the initial contract to pursue communicative ends is maintained by the implicitly recognised structure. On the other hand, the actual ways in which turns are taken and given, addressees selected, floor held and interruptions accomplished, requires an executive model. This model in turn highlights a range of behaviour that is closely bound up with the process of utterance, for example tempo, pausing, eye-contact, posture, gesture, and several phonological features. Phenomena of ambiguity and changing tack in mid-utterance are also able to be pinned down in the description of discourse execution. Without this distinction, much of the evidence cannot be precisely assigned in a description. It must be noted, however, that neither of the above domains of patterning conflates or overlaps with the domain where the basic rules of language structure are stated-the grammar, in the broad sense of the term. Turning to the applications of discourse description in language teaching, we find that there are clear converging movements in several areas. The pragmatism of teachers, and the pressures on them to respond to frequently­ changing demands, leads to a quick, if not always orderly, transfer of insights from theory to practice. The notion of communicative skills has refreshed the profession and stimulated much controversy and some innovative teaching. The new target is to make students proficient in achieving planned results from their use of language. Their command of the rules just referred to is, of course, a limiting factor, and the techniques by which structural and communicative teaching can be made complementary to each other have not yet been devised to the satisfaction of the practitioners. Almost certainly, this problem betrays a set of deficiencies in linguistic theory and description. But communicative skill implies a sensitivity to other participants, and to the general and specific constraints on the effective use of language. It demands control over the execution of utterances in both speech and writing, and alertness to control opportunities, and maintenance of co-operative structuring. It raises levels of expectation in a physical sense, to skill in con­ ducting discourse over extended instances of production, reception or both combined. It thus profits from all the areas of descriptive activity that have been mentioned. The interest in spoken interaction meets a demand for effective teaching of oral communication, and it would be simplistic to suggest that the matching is coincidental. They stimulate each other, and feedback from the classroom can be a valuable source of insights, hints and warnings to the descriptive linguist. The study of language learning brings out psychological factors that the descriptions can then be faced with. Similarly, the advances in understanding the nature of written texts or spoken monologues finds an application in the 'specific purpose' kind of language teaching, which is a major industry worldwide. Wherever language

1 88

A P P L I ED D I S C O U R S E A N A L Y S I S

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

teaching is integrated with, involved with o r parallel o r preliminary t o the teaching of something else, there are special demands made on the student and the teacher, and hence on the descriptive support. The language being taught is liable to be carrying conceptual difficulties, or problems for the student in how to derive practical benefit from his linguistic skills. Again the perspective broadens, and severe descriptive problems arise for the analyst, who may be asked to provide a linguistic characterisation of a text which accurately reveals its potential value as an instance of language for teaching students of a remote discipline. This is Applied Linguistics with a vengeance! There is, then, a particularly lively interchange going on between the classroom and the backroom. The variety of material in this volume shows a fund of usable ideas, and brings out several important principles. It also, though, indicates that language teachers must be independent and flexible in their own attitudes if they are to profit from the continuing interchange. The subject is moving fast and variously, and there is an ever�present danger that ideas are enshrined in practice before they are thoroughly evaluated and precisely explicated. But we can see immense potential in the working-out of the themes and schemes that are presented here, and that is perhaps the main justification for presenting them to the community of applied linguists.

'YOU DID S AY 'ORAL INTERACTIVE DIS COURSE'? HENRIHOLEC

Centre de Recherches et d'Applications Pedagogiques en Langues, (CRAPEL}, Universite de Nancy II, France 1. I N T R O D U C T I O N

' ... the term "interaction" is used to describe a direct communicative exchange (the direct communicative exchange being a particular type of oral com­ munication) .' (Gremmo, Holec and Riley, 1978).

For others, both written and oral forms are interactive: 'Those working in written discourse have tended to analyse it as monologue and to ignore the fact that as he reads it the reader interacts with the text and thus an interactive model might also be appropriate to a written discourse.' (Coulthard, 1977, p. 180)

while still others, like Widdowson, suggest that 'in written discourse ... there is interactivity without interaction'. (Widdowson

1977: 259)._

Such differences of approach show clearly that the distinction between the terms interactive and non-interactive is far from having been fully elucidated and it is the aim of this article to contribute to that elucidation. Since it is probable that these differences are as much due to differing in­ terpretations of the terms 'discourse', 'oral discourse' and 'written discourse' as they are to the lack of precision in the distinction between interactive discourse and non-interactive discourse, it will be necessary first to recall just what the aim of discourse analysis is. We will then go on to propose a definition, and one that we hope is both clear and precise, of what is meant by

Applied Linguistics, Vol. I, No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Discourse analysis i s a relatively new discipline, s o i t i s hardly surprising that the discussions to which it has given rise include ambiguities, inconsistencies and even contradictions, which are due partly to a lack of precision in the definition of some of the basic concepts and partly to the inadequacy of the analytical tools which are used. But though these 'imperfections' may be sources of potential confusion, the problems they pose enable us to clarify and extend the discipline, in other words, to make progress. , One such problem concerns the notion of interaction as it is applied to the distinction between oral discourse and written discourse. For some analysts, of course, this notion is not relevant and they are satisfied to subcategorise discourse into oral discourse and written discourse and leave it at that: but this attitude is becoming increasingly rare. Amongst those who think that this dimension is relevant, there are some for whom only certain types of oral discourse are interactive:

190

O R A L I N T E R A C T I V E D I S CO U R S E

oral discourse and written dis·course. Finally, some examples will be taken for discussion which seem to throw light on the problem of interactive discourse. 2. D I S C O U R S E

3. O R AL A N D W R IT T E N D I S C O U R S E A second source o f confusion is due t o the lack o f clarity in the distinction between oral discourse and written discourse: this can be seen when we talk about a written transcription of oral discourse as still being oral discourse, or when we say that a written text read aloud is still fundamentally a written text. This shows, at least, that the way we subcategorise discourse intuitively is not based on the differences between the various communicative channels .

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

The first source o f confusion which must b e briefly considered, and which may largely explain why the notion of interaction is not universally regarded as relevant, is to be found in the definition of the very object of analysis, dis­ course. There is a triple ambiguity here: I. 'Discourse' is sometimes used as an equivalent of de Saussure's parole to refer to all the realisations of the underlying langue: in this sense, the object of analysis consists of all the utterances which speakers in fact produce, ut­ terances which are then considered in terms of their relationship with langue (e.g., the relationships between langue and oral discourse, or the relationships between langue and written discourse). 2. 'Discourse' is also used to refer to the object of analysis of what British and American linguists call 'suprasentential linguistics': Householder ( 1 972); from this point of view, discourse is any series of utterances which are not simply juxtaposed, but which form part of a higher structure, being linked by semantic cohesion (whether this cohesion is realised grammatically, lexically, non-verbally, etc.). 3. 'Discourse' is also used in a wider sense by linguists who try to analyse the communicative functions of language: Widdowson ( 1977), Sinclair and Coulthard (1975); here it is used to refer to the entity consisting of the text (verbal and non-verbal messages) and of the circumstances in which the text is produced and interpreted (who is the speaker, whom is he addressing, what is his communicative aim, etc.). For instance, a printed notice saying 'No Smoking' on the wall of a theatre is an example of discourse, but the same message, put away in a drawer, would be merely a text. From this point of view, then, the object of analysis is the text-in-situation (or the message in context and in situation) and the aim of t;1e analysis would be, for example, to identify and describe the relationships between a text and the circumstances in which it is produced, or to define the structures of texts containing several messages. In this third approach extra-textual parameters are also taken into account: indeed, the text is regarded as the product, realised verbally or non-verbally, of the psychological and sociological constraints and conventions which enter into a given situation. It is this approach alone that includes the idea of 'in­ teraction', describing the collaborative production of a text by several dif­ ferent speakers. This, then, is the approach which will be adopted in what follows, the term 'discourse' being used to desi�ate the verbal and non-verbal lext used by an actor to address his interlocutor for communicative purposes.

HENRI HOLEC

191

However, the distinction is not easy to make with precision. 1 . It is not possible to define the distinction in question by taking actual examples of the two types and by describing those features which are characteristic of Type A and those which are characteristic of Type B, to avoid oral and written for the moment. This impossibility arises because of one of

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

the two following reasons: Either (a) the characteristics described in this way are not sufficiently precise for the distinction to be applied with any satisfactory degree of rigour or in any consistent fashion: this is the consequence of faulty methodology, since we start by intuitively separating the various examples of discourse into the two main sub-categories which it is our very purpose to describe, so that when we do later get round to describing them we run the serious risk of attributing significance to features which are not necessarily exclusive to one of the two types. For example, it may be found that a given formal feature occurs in all the texts of discourse type A and is therefore regarded as specific to that type, whereas in fact it cannot be proved that it can never be found in examples of discourse type B: what is improbable has been confused with what is im­ possible. Or (b) the simultaneous description of discourse along a number of different parameters gives us a cline where it is impossible to locate one and only one point where discourse type A can be separated from-discourse type B. The description of particular examples of discourse is not, therefore, especially helpful. In the same way, psycholinguistic analyses of the differences between the encoding and decoding processes, although very illuminating concerning the ways in which texts of the two discourse types are produced and interpreted, do not, as things stand, provide an adequate basis on which to establish the distinction. An examination of the differences between discourse type A and discourse type B in terms of 'why' rather than 'how' would seem to be a more rewarding line of investigation, in that it allows us to consider discourse as a process rather than as a product. We believe then that the distinction between type A and type B can be made acceptably explicit by analysing the conditions in which the discourse is produced. 2. If we take into consideration both the dynamic dimension of the production of discourse (a product being made) and the static dimension of its realisation (a finished product), it is in fact possible to distinguish two types of discourse: (i) There is one type of discourse whose finished product, the text (a set of verbal and non-verbal messages), is constructed in real time and whose production cannot be dissociated from the linear passage of time. The structure of the text and the production of the text proceed 'abreast', con­ comitantly and in real time so that for every 'after' in production there is a corresponding 'after' in the text, the production of the text being continuous and irreversible: in this type of discourse, the text has an internal temporal structure. (ii) In the second type of discourse, the structure of the text is not directly associated with the temporal linearity of the process of production: dissociated in time from the process of construction, the text proceeds parallel to it, but

192

O R A L INTE R A CT I V E D I S C O U R S E

4. I N T E R A C T I V E D I S C O U R S E/ N O N - I N T E R A C T I V E D I S C O U R S E 4 . 1 Interaction

(a) The notion of interaction would be of little use as the basis for a descriptive category if all it meant was that discourse is a phenomenon in­ volving two or more participants influencing one another: if this were the case all discourse would be interactive by definition, as indeed would painting, sculpture, music and so on. (b) Nor can the notion of interaction be limited to the fact that one of the participants, the speaker, takes into account the presence of other par-

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

not abreast of it nor i n real time, so that a 'before' in the text may correspond to an 'after' in the process of construction: in this case, the production of the text is therefore non-continuous and reversible and the text has no internal temporal structure. In order to understand this fundamental difference between the texts in these two types of discourse, it may be helpful to compare them to pictures being drawn on two strips of paper: in one case, the strip would unwind continuously, while in the second, it would be possible to control the movement (to slow it down, stop it, to rewind, to start again, and so on). The pictures on the first strip would have an internal dynamic structure which would remain even if the strip was later unwound and read. The presence or absence of internal temporal structure has many different effects on texts. For example, to mention only one of these effects-one which is extremely important but which has never been studied systematically-texts in the first type of discourse reflect the limits of human capacities for memorising, anticipating, organising, etc . , in real time, which is obviously not the case in discourse of the second type: consequently, the content of texts of the first type tends to be organised gradually as the discourse proceeds,-very often on the basis of chronological development,-with frequent repetition, switching back etc. , and the longer the text, the more noticeable all this is. Texts of the first type are usually, but not necessarily, realised acoustically as regards their verbal component and visually as regards their non-verbal component. When this is the case, we find, in addition to the characteristics related to their temporal structure, characteristics which derive from the ephemeral nature of acoustic and visual signals. For example, the limitations on human capacities mentioned above are made even tighter in this way. Similarly, texts of the second type are usually but not necessarily realised graphically (i.e., by some lasting visual signals). One of the effects of this is an increase in the number of ways in which the text can be controlled. In addition to the freedom from temporal constraints there is the fact that the material realisation is spatial in nature (which makes it possible, for example, to begin the realisation of the text by the end as well as by the middle or the beginning). 1 3 . Consequently, to us the term 'oral' to refer to the first type of discourse and the term 'written' to refer to the second is neither arbitrary nor incorrect, but simply insufficiently explicit to prevent ambiguity and misunderstanding. Since it is not possible simply to do away with such widely-used terminology, we will have to be satisfied with underlining the fact that by 'oral discourse' we mean discourse whose text is constructed in real time, and by written discourse we mean discourse whose text is not subject to this constraint. 2

HENRI HOLEC

193

'By interactive discourse, we mean discourse which is the collaborative construct of two or more participants mutually engaged in other-directed communicative behaviour.' (Riley, 1979)

The clarity of this distinction notwithstanding, the analysis of authentic discourse does throw up a number of problematic borderline cases. For in­ stance, since the degree of participation varies from individual to individual, how are we to handle discourse involving a number of participants but in the realisation of which-in terms of speaking turns-only two actually take part: is such discourse less interactive than a discourse in which all present share the task? And what about the type of discourse which involves two participants but to which one participant makes only a single contribution, as is the case in certain interviews where the interviewer sets the ball rolling with a question and then makes no other contribution other than a brief expression of thanks at the very end. And does it make" any difference if this contribution is very long-or very short? At a more general level, is it possible to talk in terms of degrees of interaction, non-interaction being just an extreme point on the scale? Only by analysing numerous examples of authentic discourse can we hope to answer these questions. The distinction between interactive and non­ interactive as it has been defined above is important, even crucial in some ways, in the analysis and description of discourse, and it is this distinction which we are going to refine by applying it to both oral and written discourse: can both these types of discourse be both interactive and non-interactive? 4.2 Oral, non-interactive discourse There is no need to show that oral discourse can be interactive; the majority of oral discourses are, which is hardly surprising since this type of discourse is usually realised in face-to-face communicative situations which, for psycho­ social reasons, generally require the interactive participation of all present (at least in Western cultures). In certain situations, however, the oral discourse

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

ticipant(s), his interlocutor(s), either in the sense that he is guided by feedback or by his own construing of the situation: if such were the case, all discourse would be interactive and the most one would be able to do would be to analyse degrees of difference along the same parameter (type and/or degree of in­ teraction); in any case, there would no longer be grounds for the distinction between interactive and non-interactive. (c) A definition of interaction, if it is to be used in discourse analysis, should be based on an analysis of the types of role played by participants. In certain cases, the discourse is the mutual responsibility of all the par­ ticipants, each contributing in turn 3 to its realisation and collaborating with the other participants to produce a coherent series of contributions. This means that each participant constructs and realises only part of the text, contributing a number of fragments in alternation with the other participants; the stretches of discourse which are realised by these fragments of text are usually, though perhaps incorrectly, called 'speaking turns'. In other cases, a single participant is responsible for the whole of the discourse. In the first case, we are dealing with interactive discourse; in the second, non-interactive discourse.

O R A L I N T E RACT I V E D I S C O U R S E

194

which i s produced i s non-interactive: an example would b e a political speech or a lecture. As an example, then, let us take the following passage from a lecture:

There is no doubt as to whether this is oral discourse or not; the text is being constructed in the here-and-now along with the discourse; it is also non­ interacti�e. since the teacher was the only participant to take the floor during this extract (and during the three-quarters of an hour which followed it-the last 1 5 minutes of these lectures being regularly given ovrr to a discussion separated from the lecture by a long pause.). The differences between a text like this and a text from a written non­ interactive discourse on the same subject (e.g., a lesson in a textbook on linguistics or a chapter in an introduction to descriptive linguistics) are so obvious that there is no need to enter into further detail here, except, perhaps, to note that this extract is essentially a recommendation to the students to do the exercises without outside help and that this would probably have been omitted in written discourse, or would appear in a general introductory chapter on working methods. To illustrate more clearly these differences between oral non-interactive discourse and written non-interactive discourse, let us consider the two examples reproduced below. The first is an extract from a didactic discourse, showing how to use a video-cassette recorder, which we will regard as non­ interactive, since it has aU the relevant characteristics (this example is taken from Hutchinson 1 978, p. 15):4 the second is an extract from the instructions for use booklet provided by the manufacturer.

Example I First of all, of course, we must put the cassette of tape into the machine.Well on the machine the frrst button here is labelled EJECf and this means that it ejects the cassette in its holder here.We press this button we see the cassette holder comes up out of the machine and this now enables us to put the cassette into the machine. We take the cassette with the label uppermost and simply feed it into the slot which is so exposed.We push the lid down firmly, make sure it goes right down and the cassette is now ready to play.

Example2 Extractfrom the instructionsfor use.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Cette fois-ci encore, hm, nous allons travailler sur un corpus, hm ... un corpus a, er, elucider ...et ya un theine et une version assortis ...Cette fois-ci c'est du turc ...hm ...Pour le theme et Ia version ne vous ....documentez pas sur Ia question, hein, faites �tout seuls ... sans grammaire, sans ... parce qu'on peut toujours trouver des ...des gens de nationalite turque dans Ia region, hein ...alors faites �a seuls, hein, c'est pour votre ...si vous faites �seuls, pour votre formation personnelle c'est bien preferable, hein, que de ... ne pas chercher, que de vous faire communiquer des reponses ... Alors vous faites comme si vous ne ..., hm, si vous connaissez deja le turc y aura done des choses que VOUS n'aurez pas a deviner, a decouvrir par VOUS-memes .... Je vais deja vous donner le ...�a c'est le debut ...(distribue les textes polycopies) ....

HENRI HOLEC 0 Press EJECT

195

(with power on).

6 Insert cassette.

Finally, let us consider a third example, this time an extract from an oral interactive discourse on the same subject:

Example3 C. A. C. A. C. A. C. A.

Dis, tu me montres comment ca fonctionne? Quoi. . . � a Ia, le magnetoscope? Oui, j'sais plus comment on installe Ia cassette (rire) Eh bien, c'est pas difficile, eur, Ia cassette se met la-dedans, eur, dans ce compartiment. . . <;a, je le sais, mais comment on !'ouvre? Eh bien, tu appuies sur le bouton Ia, . . . le bouton marque EJECT, Ia . . . tu vois, ca s'ouvre tout seul, c'est simple . . . Ah bon. . . moi j'essayais de l'ouvrir comme �a (rire) . . . en tirant. .. en soulevant le couvercle. . . Oh non, surtout pas, malheureuse, . . . tu risques de tout casser . .. non, il faut appuyer sur le bouton Ia. . .

This conversation can be rendered in English as follows: C. A. C.

Show me how that works, will you? What. . . that, the videorecorder? Yes, I can't remember how you put the cassette in Oaugh).

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

@Press down.

196

.

A. C. A. C. A.

O R A L I N T E R A C T IV E D I S C O U R S E

Oh, that's not difficult, er, the cassette goes in there, er, in this com­ partment. Yes, I know that, but how do you open it? Well, you press that switch ...the one where it says EJECT, there ... you see, it opens by itself, it's simple. Oh, right ... l was trying to open it like this (laugh) ...pulling ... lifting the lid. No, don't do that whatever you do ...you could break the whole thing ...no, you just press that switch.

/

4 . 3 Written interactive discourse Just as oral discourse can be non-interactive, written discourse can be in­ teractive. Reproduced below is an example, consisting of an exchange of letters between the organisers of a colloquium and one of the participants. lith January 1979

M. H. Ho1ec CRAPEL Universite de Nancy II 23 Boulevard Alben 1 er S400 Nancy Dear Henri,

I am writing to invite you to participate in an Applied Linguistics Colloquium to be held

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

An examination of the two passages of oral non-interactive discourse leads us to draw two main conclusions: The flrst concerns methodology: very often the distinction between inter­ active and non-interactive is ignored in descriptions of discourse, which results in attempts to compare the incomparable or to use the same model for the description of all kinds of discourse. It is clear that at least two models are necessary, one for non-interactive discourse, whether it is written or oral, the other for interactive discourse (see below); non-interactive discourse can be described in terms of its illocutionary structure and content structure, but interactive discourse must also be described in terms of its interactive structure (Cf., Riley, 1979). The second conclusion concerns the notion of interaction: the comparison between the extracts of oral interactive and oral non-interactive discourse shows clearly that to say that discourse is interactive simply because the participants mutually influence one another, fails to bring out the very im­ portant point as to whether or not it is the joint product of a number of dif­ ferent participants and it is this phenomenon of 'group responsibility' which is the essential difference between interactive and non-interactive. To say, as do Murphy and Candlin (1976) that: 'Despite its apparent one-man-show format, the lecture is an interactive situation. It is not simply a monologue delivered into space which happens to be occupied by students. Audience and speaker communicate with each other through eye contact, gaze direction, facial expression, etc .. .' (quoted and discussed in Hutchinson 1978:5) emphasises the point that face-to-face oral discourse (such as the type of lecture referred to here) has certain characteristics which are not to be found in discourse where the participants are separated by a greater distance, but it tells us nothing about the difference between a lecture and a discussion.

HENRI HOLEC

197

at the University of Berne from the evening of Wednesday 30th May to midday on Saturday 2nd June 1979... . the invited participants will present a paper at the Colloquium and agree to its being published in the proceedings. Previous years have shown that it is an advantage to have papers bear on a particular theme; and this year we would like this theme to be discourse analysis and its application to language teaching. If you would like to accept this invitation to attend I would be grateful if you could let me know very soon, and at the latest by 3 1 January 1979. Participants will be asked to let me have a copy of their papers by 12 March 1979 so that I can then arrange to have them duplicated and distributed to other members of the group well before the Colloquium takes place. There will, of course, be an opportunity to revise the papers in the light of discussion before they are submitted to press. I look forward to hearing from you.

H. G. Widdowson

Nancy, le 25th January 1979 Lll Durt:tftll'.

Professor H. G. WIDDOWSON University of London Institute of Education Bedford Way ' London, WCIH OAL

Dear Henri, Aren't I honoured and pleased at your invitation ! Thank you mille fois, mon vieux. I don't know much about language teaching and still less about discourse analysis, but this won't hamper my powerful cerebral machinery! Looking forward (and upward!) to seeing you again. Friendly-ly, Henri Holec

21st March 1979 Dear Colleague,

I wonder if I could remind you gently that I would like to receive copy of your Berne paper as soon as possible so that it can be distributed before the Colloquium. So far I have only received 1 paper. Best wishes. Yours sincerely,

Professor H.G.Widdowson

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Yours,

198

O R A L INTERACT I V E D I SCOURSE

I s this discourse interactive o r non-interactive? Before w e can answer this question, it is necessary to make clear exactly what we are talking about: (i) Is each letter taken separately-in which case, we will have to reply that we are dealing with three different examples of non-interactive discourse. (ii) Or is it the set of three letters taken together-in which case our reply will be that we are dealing with a single, interactive discourse. The analytical observations which follow show that everything points to (ii) being the case, these letters being three 'speaking turns' which form a whole.

(b) At the level of illocutionary structure (i) The first letter includes, amongst other things, an invitation and two requests. (ii) The second includes, amongst other things, an acceptance of the invitation and complies with one of the two requests (it is dated before the deadline stipulated in the first request). (iii) The third letter includes, amongst other things, a repetition of the request which has not yet been complied with. Therefore, despite extreme variations in the modalisation of the illocution­ ary acts (formal/informal), the discourse is also coherent from the illocution­ ary point of view. (c) At the level of content (i) The invitation included in the first letter is an invitation to a particular colloquium, bearing on a particular topic. (ii) The acceptance in the second letter is an acceptance to attend this same colloquium (by implication) on the topic referred to (which is repeated in­ directly in the acceptance). (iii) The request in the third letter deals with the same matter as the request in the first letter, the venue and date of the colloquium being repeated. Therefore, at the level of content, the three turns are also coherent. It is, then, the whole formed by the three letters which should be considered, and this is an interactive discourse containing three 'speaking turns'. Of course, there are many differences between written interactive discourse and spoken interactive discourse, due both to the non-temporal nature of written text and to the communicative uses to which it is generally put, e.g., communication at a distance, which results in the 'speaking turns' being spaced out in time, which further means that each turn has a certain in­ dependence (it is not possible to interrupt, for example, and there is a tendency

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

(a) At the level of the interactive structure (i) At the beginning of the exchange an interactive pair consisting of 'Speaker' and 'Addressee' is set up; the Speaker selects himself and appoints his Ad­ dressee. (Realisations include the name and address of the recipient at the top on the left, and the opening 'Dear Henri'). (ii) The Addressee appointed replies to the Speaker (the name and address at the top on the left; 'Dear Henry'). (iii) The third letter opens a new exchange, since the Addressee is no longer just the Henri of the first letter but all the participants in the Colloquium who have failed to send in their articles. The discourse is, therefore, interactively coherent (cf. Riley, 1979).

HENRI HOLEC

1 99

to include a s much information as possible in one single turn). Generally speaking, the ways in which interactive structure is realised are completely different, so that written interactive discourse is not just a written transposi­ tion of oral interactive discourse. Nevertheless, these differences do not mean that the wr:tten form excludes any possibilit of interaction, even if written discourse is not usually in­ teractive Gust as oral discourse is not usually non-interactive). As soon as a certain number of conditions are fulfilled, such as the triple coherence which has been described above to show that the three different letters are parts of the same whole, written discourse can be produced interactively. But it must be emphasised that all these conditions have to be satisfied and not just any one of them: for example, it is not interactive simply because, at the level of content, a reference is made to the content of another discourse, as is often the case in scientific articles, say, where other writers and other texts are quoted . This is not sufficient to make it a 'speaking tum' forming a coherent in­ teractive whole with the quotation. Below is an example taken from the letters page of the Nouvel Observateur, which clearly brings out what does and what does not form part of an interactive written discourse:

y

Oser affirmer, comme le fait Fabien Gruhier dans son article « Ia Fievre des pieds fourchus » (« le Nouvel Observateur » n 754), que ce sont les agriculteurs les moins touches dans !'affaire de Ia fievre aphteuse est vraiment une preuve d'ignorance de ce qu'est !'agriculture. [ . . . ) Les eleveurs sont indemnises d'apres une expertise, mais croyez-vous que le proprietaire d'un troupeau de qualite, selectionne pendant vingt ans, va trouver sur le marche un troupeau d'une valeur comparable? Jamais. II devra done repartir de zero. Et je ne compte pas le choc psychologique. [ . . ) Realisez-vous ventablement que, de 1962 a 1978, un tracteur •

.

de 75 ch a vu son prix multiplie par quatre alors que les prix agricoles - du ble au pore en passant par le ba:uf - ont presque stagne? Et i1 en est de meme pour tous les materiels et foumitures (engrais, fuel, etc. ) . [ . . . ) Je ne parlerai pas des salaires: un berger, dans Ia region, touche le S.M.I.C. mais travaille plus de soixante heures par semaine. [. . . ] Ce qui me console, c'est que cette cecite volontaire sur les problemes agricoles existe a gauche comme a droite. L'agriculture, c'est le petrole mrus on s'en fout tant que �a tient. Jusqu'a quand? OLIVIER GIRAUD, L 'Isle-Jourdam.

[Sans doute faut-il (( repartir de zero )) pour reconstituer un troupeau. MaiS en attendant, en plus du remboursement du chepte/ abattu, les agricu/teurs concemes toucheront une indemnisation pour leur manque a gagner. ns sont done ef­ fectivement moins touches que tous les autres, ceux dont le cheptel n 'a pas ete abattu, qui souffrent pourtant de /'in­ terruption des circuits commerciaux, et pour lesquels aucune compensation n'est prevue. - F.G.] The letter from the reader, addressed to the author of an article which has prompted him to write, forms, together with the author's reaction, an in­ teractive discourse. But the article in question does not form part of this discourse : it is simply a referent, in the same way as the general situation of cattle-breeders is, for example. Translated by PHILIP RILEY

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Le coOt de Ia fi�vre aphteuse

200

O R A L I NTERACT I V E D I SCOURSE

NOTES 1

It is the spatial nature of such texts which explains the

use

in meta-discursive comments on the

structure of an argument of spatial expressions such as 'above', 'below', 'in the pages which follow' etc. In texts of the frrst type the internal temporal structure results in the use of temporal expressions 'As I may show later on', 'As I was just saying . . .' 'I will go into this when we come to . . . ' 2

Obviously, i t i s also possible to distinguish between sub-ategories of oral o r written discourse

according to the type of situation (face-to-face/at a distance, 'live'/recorded transmission) or to the register (literary/scientific, academic/non-academic) and so on. ' For further details on the way in which this alternation is carried out see Gremmo, Holec and Riley, 1978a, and Riley, 1979. •

Hutchinson describes the lesson from which this extract has been taken as having in fact been

recorded in the absence of any students: ' . . . played direct to the camera with only film crew as also be noted that the transcription has been considerably 'cleaned up' or 'idealised' .

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

audience' (p. 5), which further confirms the belief that it is non-interactive discourse. It should

WHEN COMM UNIC ATION BREAKS DOW N: LEVELS OF COHERENCE IN DIS C OURSE PHIL I P R ILEY

Centre de Recherches et d'App/ications Pedagogiques en Langues, (CRAPEL), Universue de Nancy /1, France

I N T R O D U CT I O N

As the movement towards 'communicative' teaching grows, i t becomes more

A. B.

What is your name? Well, let's say you might have thought you had something from before, but you haven't got it any more

This example of schizophrenic' s talk (from Labov, 1970 also cited by

Applied Linguistics, Vol. I, No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

and more common for language teachers to say 'I only correct (interrupt, etc . ) when communication breaks down' . This reaction against the incessant nit­ picking which characterises so much teaching is a healthy one: at least it en­ courages the learner to try to speak, rather than penalising him every time he opens his mouth. But it also glosses over a very real and important problem: how do we know when communication has broken down? The classroom teacher's answer to this question is usually a simple one. 'Communication has broken down when I do not understand what the learner is trying to say . ' Now this i s a good rule-of-thumb-but it has no diagnostic or remedial value whatsoever. Moreover, it imposes the sort of either/or decision-a particular attempt at communication is judged to be either a complete success or a complete failure-of which case-hardened applied linguists tend to be very suspicious. They prefer to deal with clines, with degrees of difference, rather than with black and white opposed categories. This is particularly true for the linguist who looks back at the dossier on grammaticality, and who sees the way in which early attempts to build up a case of unacceptability against certain doubtful sentences were outwitted by fast-talking semanticists, and alibis were found for suspicious constructions. Once bitten, twice shy: this time we'll start by taking a more flexible line by talking in terms of degrees of . . . Of what? As Malcolm Coulthard (1978) has pointed out, in the description of oral face-to-face interaction, there seems to be no discourse equivalent of the concept of grammaticality: 'a next speaker always has the option of producing an unrelated utterance'. This might well seem to be discouraging to linguists whose aim is to describe the relationships holding between utterances (or, rather, between communicative acts.) If anything goes, what is the point of trying to establish neat little Newtonian models? What are the rules of anarchy?

202

L E V E L S OF C O H E R E N C E IN D I S C O U R S E

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Coulthard , 1 977), has become the locus classicus for the problem o f ap­ propriateness . But whereas it is usually taken to exemplify discourse in which communication has completely broken down, it will be argued here that the very fact that B made a contribution to/took a turn in the discourse shows that this is not the case: however attenuated, some form of communication has taken place. And then, at the opposite extreme, there is the very real sense in which all dyadic discourse is schizophrenic-it is the product of two per­ sonalities, however well they succeed in communicating. What we need, then, is some way of describing degrees of 'discoursality'. This article is a very tentative step in that direction . It argues that if we distinguish between interaction , illocution , content 1 and realisation we can show that a given communicative act may be acceptable discourse at one or more of those levels but not at another or others. By pinpointing why and how an 'utterance' is unacceptable, the degree to which it is deviant, the linguist is once again back in business as a discourse analyst, since he is (hopefully) now able to establish a model which can give a chain-exhaustive account of discourse as a series of relatively related utterances. This does not mean, though, that he can claim to be describing all the factors which enter into the structuration of discourse : the psychology and motivations of participants in an interaction will remain outside his com­ petence, as will their social roles. Although his observations may shed an indirect and interesting light on such factors when they have linguistic correlates 2 the linguist restricts himself to those aspects of the structure of observable communicative behaviour which can be shown to be rule-governed. In other words, a model for the description of discourse is not an overall theory of communicative behaviour, nor is it a theory of either communication or behaviour: it is an attempt to account for discourse, i.e., the linguistic aspects of interaction. This may mean casting his net much wider than he has previously done, to include non-verbal behaviours, for example (cf. , Riley 1 976) but this does not alter the fact that such a demarcation line has to be drawn somewhere. Here we will concentrate on two fundamental distinc­ tions- ! . Interactive v. non-interactive discourse, 2. Interaction v. Illocution. A second aim of this article is to show that work at present being carried out in discourse analysis is relevant to language teaching in an immediate, fun­ damental and detailed way. As the boundaries of linguistic science have grown wider still and wider, giving rise to a large number of hybrid applications,­ sociolinguistics, neurolinguistics, psycholinguistics,-and to exciting develop­ ments in fields as disparate as philosophy, language acquisition and semiotics, the language teacher tends to feel that applied linguistics is very interesting and all that, but hardly relevant to his needs and interests. This is one of a number of articles based on the work of a CRAPEL research team consisting of M-J . Gremmo, Henri Holec and the author i n which we have investigated the pedagogical applications of discourse analysis. Earlier articles have considered applications of our own research to the evaluation of audio-visual teaching materials (Holec, 1 975) to problems in communicative language teaching (Riley, 1 977, Gremmo , Holec and Riley 1978) to the study of the teacher's role (Gremmo, Holec and Riley, 1 977) and to contrastive analysis (Riley, 1979). Here we try to outline an approach to the description and understanding of communicative error or failure. Readers already familiar with the earlier work

PHILIP RILEY

203

are asked to excuse a certain amount of repetition, but newcomers need to know the story so far .

1.

I N T E R A C T I V E V. N O N-I N T E R A C T I V E D I S C O U R S E

There is n o generally accepted typology or taxonomy for discourse: s o far, workers in the field have been kept busy enough simply identifying the relevant parameters and resources, a task which is still by no means complete. Among the oppositions and clines "hich have already been identified are: spon­

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

taneous/prepared; oral/wmten; formal/informal; authentic/non-authentic. It can and has been shown that variations along any of these parameters will result directly in corresponding variations in the discourse type. In what follows we will be discussing spontaneous, oral, authentic discourse. A further such opposition,-but one which has been largely ignored until very recently,-is interactive/non-interactive. By interactive discourse we mean discourse which is the collaborative construct of two or more par­ ticipants mutually engaged in other-directed communicative behaviour. To put it more simply: is the discourse produced by one person or more than one? How many sources or contributors are there? The defining characteristic of interaction, then, is alternation: whatever the relative length or frequency of individual contributions, verbal or non-verbal, there must be this 'floor-sharing' if contributions from more than one par­ ticipant are to be integrated into the discourse. To be adequate, a description of discourse must account for alternation, for the rules governing the transi­ tion from speaker-state to hearer-state: this includes not just turn-taking, but rights to the floor, which is why descriptions based on the 'ideal speaker­ hearer' necessarily lack a social and interactive dimension. It should be noted that we are here using the term interaction in a restricted sense, to refer to one particular aspect of communicative behaviour. By in­ teraction, we mean the set of actions and reactions which are realised by turns ('taking the floor'), i.e . , the give-and-take which characterises this particular category of or.al communication. It follows that this particular aspect, this particular definition, of interaction could never be applied to the analysis of written discourse as is the case with certain other approaches to discourse analysis. 3 We are dealing here with interactional tactics: who speaks when and to whom, in other words, how the turns are distributed. We are not concerned for the moment with what participants say, only with the fact that it is, or is not, their turn to say it. Participants regulate their interaction by address i . e . , the transition in real time from one turn to another. The rules of address decide whose turn it is, and whose turn it will be next . The timing of turns is determined by the turn­ taking signals (most of them non-verbal, particularly key, orientation and gaze) which are emitted by the participants (Kendon 1967, Duncan 1972). When a participant makes the transition from one turn to another (e.g. , he 'hands over the floor' , or he 'takes his turn') he enacts an interactive role which can be described in terms of functions and acts. Interactive roles are not to be confused with social roles (cf. Gremmo, Holec, Riley, 1977): they define the nature of the individual's participation in a given minimal interactive structure, the Return. In very general terms, four such roles are available to participants:

204

L E V E L S OF C O H E R E N C E IN D I S C O U R S E

(a) The role o f Speaker (S), who initiates and closes the minimal unit of in­ teractive structure, the Return. (b) The role of Addressee (A) which consists in the acceptance of the turn­ taking restraints imposed by S. (c) The role of Talker (T), normally available only to the A of the previous turn and which differs from S in that there is no choice of interactive partner, who must be the previous S. (d) The role o f Listener (L), which might at first seem t o b e a purely negative role, since it consists in participating in an interaction without taking the floor, but of course any Listener is a potential Speaker or Addressee, to say nothing

These interactive roles are realised by the performance of interactive acts, i.e., particular types of turn. There are three basic types of act: Opening (0), Closing (C) and Reply (R).Acts combine to form interactive structures which may be hierarchically ordered as follows:

INTERACTION

TURN:

O, R, C

RETURN:

O + Rn + (C)

INVOLVEMENT:

[O + Rn + (C)] + [O + Rn + (C)]n

ENCOUNTER:

{[0 + Rn + (C)] + [O + Rn + (C)]n} n

[0 + Rn + (C)]n}

+

{[O + Rn + (C)]

+

In reading this table, the following points should be kept in mind: (i) Turns are 'acts', but of course this is a nonsense, communicatively speaking: ' What is the sound of one hand clapping? ' They do not enter in­ teractive discourse as there is no alternation. Return is the minimal structure. The end of a return is signalled by a C and/or by a change of Addressee. Involvement is the intermediate structure. The end of an involvement is signalled by a change of Speaker. Encounter is the maximal structure. Above this, we would be dealing with situationally-defined events. (ii) A dyadic interaction, even one of considerable length, might well be limited simply to a series of returns . That is, this model distinguishes between interactions involving two and three or more participants. The model is described in greater detail in Gremmo, Holec and Riley, forthcoming.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

of the ever-present possibility of interruption (i.e., floor-taking out-of-turn, not anticipating one ' s turn). Moreover, Listeners have an indirect influence on the interaction via the continual stream of information which they com­ municate to S or A (approval, impatience, boredom, etc .) verbally or non­ verbally.

PHILIP RILEY

205

1 . 1 An analogy It might help the reader to understand what is meant by 'interaction' here if he thinks o f it in terms of a game of tennis , a rule-governed activity which also requires a minimum of two participants. There, too, we find an alternation in the turns which players have: rallies may be of any length, but players must take it in tum to play their strokes. When it is my tum, I can play one of several kinds of stroke: a Service (if it is my tum to start), a Reply (if my opponent has just hit the ball into my side of the Court) or Out (if I fail to return the ball to my opponent's side). In each Rally, then, there is only one Service (S). This may be followed by any number of Replies (R). The end of a Rally is signalled by

The minimal structure for a Rally, then, is

S + Rn + (0) A minimum of four Rallies gives us a Game, the end of which is signalled by a change of Server. The analogy can be continued through Set and Match, but we'd better not push things too far. For the moment, let us j ust note that the description of tennis resulting from this approach makes no reference to many aspects o f the game: psychological factors, such as the will to win, physical factors such as fitness and speed, formal differences between strokes (lob, volley, smash, etc.)-all these are consciously omitted to enable us to see starkly one particular aspect of the game, the work rate or distribution o f turns.

1 .2 Rules ofperformance One of the conclusions to which the investigator of interaction (as understood here) is driven is that there are some rules of performance which are based not on some underlying linguistic competence but rather on the physical nature of the activity and the medium in question. Space, time and the materials which occupy them are subject to physical rules which are inherent in their nature. These rules impose certain restraints on the development and growth of all materials. Although very simple in themselves such restraints often result in extremely complex patterning. 4 '

Suppose you make a small disc of clay. It is obvious that with a rolling pin you can spread the clay into a larger disc. Under the action of the rolling pin, the clay spreads evenly in all directions-just like the space in which it lies. Now, suppose you press or flatten only the center of the disc. You can do that by manipulating the clay with your fingers. Flattening the center causes the center to spread and grow faster than the perimeter, and the disc naturally takes the shape of a bowl. You can also squeeze the perimeter of the disc so that it grows faster than the center. Again the disc will not lie flat. Instead, it thrusts itself simultaneously both forward and back to make a saddle. The saddle comes about just as naturally as the bowl. The clay is not moulded into those forms but the forms arise naturally, depending upon where you press the clay. From playing with the clay, we are led to the discovery of a fundamental rule: if the center and perimeter grow at the same rate, the material spreads in a plane; if the

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

(i) An Out (ii) The cessation of alternation: if the same player plays two consecutive strokes, the second must be a Service.

206

L E V E L S OF C O H E R E N C E I N D I S C O U R S E

The same type of argument can be applied to the description of meanders, branching growths-and to face-to-face interaction. Now it is true that it is possible to give highly abstract descriptions of forms of such structures-the wavy edge of the oyster shell, for instance. But we must not make the mistake of thinking that our abstraction is the underlying structure: it is only a description of the results of certain physical constraints. It is those constraints which govern the structure and evolution of the material . Similarly, the series of abstractions which we call langue or competence must not be thought of as the only rules or constraints governing the structure and evolution of com­ munication: face-to-face interaction is also a physical phenomenon subject to the constraints imposed on the various media by space and time. Alternation, for example, which we have seen to be the most fundamental characteristic of interactive discourse, is imposed by the physical nature of communication (cf. , C. F. Hockett's 'design features of language') which make it nearly impossible for two acoustic messages to occupy the same acoustic space at the same time and which prevents us from acting as emitters and receivers simultaneously (indeed, we can go further and say that this is why so many behaviours regulating interaction are in fact non-verbal .) If we ignore this physical dimension of face-to-face interaction-if we preserve a neat competence/ performance opposition, where performance is itself a level of abstraction­ we misrepresent some of the most important features of this type of discourse . For example, there seems good reason t o believe that many o f the ' I hereby's' of speech act theory are only verbal paraphrases for the interactive act o f floor-taking. I n face-to-face interaction, ' I hereby' b y taking m y turn, by performing an interactive act, by taking the floor and contributing to the discourse. That I usually do so non-verbally is neither here nor there: by concentrating on the purely verbal aspects of communication, speech act theorists are obliged to try to account for all aspects of discourse in verbal and illocutionary terms alone. This leads to the circularity of giving a verbal label to the act of taking the floor which has to be integrated into the deep semantic structure of the utterance only to be deleted later, because there appears to be no 'surface realisation' . But, there is in fact and it is related not just to an underlying competence but to the physical laws governing the distribution o f the object, discourse, i n time and space.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

center grows faster than the perimeter, or the perimeter grows faster than the center, a bowl or a saddle results. The reason for those transformations lies in the nature of space. The transfor­ mations have nothing to do with our intentions to make one form or another. Nature too is similarly constrained. She makes cups and saddles not as she pleases but as she must, as the distribution of material dictates. Take an oyster shell, for example. Since the perimeter of the shell grows at a faster rate than the center, the perimeter curls and wrinkles. No genes carry an image of how to place the wrinkles; no genes remember the shape of the shell; they only permit or encourage faster growth at the perimeter than at the center. Similar differences of growth lead to the development of more complicated structures, like the outer shell of the human ear. The convolutions of the outer ear arise like the convolutions in a piece of paper that has been sprinkled with water. The living tissue and the paper both bend and warp in accord with the differential expansion of their surfaces.

PHILIP RILEY

207

1. 3 Analysis of a passage of classroom interaction Consider the following passage: it is an extract from a transcription of a language class given by a native speaker of American English to a group o f four learners :

Turn

Participant Teacher

2. 3. 4. 5. 6. 7. 8. 9.

Mr. P. Teacher Mr. P . Teacher Mr. P . Teacher Mile X Teacher

10. 11. 12.

Mr. P . Mr. D. (?) Teacher

13. 14. 15. 16. 17. 18. 19. 20. 21 . 22.

Mile X Teacher Mile X Teacher Mile X Teacher Mile X Teacher Mile X Teacher

er Mr. P. er what's the man doing . . . he's sitting but what's he doing with his hand She's pointing their hand Pardon He is pointing his hand OK he's pointing his hand and what and he is showing the seat in front of him OK he's pointing his hand and what the menu . . . the menu The menu or (gesture) look at the picture look at the picture . . . he's pointing at his watch why is he pointing at his watch Because she's late She's she's late what OK the girl is late and perhaps (gesture) he's been (drums hands on desk imitating impatience) he's been . . . Wait wait . . . Waiting He has waited He's been waiting waiting many many times Many times (French gesture for doubt) many times Some times . . . Some times (gesture) No No he's been waiting for (gesture) for a (gesture-"fisherman's tale" long) A lot of time A long time A long time he's been waiting for a long time er let's look at the text and er Mr. D. will you read the text please the text on the next page Julia had a date a date with her new boyfriend in this restaurant at 8. He came on time but she she did not. She came in only a moment ago. It is 9. "Have you been waiting long?' she asked him when she came in She what when she came in She asked him she asked him She asked him she asked him when she came in. "Yes I have" , he is saying: he is rather angry Rather he's he's what he's rather angry he's rather angry (writes on blackboard). Because he has been waiting for an hour An hour an hour Yes for an hour Yes what time is it now on the picture =

23.

�}

24. 25.

Teacher Teacher

26.

Mr. D.

27. 28. 29. 30.

Teacher Mr. D. Teacher Mr. D.

31.

Teacher

32. 33. 34. 35. 36.

Mr. D. Teacher Mr. D. Teacher Teacher

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

I.

208

L E V E L S OF C O H E R E N C E IN D I S C O U R S E

Mlle X Teacher

39.

47. 48 . 49. 50. 51. 52. 53. 54. 55. 56. 57.

Mr. Mlle X (?) Teacher Teacher Mlle E. Teacher Mlle E. Teacher Mile E . Teacher Mile E . Teacher Mile E. Teacher Mile E. Teacher Mile E. Teacher Mlle E. Teacher

58. 59.

Mile E . Teacher

60.

61.

Mile E. Teacher

62. 63.

Mr. D. Teacher

64.

Mlle X Mr. D. Teacher Mr. D. Teacher Teacher

40.

41. 42. 43 . 44.

45. 46.

65. 66. 67. 68. 69. 70. 71. 72.

P. }

Mr. D. Mlle E. Teacher Mr. D. Teacher

}

}

It's nine o'clock Yes it's nine o'clock . . . and when did this er man arrive when did he arrive in the restaurant when did he get to the He arrived at 8 at eight . . . 8 o'clock OK. (writes on blackboard) Miss E., can you ask a question with how long How long ago How long and the man How long ago did he arrived Mmm be careful No how long did he Er ask me a question with how long has How long has he arrived What happened at 8 o 'clock what happened He arrived at 8 o'clock OK he arnved huh and what's he doing right now Now he waits Now he's Now he's waiting Now he's waiting How long has he wait he wait . . . wait How long has he waited or when did he start waiting when did he begin waiting when did he begin waiting when did he start waiting At 8 o'clock At � o'clock OK at 8 o'clock he was waiting and now he's waiting and that that's been continuing huh so how long has he has has he what what can we put here (gesture indicating sentence on blackboard) In the restaurant in this restaurant I wanted you to I want you to change something with this with wait OK at 8 o'clock he was waiting now he is waiting how long has he (whistle to indicate blank that they have to fill in) look at the text look back at the text look back at the text look at Julia's question What exactly is that question No look at her question Mr. D. Look at the question what is the what is the question She asks when he came No no look at the text not not the question look at the question Have you been waiting long? Yeah have you been waiting long? . . . Mmm OK now this question is very similar you can change this question so that it's it's looks like \ How long has . . . waiting long Has be been . . . wait How long has be been Waiting Waiting do you see that he started waiting at 8 o'clock he's still wciiting now and he's he's been waiting for an hour he started an hour ago and it's continuing he has been waiting for an hour

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

37. 38.

209

P H I L I P RILEY

If we apply our system of analysis to this passage we obtain the following description:

Tum No Teacher M' P. M"' X M' D M"' E. Act

3

2

I

A T L L L

s A L L L

4

s

A T L L L

s A L L L

8

7

6

A T L L L

s A L L L

return

23

c .,

.J : • c

<� R C 0 ,______.

42 A L L L T

0

R

IS

14

A L T L L

A L L T L

0

A L L T L

R

A L L T L

s L L A L

R

R

31

30

29

28

s L L A L

R

32

R

R

34

33

A L L T L

s L L A L

R

s L L A L

A L L T L

R

R

3S

36

s L L L L

44

4S

A L L L T

s L L L A

R

R

R

A L T L L

18

19

20

s L A L L

A L T L L

s L A L L

38

39

40

C

0

R

C

R

0

.___.

21 A L T L L

48

49

so

Sl

S2

s L L L A

A L L L T

s L L L A

A L L L T

s L L L A

A L L L T

R

R

R

R

R

R

R

S3

S4

s L L L A

A L L L T

return

R

R

(VIi)

S6

S7

s L L L A

A L L L T

s L L L A

R

R

R

ss

s L L L L

._______.

(VI)

47

A T T L L

s A L A A

return

A L L L T

46

37

s L A L L

return

s L L L A

A L T L L

s L A L L

(v)

43

17

16

A L T L L

s L A L L

(iu)

(IV)

s L L L A

13

s A A A A

return

return

41

12

A T L T L

SB

S9

60

A L L L T

s L L L A

A L L L T

R

R

R

return

(VIii)

62

63

64

6S

67

70

72

s L L A L

71

A L L T L

s L L L L

s A A A A

A L L T T

73

s L L A L

A I T T L

68

69

A L L T L

66

s A A A L

s L L A L

s A A A A

R

R

R

R

R

0

R

T

s L L L L

0

C

A L L T L

61

return

(ix)

R

return

(X)

C

.£.

S. Speaker T: Talkc:r A: Addressee L·

lulcnc:r o· Open••• R: Reply C OOSins

return

(XI)

Since most of the signals or behaviours which realise Address-and which enable us to assign symbols to utterances-are non-verbal, this analysis could only really be justified by providing the reader with a copy of the videotape and by annotating each separate tum with a description of the relevant NVC. In the absence of any satisfactory notation, this description would have to be verbal and lengthy and therefore no attempt has been made to provide this detailed information here. Nonetheless, certain turns and their attributions to particular participants probably need further clarification. How do we decide, for example, that Tum 1 is an 0 produced by the Teacher as S? Our reply would be that he takes the floor, which is vacant, spontaneously; that is, no other participant was occupying the floor and he himself was under no interactive obligation to do so (he was no one's Ad­ dressee). He initiated this stretch of interaction by taking the floor and choosing his Addressee, his interactive partner 'er Mr. P . ' (-an example here

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

0�

27

26

s L L A L

s L L L L

II

s A A A A

(li)

2S

24

A

s A L A A

10

T L A L L

return

(i)

22

9

A L s L L

s A L L L

210

L E V E L S O F C O H E R E N C E IN D I S C O U R S E

o f Address being realised verbally.) When Mr. P . fulfills his interactive duty in Tum 2 he is a Talker not a Speaker precisely because he did not initiate this return and because he has no choice of Addressee: he must reply to the Teacher. As long as this particular interactive S and T partnership continues we are dealing with the same return. This occurs in Tum 8, where Mile X takes the floor without having been the Addressee of the previous utterance: i .e . , she performs an 0 . This is a clear example of the difference between interactional

2.

I NT E R A C T I O N V . I L LOC UTI O N

The major aim o f the discourse analyst is to describe communicative behaviour in terms of the constraints placed by participants on other and subsequent contributors to the discourse. This is a complex task, since there are many types of constraint: there are the interactive constraints wnich we have just been looking at which, roughly speaking, oblige a participant to speak or to keep quiet . Other kinds of restraint, though, can be placed not on a participant's right to speak, but on what he says: a Speaker can impose the duty to Reply at the same time limiting the choice of his Addressee as regards the illocutionary value and the propositional content of his reply. Interactive acts, then, such as taking or giving the floor, are to be distinguished from the illocutionary values which messages occurring at those points in structure may happen to have: an Opening turn may be occupied by a Greeting, Requesting Information, Ordering, etc. The relationship can be seen as a tagmemic, or slot-and-filler one: in each interactive slot there is at least one illocutionary filler. At the same time, sequences of illocutionary act give us illocutionary structure.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

structure, where this is an initiation and illocutionary structure, where it is the second element in an exchange. Turns 9 and 10, although forming one utterance, are separate interactive acts , Turn 9 being in reply to Mile X, Turn 10 marking a change of Address and, consequently, the beginning of a new Return: it is an example of General Address , where the Speaker imposes the duty to reply on several or all of the other participants. In this case, the transition from Address to Mile X to the whole group was marked by gaze, key, posture and gesture: he looked all round the group, went from low to high key, took up a step-back stance and held out his right hand palm upwards. A Speaker who has signalled a General Address has the choice of continuing to do so as in Turn 12, or of 'narrowing down' his Address to those participants who reply, as in Turn 14. At the end of that same Return, though, the process is reversed: here (21 ,22) the Teacher switched his Address from Mile X to the rest of the group. In such cases, the change of Address almost always implies the exclusion of the previous Ad­ dressee (it is a form o f evaluation or sanction of Mile X's performance) which is the reason for coding her role in 22 as Listener and not including her in the General Address. Return number (iv) contains a Closing (24). The Teacher, who was the Addressee of the previous turn, fulfills his interactive duty by taking the floor but does not himself choose an Addressee, that is, he does not impose on any of the other participants the duty to reply. This C is signalled by a drop to low key, by the teacher's breaking off eye-contact and by a pause before the selection of a new Addressee (in 25, the Opening of a new Return .).

PHILIP RILEY

211

The adjacency pair can b e cited as an extreme example o f illocutionary constraint. Normally, requesting information is followed by informing, greeting by greeting:

Good morning Good morning

Greeting Greeting

}

Exchange

But we can also attest

Good morning You're late

Greeting Remonstrating

}

A . - Good morning B. Good morning A. B.

Good morning You're late

�} �}

Return Return

The same would hold true even for pathologically deviant sequences:

A. What is your name? B. Well, let's say you might have thought you something from before, but you haven't got it any more

�}

Return

We are dealing, then, with two 'parallel' structures: as it happens in these examples there is a congruence in the units of those structures-the boundaries of the illocutionary Exchange and the interactive Return coincide exactly. But this is by no means always the case: let us look back at the long Return between the Teacher and Mile E. which occurs between turns 41-60 in the passage we took for analysis. This is one Return because it is an unbroken stretch of in­ teraction between two participants: there is no change of Speaker nor of his Addressee. At the illocutionary level, however, it is a very different kettle of fish: in his (ultimately unsuccessful) bid to get a particular answer from Mile E . , the Teacher initiated a long series of seven illocutionary Exchanges, thus:

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

It is obviously not going to be possible to set up hard-and-fast rules for discourse structure based on illocutionary forces alone or their positions of occurrence alone. We do not need to take the extreme case of schizophrenic discourse cited earlier to show that any illocutionary force may be followed by any other, although there are of course statistical correlates . This presents the discourse analyst with an insuperable problem if he thinks that illocutionary constraints and structures are the only ones there are. But if he recognises interactive structure, he can show that although the second Exchange may be 'deviant' at the illocutionary level, it is perfectly coherent at the interactive level.

L E V E L S OF C O H E R E N C E I N D I S C O U R S E

212

INTERACTIVE

ILLOCUTIONARY

TURNS .....

41

42 43 44

45 46

-

52 53 54 55 56 51 58 59

60

61

MOVES

Nominating, elicit

Initiating

reply elicit reply evaluation reply elicit reply elicit reply evaluation, elicit reply evaluation reply ' accept reply elicit reply accept, commen oticit

Responding Initiating Responding Follow-up Responding Initiating Responding Initiating Responding

} }

reply comment, elicit

�

I

Initiating Responding Follow-up Responding Follow-up Responding Initiating Responding Initiating Responding Follow-up

!

EXCHANGES Eliciting (i) exchange

Exchange (ii)

} }

Eliciting (iii) exchange Eliciting (iv) exchange

Eliciting (v) exchange

}

}

Eliciting (vi ) exchange

Eliciting (vii) exchange

This is an attempt to apply the Sinclair-Coulthard system of analysis­ rather a ham-handed one, as we are not always sure that we apply their categories as they understand them. Questions of detail aside, though, one thing is clear: there is no one-to-one relationship between their units of illocutionary structure (act, move, exchange, transaction and interaction) and our units of interactional structure (turn, return, involvement and encounter). The Sinclair-Coulthard system gives an analysis of the sequences of illocutionary acts which occur in the language of the classroom and of the constraints which participants place on one another at the illocutionary level. As such it is different from, but perfectly compatible with, the interactive analysis we looked at earlier_ We have here, then, a series of 20 interactive turns which form one in­ teractive return: the same passage (approximately) includes 26 illocutionary acts forming 20 moves, forming 7 exchanges. Just as an utterance may include more than one interactive turn (72, 73) it may also include more than one illocutionary act (59). Moreover the boundaries between units may coincide but they may also occur at quite different points: the final Eliciting exchange, for example, continues beyond the end of the interactive Return.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Return (viii)

47 48 49 50 51

ACTS

P H I L I P RILEY

213

2.1 Interruptions

An interesting aspect of discourse on which much light is shed by the in­

What's the time Why do you want to know Well it's for my bus Oh , I see. It's 6.30 'fl!�ks

Return

We have here an adjacency pair whose members are not in fact adjacent. Yet any speaker of English can pair them off. It clearly cannot be position of occurrence alone which justifies his doing so-which is a further reason for distinguishing the turn from the message. In the interruption exemplified by Turn 8, the interrupting Speaker ad­ dressed the interrupted Speaker. This seems in some sense to be less of an interruption than one where the interrupting Speaker addresses the original Addressee (i. e . , the participant whose turn he has just taken) . And this in turn is less of an interruption than would be the case if the interrupting Speaker were to address a previous listener. This would be bordering on 'talking in class' . All these interactive interruptions may or may not coincide with illocutionary and content interruptions: to take just one example, the in­ terrupting Speaker who addresses a previous Listener may suggest an answer or he may talk about last night's football match. In this case the discourse is incoherent at the levels of interaction, illocution and content-and the 'talking in class' has become a separate but simultaneous discourse. By looking at who is the Speaker of an interrupting 0, then, and who is the Addressee, we are able to describe degrees in interactive coherency which go from the over-enthusiastic floor-grabbing of the learner who answers teacher's question out of turn, but whose participation in the same discourse cannot be questioned, to the 'aside' remark .

3.

R E A L I S AT I O N A N D C O N T E N T

Two further levels o f structuration are . posited in t h e model of discourse presented here: realisation and content. Neither will be discussed in detail since, although both levels present considerable problems to the analyst , at least the distinctions which they summarise are pretty generally accepted, which is not the case with the interactionlillocution opposition.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

teraction/illocution distinction is the problem of interruption. A purely illocutionary description such as the Sinclair-Coulthard model does not distinguish between contributions which are in turn and those which are not: it assumes that every successive utterance was in turn. Our intuitive awareness that this is not the case is a major justification for recognising the interactive dimension of discourse. For example, in the passage we have taken for analysis, the Teacher addressed Mr. P . in turn 7, but it was Mlle. X who replied. It was not her turn to speak, yet she was clearly answering the Teacher' s question . A full description of discourse must account for both these facts : it must show that there is a lack of coherency here at one level­ interaction-but that the discourse is acceptable at another level-illocution. We can now see that it is also perfectly possible to have an interruption at the level of illocution in an interactively well-formed passage:

L E V E L S OF C O H E R E N C E I N D I S C O U R S E

214

(1)

A.

( 2)

B. A. B.

I forgot to give John his car keys back and I can't get them back before tomorrow. I'm going to John's for dinner tonight. Do you fancy dropping round for a game? I ' m going to John's for dinner tonight.

The semantico-logical or pragmalinguistic operations which allow us to un­ derstand and label B's utterances as, say 'Offering help' and 'Refusing' have been the subject of intense study recently (in particular, see Cole and Morgan 1 975) and we do not feel competent to enlarge on them. However, two fun­ damental points need to be kept in mind: (a) If illocutionary force and propositional content can be shown to vary independently, they should, as far as is possible, be described separately. To describe illocutionary acts uniquely in terms of their relationship to content or information is to privilege the purely informative and verbal elements of communication at the expense of the psycho-social, pragmatic and dynamic aspects. It is true that such an approach may seem to account for certain types of discourse: Coulthard (1 977) describes teaching acts in terms of their relationship to information-'an accept . . takes information offered into the discourse and evaluation assesses its worth and relevance'. However, this does more to bring out a very idiosyncratic feature of classroom discourse than it does to provide us with a basis for wider generalisation. Sinclair and Coulthard were obviously fully aware of this (and besides they had a number of other criteria for the definition of their acts). But we are by no means setting up an Aunt Sally, as the following grotesque misrepresentation shows: .

The communication fallacy Language activity is too often viewed as a straight process of com­ munication, by which term is usually meant the conveying of clear-cut in­ formation ("What time is it? " , "The book is on the desk" , "don't forget to shut the window") from one sender to a receiver. Thus right from the beginning we are faced with a certain number of fallacies, or still worse, half­ truths, allegedly reasonable claims , unsubstantiated metaphors, which un­ derlie and shape the conception many linguists have of what language activity is; the metaphor or language as a conveying device, or a mental organ, with an

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

For the moment, then, we will limit ourselves to a few brief and rather general observations : By realisation we mean the set of message-bearing elements (verbal, paralinguistic, non-verbal) in a situation. These elements have substance and are realisations of various systems and structures whose organisation can be described in terms such as class, units, structure and distribution. The textual function of such elements is described in terms of their internal relations (and without reference to the meaning they carry) . We use the term 'content' in the sense of 'propositional content' . For our present purposes, the important thing is to distinguish between the illocu­ tionary value and propositional content of utterances; there is nothing new in this, of course, it is fundamental to the whole concept of the speech act. It is easy to show that content and illocutionary value can vary independently i.e., that the 'same' content can be present in acts having different illocutionary values (and vice-versa):

215

PHILIP RILEY

input and a n output, interlocks with another fallacy (referential values equated with truth values), thus cross-breeding a certain number of misconceptions' (Culioli 1 979) . For the understanding of our objection, it is important to realise that the 'conveying of author of these lines is using this view of communications ( information') as a stick to beat the concept of the speech act and its ap­ plication, the communicative approach. (b) If we separate content from illocution, interaction and realisation, and if within content operations we distinguish between (i) implications, (ii) presuppositions, (iii) notions, we have at our disposition a further set of tools for the description of communicative breakdown. It is most unfortunate that almost all the work which has been carried out so far in this area is based on artificial, isolated and decontextualised examples, and is therefore useless from our point of view. We can do no more here than hint at the sort of ap­ proach which might be adopted for the description of misunderstandings: =

Realisation

Illocution

Content

A. What's the time B. 6.30

,

,

,

,

(a)

A. What's the time B. I like fish and chips

,

,

,

X

(b)

A. What's the time B. Why?

,

,

X

X

(c)

A. What's the time B. nurdlenurdle

,

X

X

X

(d)

A. What's the time B.

X

X

X

X

(e)

A. What's the time c. 6.30

X

,

,

,

(0

(a) is coherent both as an interactive Return (OR) and as an illocutionary Exchange (Requesting information, Informing). It is also realised by well­ formed sentences and the propositional content of the reply is that requested in the elicit. (b) is coherent as an interactive Return (OR) and as an illocutionary ex­ change (Requesting information, Informing). It is realised by well-formed sentences but the propositional content of the reply is not that which was requested in the elicit. (c) This is a coherent Return, 'consisting or two well-formed utterances. But the illocutionary adjacency pair is interrupted, so that we have a Request for information followed by a Request for information. (d) Here B makes a complete hash of his reply. It is, therefore, incoherent at the level of Realisation, illocution and content. But a reply has been made, the tum taken, so that at the interactive level we still have a coherent return. (e) Here, B fails to reply. Interaction has broken down and the discourse is incoherent at all four levels. (f) An interruption. It was not C's tum, so the interaction is incoherent. But

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Interaction

216

L E V E L S O F C O H E R E N C E IN D I S CO U R S E

a request for information has been satisfied, so that at the levels o f illocution, realisation and content the discourse is coherent. This is sketchy in the extreme and an enormous amount remains to be done, particularly in the definition of illocutionary forces and in content and topic analysis (columns for implication and presupposition could be added, for example). Nonetheless, we feel that this approach has much to offer the ap­ plied linguist and the language teacher interested in the description of com­ municative error. NOTES 1 A similar distinction is made in Widdowson ( 1977). By drawing certain distinctions (e.g., that between linguistic and social roles) or by providing descriptive tools. Cf., Gremmo, Holec and Riley, 1 977. ' Although for some kinds of written discourse-e.g., an exchange of correspondence-a very similar description could be used, the main differences being along the other parameters and in realisation. ' For a detailed discussion of these matters see Stevens, 1974, from which this paragraph is taken. z

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

LE S OC I AL P E S E LOURD S UR LE D I S COURS, UN CAS D'I NE G ALI TE I NTERACTI O NELLE C H R I S T I A N B AC H MA N N

Universite Paris XIII A BSTRACT

L E mode d'approche du 'discours'-au sens de langage utilise dans les situations sociales quotidiennes-developpe dans les pages qui suivent, risque d'etre en decalage par rapport aux modes d'apprehension habituels. C'est en effet un point de depart proprement sociologique qui sera privilegie, sans que ses implications linguistiques en soient integralement tirees. La tradition d'analyse de discours a, Ia plupart du temps, repris sans le critiquer un des postulats de base de Ia linguistique: celui de Ia superposabilite du locuteur et de l'auditeur. Le fameux 'circuit de Ia parole' introduit par Saussure au debut du Cours ('Soit A et B . . . ') n'a pas ete fondamentalment remis en cause lorsque Ia linguistique a decouvert Ia theorie mathematique de Ia communication. Le postulat de symetrie, qui veut que l'emetteur puisse devenir recepteur et vice-versa, est demeure inchange. Ce qui est entendu par 'communication' est un echange entre entites socia/ement abstraites. n ne s'agit en aucun cas de nier Ia fecondite de ce point de vue: Ia linguistique est nee de cette reduction. Dans le cas de l'analyse des discours sociaux, il n'est pas evident que cette option ait Ia meme valeur heuristique. Une hypothese radicalement differente peut etre posee. C'est ce a quoi s'exerce P . Bourdieu, entre autres chercheurs (Bourdieu 1 977). Contre Ia philosophie analytique, il affirme que les mots n'ont pas de force intrinseque: leur pouvoir essentiel leur est donne de l'exterieur. 'Parmi les presupposes de Ia communication linguistique, qui echappent le plus compli:tement aux Iinguistes, il y a les conditions de son instauration et le contexte social dans lequel elle s'instaure, et en particulier le groupe dans lequel elle s'accomplit' (ibid. p. 21), dit Bourdieu, qui demande ainsi un deplacement de prob­ lematique. L'angle d'attaque theorique n'est pas celui des rapports de commu­ nication, ou d'echange linguistique, mais celui des rapports de force sym­ bolique, qui donnent Ia parole ou roouisent au silence. C'est une perspective

Applied Linguistics, Vol. I , No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

This paper focuses attention on the social factors that influence discourse by associating its development with the symbolic balance of power which gives or denies participants the right to speak. The central concern is with social in­ teractions and the structures within which they take place: with the strategies used by participants when, by means of social rules of discourse at their disposal, they achieve self-expression by clearing a way through the institutional con­ straints imposed by established norms of social behaviour. The author demonstrates the working of such strategies by an analysis of two samples of discourse data from the context of industrial relations in France.

218

LE SOCIAL PESE L O U R D S U R L E DISCOURS

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

semblable qui pourrait ouvrir une direction (et non Ia seule) d'analyse du discours: les interactions sociales, et les structures dans lesquelles el/es s'in­ serent, produisent et modelent le discours. Nulle societe n'etant homogene, l'echange discursif est, dans tous ses aspects, fondamentalement inegal . ll ne s'agit cependant pas de rallier, pour l'etude du discours, un point de vue objectiviste, comme tend a le faire Bourdieu dans ses analyses . La dimension de l'acteur et de ses strategies demeure fondamentale. Le point de vue developpe ici est le suivant: une etude du discours dans ses utilisations quotidiennes pourrait etre effectuee en termes d'actions accomplies par des agents sociaux qui construisent, par ces actions memes, leur rapport au champ psycho-social. Cette 'construction' , pour reprendre Ia formule qui court de Schutz aux ethnomethodologues, en passant par Berger et Luckman, s'opere sur Ia base des structures objectives dans lesquelles l'acteur se trouve insere, et qui determinent elles-memes les modes d'apprehension du reel, et les types de reaction qu 'il peut avoir. Par voie de consequence, les structures discursives, et, de fa�on plus complexe, les structures linguistiques elles-memes, seront determinees par le champ psycho-social, mais mises en action par les agents sociaux. Analytiquement, c'est done un double mouvement qui est reconstruit: celui qui, partant des structures globales de la societe, impose a ses membres des normes implicites ou explicites; c'est en particulier le niveau des institutions, qui ne sont rien d'autre, en fait, que des normes ideologiques cristallisees et matenalisees . Le second mouvement est celui qui, issu de la subjectivite des acteurs, se formule dans des regles sociales de discours, qui sont autant de strategies individuelles au service de procedures d'interaction que le sujet met en oeuvre pour se frayer son chemin a travers les contraintes institutionnelles. Si de telles strategies ne sont jamais completement 'libres ' , puisque les in­ stitutions n'en permettent qu'une gamm e limitee, elles n'en demeurent pas moins I' objet d'un choix 'politique' , au sens oil il determine, et oil il est determine, par des rapports de force. Les propositions qui precedent sont allusives, et demanderaient a etre detaillees du point de vue sociologique comme du point de vue linguistique, ce qui ne peut etre fait dans un nombre de pages restreint. Certes, elles tendent a n'envisager le discours que sous I'angle de Ia machinerie sociale qui le produit; cependant, dans le logicisme dominant, une approche teintee de sociologisme ;: ne peut faire naitre que des interrogations salutaires. Deux exemples, brievement resumes, vont tenter de concretiser cette demarche. Tous deux sont issus du domaine industriel. Dans le cadre d'in­ tervention de type ergonomique, un groupe de chercheurs de Paris XIII Oaboratoire 'Communication et Travail') y a applique des procedures d'analyse du discours. Le premier exemple est une na"ation de pratique. La technique utilisee s'inspire du 'recit de vie' au magnetophone tel que les anthropologues l'ont mis au point. Simplement, dans le cas qui nous occupe, c'est le recit d'une pratique particuliere qui est demande a une ouvriere Jeanne: sa premiere prise de parole en public. Cette ouvriere travaillait dans une usine qui a connu la notoriete des media, en France: Lip, une usine de montres du Jura. Rappelons en deux mots le contexte. En avril 1973 , l'usine est mise sous tutelle j udiciaire, apres depot de

C H R I ST I A N B A C H M A N N

219

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

bilan . ' N i l'emploi, ni l'integrite d e l'entreprise n e soot garanties' , proclament les administrateurs designes. Les ouvriers de Lip, refusant cette decision, s'organisent en comites, montent a Paris , prennent comme 'otage psy­ chologique' 500 millions d' A. F.- de montres qu'ils vont vendre a leur compte, organisent une solidarite active a leur action par la tenue de meetings, negocient directement avec M. Giraud, 1 le mediateur designe par le gouver­ nement, et remettent en marche la production. Cet evenement exceptionnel, meme si il a ete loin de satisfaire toutes les utopies dont il etait porteur, n'en a pas mains fait emerger, pour le personnel de production, une situation nouvelle. De telles moments, ou l'on imagine que 'tout est possible' , comme certains l ' avaient imagine lors des greves de 1 936, creent la possibilite de n!gles d'interaction nouvelles. Le discours de Jeanne decrit le processus qui l'a mise en situation de prendre la parole. La mise en marche du processus ne peut qu'etre difficile. On peut dire des femmes ce que Sacks disait des enfants (Sacks 1 972): leur droit social a la parole est limite. Pour une ouvriere, cette formulation est d'autant plus vraie. Decrite par Jeanne, sa prise de parole s'effectue scht:matiquement en trois temps: -tout d'abord, !'integration a un systeme de contrainte institutionnelle. Le cadre materiel du meeting, et les relations de face a face qu'il amene, suscitent les possibilites initiates de la parole. Jeanne pourrait cependant rester muette, et reduire son role a celui de figurante. La premiere prise de parole est fondee sur une auto-categorisation valorisante ('ils demandaient des questions pratiques' . . . 'j 'ai commence a me lancer'). Le deni de competence ('s'ils commen�ent a demander une question economique surtout ou politique') constitue un interdit interactionnel d'une efficacite eprouvee dans le domaine industriel. D'autre part, la sociologie de la deviance a mis en evidence la relation qui unissait la categorisation que l'on donne d'autrui et celle qu'il adopte de lui-meme (Cf. 'labelling theory'). L'acquisition, aux propres yeux de Jeanne, d'un domaine d'expertise qui lui est reconnu-quel que soit l'aspect initialement exigu de ce domaine-assure une premiere experience positive. -enfin, Jeanne pourrait n'avoir pris la parole qu'a titre exceptionnel, et son audace pourrait rester sans lendemain. Une pedagogie de la gradation, meme si elle se revele quelque peu sauvage-assure le controle strategique de Ia situation. Son cadre de perception , en un sens different du concept goffmanien de Frame, lui fait considerer qu'une place lui est assuree en tant qu'acteur dote de capacites d'intervention discursive. Ce cadre regroupe de maniere quasi­ systematique les elements que Hymes detaille dans sa description ethnographique de la parole (Hymes 1 974). On y trouve en particulier une composante comme la 'scene' (scene) qui designe le contexte psychologique ('psychological setting') de l'acte de langage: 'la salle etait sympathique' . . . 'des femmes qui venaient d'une usine de Lyon'). Meme les intentions ('out­ comes' dit Hymes) des participants soot analysees par Jeanne: elles venaient non pas pour acheter des montres a bas prix mais pour s'informer. Le jeu interactionnel qui en resulte ('un dialogue, quai') est lui-meme le produit de ce cadre, en terme de procedures d'interaction sequentiellement organisees. Dans le deuxieme exemple, la prise de parole n'est pas racontee, elle est

220

L E SOCIAL PESE L O U R D SUR L E DISCOURS

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

saisie sur le vif . Deuxieme difference: le processus semble avorter. ll s'agit d'un fragment de reunion, d'une duree de quelques minutes. La situation est Ia suivante: dans une usine de savon de Ia Region parisienne, une experience d'enrichissement aes taches a ete decidee par Ia direction generale. Une nouvelle organisation technique vise a donner aux ouvriers du temps libre pour qu'ils assurent des operations simples de gestion de leur production. La direction demande done aux contremaitres de reunir les ouvriers, et de leur donner, sous Ia forme de tableaux commentes, des indications sur Ia gestion de l'usine. Les contremaitres soot plutot favorables a !'experience, et l'un d'entre eux adhere meme a un syndicat qui se reclame de l'autogestion. Les ouvriers, faiblement syndiques, ne temoignent ni enthousiasme extreme, ni hostilite a l'egard de cette innovation. Dans un sentiment proche de !'adhesion a eclipse' decrite par Hoggart, ils attendent. La reunion d'information, preparee par les contremaitres avec le chef de personnel et le chef d'atelier, se deroule un matin, pendant le travail, dans un local proche des chaines. L'analyse de ce fragment montre que Ia hierarchie ne tisse pas, dans les relations qu'elle etablit, un continuum qui irait de l'autorite au laissez-faire. A cette conception, frequente chez les psychologues et les pedagogues, il est preferable de substituer, comme dans le cas precedent, celle de structure d'interaction, institutionnellement reglee. Cette structure, en tant que realite sociale, imprime ses modes relationnels aux contremaitres, que ces derniers soient autoritaires ou liberaux, en ne leur laissant qu'une marge de manoeuvre etroite. Dans les institutions comme dans Ia vie quotidienne, ce soot de tels mecanismes qui expliquent ce que Ia sociologie des organisations designe comme 'effets pervers' ou 'effets contre-intuitifs ' : les intentions des acteurs, lorsqu'elles s'actualisent dans des institutions, semblement magiquement deviees et obtiennent des resultats inverses de ceux qui etaient attendus. L'ecole est un des champs de predilection ou les sociologues ont pu reperer ces effets pervers. La transmission d'inforrnation est traduite par les contremaitres dans Ia forme d'interaction qui leur est familiere: celle de Ia pedagogie Ia plus traditionnelle, qu'ils ont eux-memes vecues dans leur jeune age. La structure d'interaction qu'ils adoptent est done celle d'une classe , avec les regles discursives que cette structure implique, et que de nombreux travaux d'analyse du discours pedagogique comme celui de Sinclair et Coulthard, ont mis en evidence. Leur maniement malhabile a Ia fois des contenus techniques et du savoir-faire pedagogique jette ce que Goffman appelle un discredit sur leur representation. Lorsqu'un ouvrier, face a ce rituel, ressent !'impression d'etre dupe, et proteste ('non je ne suis pas d'accord c'est pas evident il faut le decompte'), Ia construction de !'interaction s'est deja effectuee de fa�on telle que c'est le contremaitre le plus autogestionnaire qui menace de rompre !'experience ('si � ne vous interesse pas les prochaines fois on en reparlera pas'). La protestation, meme sous forme de demande de precision, d'un ouvrier est inacceptable dans ' l'equilibre precaire du rapport de force edifie par les contremaitres. Leur superiorite interactionnelle deja chancelante est menacee. Leur sanction sera Ia reduction des ouvriers au silence. D'un commun accord, les reunions seront supprimees. Ces reflexions sur l'entreprise n'ont sans doute qu'un interet limite pour des

C H R I ST I A N BACHMANN

221

specialistes de l'apprentissage des langues secondes. Cependant, !'acquisition d'une langue etrangere est aussi un cas d'inegalite interactionnelle. L'inegalite est certes d'une nature toute differente. ll n'est plus question, dans ce cas, de Ia subordination hierarchique et des droits limites a Ia parole qu'elle impose, mais bien plutot de Ia meconnaissance linguistique et culturelle (au sens an­ thropologique du terme) et des possiqilites limitees de moyens d'action qu'elle implique. Par exemple, il pourrait etre pedagogiquement eclairant d'examiner quelles strategies developpent des acteurs sociaux pris dans de telles structures

NOTES 1 Nom mentionne par Jeanne dans son recit (Cf. annexe 1).

REFERENCES

Bourdieu, P . 1977. L'economie des echanges linguistiques, Langue Fran�aise, n • 34, pp. 1 7-34, Larousse, Paris. Goffman E. 1 967. Interaction Ritual. Essays on Face-to-Face Behavior; Doubleday and Co. Anchor Books, Garden City N.Y. Hymes D. 1 974. Foundations in Sociolinguistics, U. of Pennsylvania Press, Philadelphia. Sacks H. 1 972. On the Analysability of Stories by Children, in Gumperz and Hymes 1 972, Directions in Soczolinguzstzcs, Holt, Rinehart and Winston. ANNEXE 1

Interview d'une ouvriere de Lip (greve de 1974)

1

-ohlouilmais au debut que j 'allais au meeting moi j'etais drolement genee puis ennuyeel . . . en etait parti a quatre . . . heu . . . I quatre . . . trois hommes et moi I pour preparer les six heures a ( . . .) et puis le premier meeting c'etait le soir a Ia bourse du travail I j 'etais dans le hall je savais qu'il y avait deja plein de gens dans Ia salle qui attendaientlon s'est assis au milieu des . . ./des . .I rangslet puis I au face a face de pres I OH: j 'etais mais alors hein: : I il n'aurait pas fallu qu'ils me disent seulement de me presenter je n'aurais meme pas pu dire mon NOM I . . tellement que ca se nouait ici I I ah je savais bien que je pouvais etre coincee s'ils commencaient a I demander une question economique surtout II ou politique I mais quand il a I quand ils demandaient .

.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

d'interaction. Goffman (Goffman 1967) expose que les conversations quotidiennes supposent une obligation d'engagement. QueUes strategies des etudiants etrangers, places dans une situation d'inferiorite interactionnelle, peuvent-ils mettre en oeuvre face a cette obligation d'engagement le repli sur soi et le mutisme, que mentionnent Peter Harder dans l'article inclus dans ce numero, mais aussi le repli sur !'interaction Oe doute systematique et !'in­ flation metalinguistique, ou alors le discours vide decalque du manuel), le repli sur autrui O'election d'un interlocuteur privilegie auquel on s'accroche), !'engagement simule Oa comprehension feinte) ou le surengagement Oe discours logorrheique et incomprehensible). Frequemment, ce sont des strategies mixtes qui sont adoptees. Leur inventaire pour telle ou telle population, et leur prise en compte dans Ia pedagogie, permettraient peut-etre d'alleger le poids du social pour les apprenants d'une langue seconde.

L E S O C I A L PESE L O U R D S U R L E D I S C O U R S

222

des questions pratiques II c'est I a que j e me suis I que j'ai commence a me lancer I par exemple sur I heu I comment I heu I fonctionnent vos commissions -oui -ben Ia I je me suis remis en memoire Ia salle avec les . . .I pieces de chaque cote I tout ce que les gens faisaient I et voila j ' /j 'ai commence a raconter comme � I a partir de trues pratiques I autrement si je commen�ais par Ia politique l' economique que je ne connais pas ASSES II -alors Ia premiere fois tu as parte longtemps -oh non (petit rire) . . . une phrase I j 'ai dit que . .I c'etait/pour Giraud I j 'ai dit que . .I il nous avait dit que/ les ces soixante licencies I heu . . .lseraient repartis a l'exterieur mais sans garantie . . .I sans VRAIMENT . . .I heu I une garantie I heu I absolue I de I rentrer a l'usine APRES/ alors que c'etait vraiment I PAS I pas positif I voila ce que j 'avais dit surtout I I (c'est tout ce que j 'ai dit) I le soir ils m'ont fait peur I X m'a dit demain je te previens heu/ moi je dis pas un mot c'est toi qui tiens le meeting entierement I hein -et tu l'as tenu -non non il s'est trouve un meeting I heu I Ia salle etait sympathique le decor etait sympathique I les gens il y avait beaucoup de femmes I dans cette salle il n'y avait que des femmes I des SIMA de Lyon qui etaient en greve I qui sont I des petites jeunes filles qui ont raconte un petit peu I ce qu'elles faisaient comment � se passait elles recevaient les clients dehors I les clients venaient pour s'informer I mais pas pour acheter I . . . a l'interieur c'etait pour s'informer pour soutenir les grevistes I alors � m 'a donne un peu con fiance I hein I �a m 'a donne un peu I �a m 'a II et puis I I apres I . . . c'est Ia qu'elles ont demande comment fonctionnent les commissions . . . elles s'adressaient VRAIMENT a moi voyez que . . . elles causaient a; Ia femme Ia I je me suis dit oui �a va aller I c'est pratique I je vais pouvoir repondre . . . puis je ne me suis pas levee alors c'est pas si intimidant de se lever deja I c'est un DIALOGUE quoi. .

.

1 L e code d e transcription est proche de celui adopte pars les ethnome thodologues (Sacks, Schegloff, Schenkein, etc.): I etlI pauses pause de 4 secondes (4) NOM segment accentuee c'est allongement segment non interpretable interpretation supposee (je crois) commentaire du transcripteur. ((toux))

A N N EXE 2

Reunion d'atelier (usine de savon)

Contremaitres: Marcel Andre 40 ouvriers, dont Julien

Marcel:

Andre:

) 1 59 000 F depenses ((toux)) I il y a Ia remise en etat de Ia P ( 1 I 50 000 Ia chaine et ils ont achete une chaine pour l' I. W. K. vous vous souvenez qu'elle i:tait completement secouee celle-la I et puis le moteur de la broyeuse I pour Ia P 4 I pour Ia P 5 et Ia C 4 I est-ee que vous avez bien pige ce tableau-la?

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

NOTES

CHRISTIAN BACHMANN

Marcel:

Andre: Julien: Marcel: Julien: Marcel: Julien:

Marcel: Julien: Marcel : Julien:

Marcel : Andre:

ce que ce qui est prevu au budget II ce qui est prevu au budget vous voyez ce qui a ete realise hein et ce qui a ete depense en plus III (4) en pourcentage parce que finalement ( ) non je suis pas d'accord (1) c'est pas evident (1) il faut le decompte pardon? faudrait un decompte total heu pour savoir exactement COM­ MENT s'est passe alors si vous voulez un decompte total heu heu je crois que: : (en fin d'apres-midi) on sera encore Ia no:1 non mais non de toute fa�on c'est I c'est normal I par exemple ou j 'habite j'ai demande a voir le les livres de compte on me l'a refuse I evidemment ce serait trop long ben ecoutez vous pouvez toujours ( ) (� revient au meme hein) non mais non mais je crois ce qui compte non mais non mais c'est trop facile de dire aux gens voila c'est pas �a c'est �a et pas autre chose ah mais n'importe comment les details I les details sont chiffres hein ben je m'en doute I non mais je voudrais vraiment voir a ce moment-lei ( ) ici hein c'est pas facile hein non mais � je m'en doute I vous avez evidemment mais I il en aurait pour 24 H I enfin de toute fa�on c'est : : : histoire de savoir ) 328 par quoi de de voir tout/ pour justifier II les chiffres ( exemple ( ) non mais heu il est certain que !'on vous donne des chiffres comme �a et c'est pour ca que je posais Ia question de savoir si II ca peut avoir un interet voila est-ce que ca peut avoir un interet pour vous I en fait c'est peut-itre pas les chiffres en eux-memes qui ont de !'interet parce que ce qu'on a depense pour faire des ameliorations ou des reparations II c'est simplement pour voir disons Ia Ia Ia philosophie du systeme hein? I on a un budget global d'entretien I et simplement pour savoir si � vous interesse I de connaitre heu : : I dans quels domaines passe cet argent/ I c 'est pour ca que si ca vous interesse pas les prochaines fois II on n'en reparlera pas I I c'est tout bon au contraire II comme vous le dites si � si � vous interesse d'avoir les details//( ) ((brouhaha))

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Marcel: Julien: Andre: Julien:

223

INTERACTIONAL M ARKERS IN DIALOGUE E D D Y R O U L ET University of Geneva

The most important characteristic of the linguistic research carried out during the last 1 0 years into mother-tongue and second-language teaching focused on the learner's communicative needs has been the description of features in the discourse which relate the speaker to his utterance. These features include

These three types of marking have already been studied and discussed as far as French is concerned in a number of publications (see the works by Bally, Benveniste, Anscombre and Ducrot, Martins-Baltar, Perret and Roulet mentioned in the bibliography). However, there is a fourth type which has so far been generally ignored but which plays a fundamental role in dialogue; these we will call interactional markers. In this article we are going to try to describe how these markers function in the dialogues of a novel by Simenon, a type of text which is often used in teaching French as a second language, though, unfortunately, not from a communicative point of view. But before we do so, in order to understand the role of interactional markers, let us make a brief excursion outside the traditional domains of linguistics and language-teaching to take a look at the processes of verbal, face-to-face interaction. According to Goffman (1973), every individual in society is trapped between two necessities: to defend the territory of his self and to establish social relationships with other people. These necessities are difficult to reconcile, since establishing social relations means removing certain barriers, leaving the way open for territorial intrusions . Now if we accept the line of argument put forward by Ducrot ( 1 977) and Wunderlich ( 1 977) that the aim of illocutionary acts is to impose rights and duties on speakers and/or hearers-that is, to modify the deontological relationships holding between the participants in the interaction, -then such illocutionary acts must necessarily imply a loss of face for those participants. The clearest example of a type of act which an interlocutor might consider to be an invasion of his territory is the Request; in order to prevent the oc­ currence of such a socially counter-productive interpretation, the speaker has at his disposal a number of strategies which allow him to change the meaning that otherwise might be given to an act, transforming what could be seen as offensive into what can be seen as acceptable Applied Linguistics, Vol. I, No. 3.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

(i) modalisations (epistemological, evaluative, etc.)-features which mark the ways in which a speaker commits himself to or accepts responsibility for what is predicated in his utterance. (ii) illocutionary markers which make explicit the particular speech act per­ formed by the speaker. (iii) argumentation markers which indicate the type of effect which the speaker wishes to have on the opinions of his interlocutor.

E D D Y R O U L ET

225

or, in other words to

correct for incursive violations (Goffman, 1 97 1 , pp. 1 39, 1 93)

Positive face

Negative face

Speaker

Confession Excuse

Offer Promise

Interlocutor

Criticism Accusation

Offer Request

The speaker cannot admit an error or apologise for having forgotten something without endangering his own image; he cannot offer or promise his help without endangering his independence. In the same way, for the in­ terlocutor, every criticism or accusation is an attack on his image and every offer or request a threat to his territory. It is therefore necessary to develop interactive strategies to reduce this threat, and these strategies are of interest to the linguist and the teacher since they are realised with conventional linguistic forms. According to Brown and Levinson, the speaker who wishes to carry out a FTA such as a Request can adopt one of the four following strategies, his choice being determined by the degree of risk of losing face involved in the act: /

on record Do the FTA

<

� without redressive action, baldly (I)

c:::::.___

with redressive action (II) off record (Ill)

Don't do the FTA (IV) The direct realisation (e.g. , 'Shut the door ! ') is distinguished from the allusive realisation ('There's a draught ! ') by the fact that its intention is clear: the speaker does not risk being misunderstood or taken for a hypocrite-risks which are inevitable in the case of an allusive realisation; on the other hand, a

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Unfortunately, with the exception of Lakoff ( 1 977), even those linguists working on speech acts have paid little attention to the influence of social interaction on linguistic structures, and the ethnomethodologists have not carried out systematic study of the linguistic strategies which allow· speakers to neutralise territorial violations. The only detailed attempt to integrate the linguistic approach with the sociological is Brown and Levinson's work (1978) on face-threatening acts (FTA). Following Goffman, they start from the hypothesis that the main aim of an individual engaged in social interaction is to save face; they distinguish between 'negative face' , which is the need, mentioned above, to defend one's territory, and 'positive face' , which is the need to be recognised and valued by other people. Generally speaking, it is in each participant's interest to avoid causing the other to lose face, since this would endanger his own position. But certain acts are in themselves threatening for the positive or negative faces of speakers and interlocutors, as can be seen in the table below:

226

I N T E R A CTIO N A L M A R K E R S IN D I A L O G U E

direct and unattenuated realisation clearly trespasses o n the interlocutor's territory and risks making him lose face; for this reason an intermediate strategy is both necessary and fundamental in verbal interaction, since it must be one which reconciles clarity of intention with respect for the territory of the other, by combining in the utterance both illocutionary and interactional markers. To the best of our knowledge, the linguistic aspects of these interactional strategies have not so far been studied systematically, nor have they attracted the interest of applied linguists with the exception of an article by F. Roussel ( I 974), where the writer shows the interpersonal role of modalisations, i.e., the

GROUP I

I 2

Je vous remercie (p. 1 1 2) (Thank you) Je m 'excuse de ces precautions (p. I I 3) (I'm sorry about these precautions)

Here we have utterances of Maigret's which are marked by the presence of potentially performative verbs used as performatives (cf. Roulet I978) and which determine the value of the act. The very nature of the acts in question tells us why the direct form is used-they are acts bearing on the speaker's behaviour which could not possibly threaten the face of the interlocutors and where, consequently, no recourse is had to any process of attenuation. GROUP 2

3

-Je VOUS prie de VOUS presenter demain a dix heures au Quai des Orfevres . . (Kindly .come to the Quai des Orfevres at 1 0 tomorrow .

. . .)

-Et d'ici Ia? (And in the meantime?) -Je vous prie de ne pas quitter cette maison, ou un de mes inspecteurs continuera a vous surveiller. (p. 129) (Kindly do not leave this house, where one of my officers will continue to guard you.) The form is identical to that of I and 2 with a performative verb determining the value of the act, but its occurrence is more surprising here, in what is, after

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

way in which the speaker uses them to establish or maintain a certain type of interaction with his interlocutor. Using Brown and Levinson's model, let us now try to describe in more detail the nature of the three types of marking set out above and how they function in the dialogues of one of Simenon's novels, Maigret et /'Affaire Nahour, Oeuvres Completes, Vol. XXIV, 7- I 54, Lausanne, Editions Rencontre, I 969. We will first look briefly at direct and allusive speech acts (which correspond to strategies I and III) before considering in greater detail those acts which make use of interactional markers. Direct speech acts are characterised by the presence of an illocutionary act marker, such as a performative verb or syntactic structur�, which makes explicit the force of the act, and by the absence of any attenuating in­ teractional marker. The examples in our corpus can be divided into three groups, according to the type of act and the type of interaction between the participants (here we will simply quote a limited number of typical examples).

E D D Y R O U L ET

227

all, a request, the clearest sort of FfA. Of course, the direct form is explained by the fact that we are dealing here with a certain type of interaction-the speaker, a police officer engaged on a case, is exercising his authority over an interlocutor who is under arrest and whose face no longer has to be saved. GROUP 3

4

Tu Vas' te rendre au boulevard Saint-Michel . . . Tu trouveras' un certain bar des Tilleuls (. . .) Dis au patron que tu viens de Ia part de Lardois et qu 'on ne cherche pas' a lui faire des ennuis . . . Essaie seulement de savoir si Fouad Oueni est aile au cercle Ia nuit demiere (. . .) (p. 68-69) -

In the third and fourth utterances in this example, the role of illocutionary marker Request is played not by a performative verb but by the syntactic structure (imperative). Here, too, the occurrence of the direct form, which at first glance seems a somewhat surprising choice for an FfA such as Request, is explained by the type of interaction in question. The person making the request, Maigret, wields authority over his colleagues and there is even less need for attenuation since, unlike the preceding example, this enters into a framework of routine hierarchical relationships. This is a special characteristic of the relationships between Maigret and his subordinates and explains why we find allusive forms used so much; e.g. , Tu Vas' te rendre . . . Tu trouveras'. These forms are allusive because the utterances contain no illocutionary marker of Request (except, perhaps at the prosodic level, where the markers are vague and rarely have a single, clear illocutionary value). In general, allusive acts run the risk of being misinterpreted (sometimes deliberately) by the interlocutor and Maigret would certainly not have addressed the accused in example 3 in this way; but with colleagues he works with regularly, this risk is excluded because the interpretation to be given to the act is given by the context of hierarchical relationships and routine professional contacts. Just as the presence of a performative verb making explicit the illocutionary value of an act seems necessary to avoid any misunderstanding when a new relationship based on authority is established (police officer/accused), so it would seem preposterous when carrying out routine acts (at least as far as exercitive acts are concerned). We can summarise our brief analysis of these three first groups of ut­ terances, then, by saying that the absence of attenuating interactional markers is related to the type of act and in the case of FfA's, to particular types of interaction, such as the establishment of a new hierarchical relationship or the carrying out of routine tasks. Let us now go on to look at those speech acts which correspond to Brown and Levinson's Strategy II. These acts are the most interesting of all from the interactional point of view, since the simultaneous presence of both illocution­ ary and interactional markers means that they are able to express the speaker's intention clearly whilst at the same time respecting the interlocutor's face.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

(You're going to go to the Boulevard St. Michel . . . You'll find a bar called 'The Lime Trees' . . . tell the boss that Lardois sent you and you're not there to make trouble . . . Just try to find out if Fouad Oueni went to the club last night . . .)

I NT E R ACTIONAL M A R K ERS I N D I A L O G U E

228

These acts fall into four further sub-groups: GROUP 4

5

Passe-moi Pierre Nahour, veux-tu? (p. 76)

(Put me through to

Pierre Nahour, will you?)

6

Reste avec e/le, veux-tu ? (p. 39) (Stay with her, will you?)

GROUP S

7

Pouvez-vous me dire ce que sont ces listes de chiffres?

(p . 49)

(Could you tell me what these lists of figures are?)

8

J'aimerais que vous restiez ici aussi, car . . .

(p. 62)

(I would like you to stay here too, because . . . )

9

Voulez-vous raccrocher, s 'il vous plait?

(p. 53)

(Would you ring off, please?)

10

Vous voulez bien m 'appeler un taxi, Pierre?

(p. 93)

(Call a taxi for me, would you, Pierre?)

Although these utterances are obviously meant and interpreted as Requests, unlike the preceding examples they contain no illocutionary marker (such as a performative verb like prier or an imperative structure). It is generally accepted that, in examples like 7 and 8, the illocutionary value Request is derived from the literal illocutionary value Request for Information or Assertion through the application of Grice's conversational maxims (1975) and of the principle put forward by Heringer (1972), Gordon and Lakoff ( 1 973) and Searle (1975) according to which indirect acts can be realised by calling into question or by asserting one of their felicity conditions. (See Roulet ( 1 977) for a discussion of how this applies to French). But this ex­ planation is inadequate: according to Grice's principle of co-operation, to which all the writers mentioned above refer, it is only if the literal in­ terpretation is irrelevant in a given communicative event that the interlocutor is led to infer the derived illocutionary value Request; now in modern French it is obvious that the form pouvez-vous is a conventional form for a request and that it is unquestionably interpreted as such by speakers without any recourse to processes involving pragmatic inference. Moreover, this expression here plays a role which is just as important as that of the content (a point which is missed by the conversational postulates of Gordon and Lakoff, 1 973), since one cannot call the interlocutor's capacity into question using any old form ; the utterance

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

This case is the clearest and the simplest to analyse because the utterance in­ cludes both an illocutionary marker, the imperative structure, which directly determines the value Request, making the speaker's intention explicit, and what we call an interactional marker, whose function is to neutralise the territorial violation implied by any request. Veux-tu ? (will you?) as used in these utterances, is not just a simple request for information which could be replaced by the form est-ce que tu veux? (do you wish?) and to which one could reply Je ne veux pas (I do not wish); rather, it is a reference to the in­ terlocutor's free will which, however formal, is aimed at attenuating the threat to his face contained in the act of requesting. We can call interactional markers which refer to the interlocutor's will in this way markers of deference.

E D D Y R O U LET

229

1 1 Etes-vous capable de me dire ce que sont ces listes de chiffres? (Are you capable of telling me what these lists of figures are?)

semantically synonymous with 7 is not so pragmatically because it is in­ terpreted spontaneously as a Request for Information. If we are going to account for these characteristics, we have to make use of and integrate the notions of convention of usage and marker of il/ocutionary derivation. Morgan (1978) distinguishes between conventions of language which determine the literal sense of words and utterances, and conventions of usage, which regulate the uses of utterances for specific ends. In the case of the utterance

7

Pouvez.-vous me dire ce que sont ces listes de chiffres?

·

12

Voulez.-vous prendre ce livre? (Do you want to take this book)

is easily interpreted as a request for information.

13

Voulez.-vous prendre ce livre, s'il vous plait ?

(Will you take this

book, please?) or

1 4 Voulez.-vous bien prendre ce livre? (Will you please take this book?)

can only be Requests, s'il vous plait and bien acting as illocutionary derivation markers. Another interesting example of an illocutionary derivation marker is the conditional, which results in an assertion being interpreted as a Request, e.g . ,

1 5 J'aimerais qutJ tu m 'embrasses. (J'aime que tu m 'embrasses.) (I would like you to kiss me. I like you kissing me) (Diller., 1977) or utterance 8 in our corpus. Moreover, if we compare 7 and 9, which, as we saw, are literally synonymous, we will note that only the first has the conventional illocutionary value Request: which leads us to conclude that it is the form pouvoir which plays the role of the illocutionary derivation marker of request. These indirect speech acts which are characterised by the presence of an illocutionary derivation marker satisfy perfectly the apparently contradictory

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

the native speaker, applying the conventions of language knows -that the ut­ terance has a particular literal sense; but he also knows that Pouvez.-vous is a standard means of expressing an indirect request, following the conventions of usage in our society. Consequently, he has no need to make use of inference; he takes a short-cut, whereas a foreigner might well follow the longer route. For the linguist, the concept of convention of usage is only interesting in­ sofar as he can relate it to specific linguistic markers, that is, only if he can accept Brown and Levinson's statement (1978, 265), that 'some aspects of usage-namely pragmatic inferences of an implicatural sort-are likely to acquire structural correlates ' . This is where the concept of illocutionary derivation marker (introduced by Anscombre (1977) and taken up by Diller (1977)) comes in: it denotes those elements of the utterance which result almost automatically in a derived interpretation. Classical examples are: s'il vous plait or bien. While

230

I N T E RACTIONAL MARKERS I N D I A L O G U E

GROUP 6

16 Puis-je vous demander votre nom exact? (p. 43) (Can I ask you your full name?)

17 Puis-je vous demander depuis combien de temps vous travaillez pour cette famille? (p. 84) (Can I ask you how long you have been working for this family?)

In these utterances, the use of the deontic pouvoir in the 1 st person present interrogative form specifies a literal illocutionary value, Request for per­ mission : 1 8 Puis-je sortir? (May I go out?) 1 9 Est-ce quejepeux aller au cinema? (Can I go to the cinema?) But here, the literal request for permission concerns a speech act, a Request. Often, it applies to other types of act: 20 Puis-je vous offrir un whisky? (Can I offer you a whisky?) 2 1 Puis-je vous suggerer de prendre un. taxi? (Can I suggest

that you

take a taxi?)

22 Puis-je vous rappeler que nous fermons a 6 h. ? (Can I

remind you

that we close at 6 o 'clock?)

In practice, these utterances are interpreted directly as Requests , Offers, Suggestions or Reminders. . However, contrary to the examples in Group 3 , it is not the modal which plays the role of illocutionary derivation marker, since it appears in the realisation of different FTA's, but the performative verb. Here it is pouvoir and interrogative structure which is the illocutionary marker of Request for Permission and which plays the role of interactional marker of deference. But, unlike the markers of deference in Groups 4 and 5, this marker refers not to the act which the interlocutor is requested to carry out, but to the speaker's own act of Requesting. Further markers of deference which appear in the corpus are 23

Veuillez lui dire que je desire Ia voir (p. 1 33) want to see her)

(Would you tell her I

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

requirements of FfA's (namely, to express the speaker's intention clearly and to preserve the interlocutor's freedom of action). Due to the presence of the derivation marker, the derived illocutionary value is clear, but at the same time, the very occurrence of the derivation process is enough to act as an in­ teractional marker which attenuates the Request. In 7, 9 and 10, the at­ tenuation resulting from the presence of an illocutionary derivation marker is emphasised by the presence of an interactional marker of deference. The presence of interactional markers in the acts in Group 4 is required both by the type of act and by the type of interaction, which is very different from those exemplified in the first three groups. Although we are still dealing with the FfA Request, in this case they are addressed by Maigret to interlocutors who, for the moment at least, are only witnesses (7, 8, 9) or by a distinguished woman to her brother-in-law (10), relationships where care has to be taken to preserve the interlocutor's face.

23 !

E D D Y R O U L ET

which belongs to the first type, and

24 Permettez-moi de vous repondre, M. le Commissaire, que le piege est grassier . . . (p. 1 1 5) (Allow me to reply, Inspector, that the trap is a crude one)

GROUP 7

25 Je dois vous demander Ia permission de vous suivre a l'interieur (p . 88) (I must ask your permission to follow you inside) In this utterance neither the deontic modal verb nor the performative fulfills the conditions of usage which would allow it to operate as the illocutionary marker of an exercitive act; therefore, the utterance has as its literal value Assertion, and its illocutionary value Request for Permission is derived. If we look at similar realisations of other FfA's such as Warning or Request

26 Je dois vous avertir que je me vengerai (I have to warn you that I will have my revenge)

27 Je dois vous demander de sortir immediatement (I have to

ask you

to leave immediately) we can see that it is the performative verb, functioning as an illocutionary derivation marker; which determines the illocutionary value o f the act and that devoir plays a role which is modal, since the speaker uses it to express a certain attitude to the act which he is carrying out. As with the utterances in the last three groups, the indirect form of the act and the occurrence of the process of derivation together play the role of in­ teractional marker of attenuation, which is reinforced here not by a marker of deference but by a hedge (cf. , Fraser 1 975). This is the term used for all those procedures for attenuating FTA's where a speaker expresses reservations of some kind or another about the language act he is performing. Hedges are not always deontic modalisations (as in 26 and 27), they can also consist of temporal modalisations e.g . ,

2 8 Je vous demanderai, Monsieur Oueni, de n e pas quitter l'im­ meuble avant queje vous en donne l'autorisation (p. 49) (I shall have to ask you, M. Oueni, not to leave the building before I give you permission). which is quite obviously a Request at the moment when it is uttered, but which is attenuated by the future marker. Similarly

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

which belongs to the second (-this is not a Request, it is a Reproach, which is also a threat to the interlocutor's face). With the exception of 5 and 6, which are rendered more direct by the presence of an illocutionary marker, the other utterances are equivalent from the interactional point of view and they all neutralise the threat to the face of the interlocutor which the act of Requesting (or Reproaching) represents . They all also differ in another way which is important from the communicative point of view, _which is their discursive value, that is, the range of verbal reactions which they leave open to the interlocutor, but this is a topic which is outside the scope of this article.

232

I N T E R A C T I O N A L M A R K E R S IN D I A L O G U E

29

Je m e permettrai de vous adresser une remarque, Mme. Nahour; c'est qu 'il n •y a personne pour conjirmer vos declarations. (p. 1 10) (I'll tell you one thing, Mme. Nahour, and that is that there is no one to confirm what you have said)

30 Dois-je vous demander de vous depecher? (Must I ask you to hurry up?)

Instead of being an attenuated Request, this is a Request aggravated by a reproach or even a challenge, and this leads us to interpret dois-je as an in­ teractional marker of aggravation. (See Roulet 1 980). It seems reasonable to suppose that the predominance of interactional markers of attenuation or of interactional markers of aggravation in a given corpus is related to the type of relationship which holds between the par­ ticipants. The constant personal conflicts which we find in the Labov and Fanshel corpus occurring between the patient and her mother and the patient and the psychiatrist increase the number of aggravation procedures as com­ pared to the cool, professional relationships holding between the protagonists in Simenon's novel. For this reason, it would be interesting to carry out a systematic study of these two fundamental types of interactional markers in different types of dialogue. 1 It is worth asking what the relevance of these interactional markers is to first and second-language teaching. The recognition of the role of speech acts in communication and of their wide variety of realisations in any particular language (cf. , Un Niveau Seuil) can only confuse the teacher if he is not aware of the interactional and discursive values of these acts and of the ways they are marked in utterances . In this area, as in so many others , the teacher does not have to wait until linguists have produced descriptions based on a highly sophisticated theoretical framework (there is clearly an awfully long way to go as far as discourse analysis is concerned); a simple framework such as the one

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

This brief examination of the realisations of certain FTA's in the dialogues of one of Simenon's novels shows that in addition to modalisations, illocutionary markers and argumentation markers, we also have to recognise the existence of interactional markers, whose function is to regulate face-to-face interaction and in particular, in the examples we have been looking at, to attenuate face­ threatening acts. The presence in the utterances we have examined of forms such as . . . veux-tu?, Voulez-vous . . . ?, Pouvez-vous, Je dois . . . or the future Je vous demanderai . . ., can only be explained by considering the usage of these utterances within an interactional framework such as that developed by Brown and Levinson. At first sight , this framework is adequate since it allows us to account for the interactional markers of a limited corpus. We should be careful to note, however, that Brown and Levinson deal only with those strategies which aim at making an interaction more harmonious, at reducing conflicts. But besides these attenuating procedures, which are certainly very important, we also find aggravating procedures such as those studied by Labov and Fanshel (1 977) in their description of psychiatrist/patient interviews. No example of an aggravating procedure occurs in our corpus, but it is not difficult to think of a form of Request, similar to those we have studied, such as

EDDY ROULET

233

Brown and Levinson have put forward can help the teacher understand certain aspects of the realisation of speech acts in the dialogues of a type of novel often used in teaching French as a second language. Translated by

PHILIP RILEY

NOTES ' The use o f interactional markers i s not limited t o oral face-to-face interaction, as i s demon­ strated in the letter addressed to authors of these papers by H. G. Widdowson:

I wonder if I could remmd you gently that I would like to rece1ve a copy of your Berne paper as soon asposs1ble . . . The indirect form of the request is itself an interactional marker of attenuation (or even two, if we m

an act of Remindmg which is expressed by a per­

formative verb whose illocutionary value is attenuated by three hedges, can, the tense or mood of

can)

and I

wonder if.

This hyper-attenuated form can only be understood in the context of the

mteraction: it is not just a simple Request, it is a Reminder sent out on the 21st March, concerning a Request which had not been complied with by the time the deadline was reached ( 1 5th March); now Labov and Fanshel (1977) have clearly shown that any reminder of a request which has not been complied with automatically acquires the extra illocutionary value of a Reproach. Hence the Plethora of interactional markers, aiming to neutralise the double threat-to the positive and negauve faces of the recipients-which a request aggravated by a reproach can 1mply.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

mclude the adverb) and it is embedded

CONCEP TUAL AND COM M UNI CA TIVE F U N CT I ONS I N WRI TTEN DIS COURSE H. G. WIDDOWSON Umverszty ofLondon

. . . communication is one of the functions of language. In communicating we express our thoughts in the hope that the listener understands what we are saying. We may be hoping to persuade him, to inform him that we believe such­ and-such, and so on. The function of language for the expression of thought is not "opposed" t.o its communicative function; rather it is presupposed by the use of language for the special purposes of communication. (Parret 1 974:52) Applied Linguistics, Vol. I, No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

All movements which attempt to set up a new scheme of values, whether these be political or pedagogic or whatever, are subject to distortion and excess. Practical action requires the consolidation of ideas into simple versions which can be widely understood and applied. This is a necessary process if the movement is to have any kind of stability and substantial effect in the practical domain. We cannot all the time be tinkering with theory 'besely seking with a continuell chaunge' , as Sir Thomas Wyatt puts it. The problem of application is: how can we consolidate without misrepresentation? How can we prevent our simple versions from being misleadingly simplistic? The movement I am concerned with here is that which proclaims the primacy of communication in language. Its manifesto (to pursue the metaphor}, which can be collated from a range of writings by different hands, contains expressions like 'notions' , 'functions' , 'speech acts' and assertions like 'There are rules of use without which the rules of grammar would be useless' (Hymes), 'The object of linguistics must ultimately be the instrument of communication used by the speech community' (Labov), ' Languages are learned for the purposes of communication' (Wilkins) etc. I am a member of this movement myself, and have indeed contributed to its general manifesto. But I am beginning to get a little worried about how it is being interpreted and implemented. There are signs, I think, of distortion and excess in the un­ derstanding of ideas and their application to practical pedagogy. What I aim to do in this paper is to reconsider certain issues relating to a communicative approach to language teaching, with particular reference to written discourse, and so to rewrite part of the general manifesto in my own terms . I intend this to be not another example of continual change but a consolidation. The first problem to tackle is a terminological one. The term 'com­ munication' is commonly interpreted in a narrow sense to refer to what language users do with their language when engaged in social activity. That is to say, there is a tendency to equate it with 'categories of communicative function ' , illocutionary acts: promises, requests, orders, descriptions, definitions and so on. But as Chomsky points out, the ability to communicate ill this sense presupposes the ability to use language in the formulation of thoughts:

H . G . W I D DO W SON

235

Language serves for the expression of 'content': that is, of the speaker's ex­ perience of the real world, including the inner world of his own con­ sciousness . . . . In serving this function, language also gives structure to ex­ perience, and helps to determine our way of looking at things, so that it requires some intellectual effort to see them in any other way than that which our language suggests to us. {Halliday 1970: 1 43) The second function serves a social purpose. The individual is necessarily

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

The dependence of communication, in a wider sense, on this conceptual function of language has not always been fully recognized. Thus although Wilkins talks about semantico-grammatical categories, which are conceptual, as well as categories of communicative function, it is the latter which have attracted most attention and their relationship with the conceptual categories has been largely left unexplored and unexplained. Again, in Searle's work, it is the illocutionary element in the speech act that has been seized upon, even to the extent in some quarters of assuming an equation between speech acts and illocutionary acts . The proposition, the conceptual element , in the speech act, has not been very much considered. This neglect is to some extent sanctioned by Searle himself since he represents the proposition as essentially only a condition on the effective performance of the illocution: it serves, as it were, a facilitating function . But one could shift the emphasis, as Chomsky appears to do, and say that it is rather the illocution which facilitates the expressing of the proposition. There seems to be no obvious way of deciding, in principle, whether it is preferable to think of the illocution as primary, with the proposition serving as the means for performing it, or of the proposition as primary, with the illocution serving as the means for conveying it. I want to argue that neither proposition nor illocution is primary and that we shall continue to run the risk of distortion and excess if either is given emphasis at the expense of the other. Language is naturally used both for the framing of thoughts and for their conveyance for some purpose in social interaction. The central issue is how these two basic functions operate in communicative use, or to put it another way, how the language user reconciles the operation of these functions in discourse . My agenda is as follows. First a discussion of these functions, then a consideration of how they are realized in written discourse, followed by a suggestion of how they relate to the procedures of reduction and expansion in reading and writing. My general purpose is to reinstate conceptual activity in the context of communication as a whole. There have, of course, been a number of proposals for defining the func­ tions of language-Buhler, Malinowski, Jakobson, Hymes; Halliday have all had a turn at it. I , in my turn, want to suggest that there are (or at least, it is helpful to postulate that there are) two principal functions: the conceptual and the communicative. The first of these functions provides the individual with a means of establishing a relationship with his environment, of conceptualizing and so, in some degree, controlling reality. This is language used for thinking, for­ mulating concepts, fashioning propositions. It is essentially, to use Halliday's term 'ideational' , and it provides for the private security of the individual by enabling him to define his experience. Halliday puts it this way:

236

C O N C E P T U A L A N D C O M M U N I C A T I V E F U N CT I O N S

involved with his fellow men so that he needs language not only t o formulate his ideas but to convey them to others in the process of performing social activity of different kinds. So language has also to have a communicative function so that the individual can do as well as think, can engage in social interaction as well as in private cognitive activity. This function of language is essentially (to use Halliday's term again) 'interpersonal' : Language serves to establish and maintain social relations: for the expression of social roles, which include the communication roles created by language itself­ for example, the roles of questioner or respondent, which we take on by asking or answering a question; and also for getting things done, by means of the in­ teraction between one person and another. . . . (Halliday 1 970: 1 43) Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

I have mentioned Halliday's ideational and interpersonal functions . He postulates a third function: the textual . Where does this come in? According to Halliday it provides the means whereby language makes links with itself so that individual sentences are fused into texts. This oddly anthropomorphic notion seems to be at variance with the descriptions of the other functions, which are based on human agency. Indeed, Halliday sets up his functional grammar in direct opposition to the idea that language structure develops in detachment from its use in servicing human needs. The textual function is surely more consistently considered as the means whereby the language user organizes propositional content so that it is effectively conveyed. The textual function, in other words, serves a communicative purpose: its business is to provide alternative versions of propositions so that they are appropriate to the state of shared knowledge and the dynamism of sharing knowledge at a particular point in an interaction. Here, then, is the simple scheme I want to draw up . Language serves the individual as a means of conceptualizing reality, of establishing some control over his environment. In this role it formalizes knowledge and facilitates thinking. This is the conceptual function. Language also provides the means for conveying basic conceptual propositions, for setting them in correspon­ dence with those in the minds of other people, and for using concepts to get things done in the business of social interaction. The adjustment of propositions so that they fit into the changing situation of shared knowledge is the 'textual' aspect of the matter. The use of such propositions to conduct social business, to perform illocutions of different kinds, is the 'interpersonal' aspect of the matter. Both are features of the communicative function of language. Linguistic theory has, of course, tended to focus on one or other of these two basic functions , with language teaching often following suit. Thus Firth objects to de Saussure because he neglects the communicative in favour of the conceptual (Firth 1 957) and Lyons objects to Firth because he neglects the conceptual in favour of the communicative (Lyons 1 966). The basic difference between Wittgenstein of the Tractatus and Wittgenstein of Philosophical Investigations can be referred to a shift of focus from the conceptual to the communicative. It is interesting, I think, to review transformational generative grammar with reference to this distinction. In Chomsky ( 1 957) for example, we can regard the kernel sentence as the formulation of a basic proposition, a· conceptual unit (which is why it has the 'psychological reality' that Chomsky

H . G . W I D DOWSON

237

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

claimed for it). Transformational rules can be seen as devices for preparing this proposition for communicative use as an element in text (they are, in this sense rules of performance, as Labov has pointed out. See Labov 1 972:226). The illocutionary, as opposed to the propositional, aspect of the com­ municative function came late into the field. It is this aspect which Ross (1 970) and Sadock (1 974) attempt to account for in their proposals for the per­ formative analysis of sentences. So transformational generative grammar, like the systemic grammar of Halliday, can be seen to have a functional basis, even though it may not be explicitly acknowledged. The distinCtion between conceptual and communicative functions I have been making is also quite useful, it seems to me, in a consideration of early work in psycholinguistics which attempted to test the derivational theory of complexity. Experimenters were anxious to get subjects to treat sets of sen­ tences as conceptual units but the natural inclination of the subjects, of course, was to treat them as communicative units. And researchers came to realize that communicative factors such as difference of attention focus (Olson and Filby 1 972) and contexts of plausible denial (Wason 1 965) had an effect on in­ terpretation time (see Greene (1 972) and McRae (1 978) for a useful survey of this research). Greene refers to these factors as 'semantic ' . They are not, they are pragmatic: they have to do not with the conceptual formulation of base propositions but with conditions on their conveyance for some communicative purpose. They have to do, therefore, not with semantic meaning, the con­ ceptual signification of sentences but with pragmatic meaning, the com­ municative value of utterances. Transformational rules, in the view I am proposing, are communicative and not conceptual in function and they are naturally only used when some base proposition needs to be fashioned to fit a particular context. If there is no context and subjects are asked to interpret linguistic expressions in parallel by internal reference, then they are forced to treat them conceptually. The extent to which they are able to actually do this will be a measure of their ability to solve a certain kind of cognitive problem which may have little or no relationship to the processing of natural language. I said that both transformational-generative grammar and systemic grammar have a functional base in that their different components can be associated with the two primary functions I have talked about. But this association is historical and indicates how the internal structure of language reflects the functions it has evolved to serve. Once evolved, this structure provides a resource for both conceptual and communicative functions, so that in use one cannot identify parts of a linguistic unit as having an exclusive connection with one function or another. Thus, Halliday shows how the following expression 'John was throwing the ball' can be analysed as a con­ vergence of choices from systemic options which derive from different func­ tions. Ideationally we have Actor/Process/Affected (John/was throwing/the ball), interpersonally we have Modal/Residual (John was/throwing the ball) and textually we have Theme/Rheme (John/was throwing the ball) and Given/New (John was throwing/the ball). (Halliday 1 968:21 1) . But this does not mean that the ideational structure is only relevant to the formulation of a conceptual proposition. Once a language has developed as an instrument for social interaction, its total resources are available for private purposes as well. But of course languages continue to evolve and their evolution is the con-

238

C O N C E P T U A L A-N D CO M M U N I C A T I V E F U N C T I O N S

Language is always under competing pressure to conform to all four of these charges. Because the pressures are inherently competitory, languages are con­ stantly changing. . . . A gain in compactness or expressiveness of com­ munication is often purchased at the expense of ease of processing or semantic transparency of message. (Siobin 1 975 :4-5) Now I want to suggest that the first two of Slobin's charges are essentially conceptual and the second two essentially communicative. That is to say, the requirements of clarity and processibility are basic to the formulation of propositions and relate to cognitive processing and storage, whereas the other two requirements relate to the conveying of propositions on communicative occasions. These latter charges can, I think, be associated with the Gricean maxims of the co-operative principle (Grice 1 975). As I see it, then, language works in two ways. On the one hand it provides for conceptual activity whereby clear and processible propositions are for­ mulated in the mind. On the other hand it provides the means whereby such propositions can be conveyed in the most effective manner for particular communicative purposes. These two functions, potentially in conflict, have to be reconciled by negotiation on every occasion of social use. This. negotiation is realized in discourse. We come to the second item on the agenda. To engage in discourse is to try to find ways of expressing propositions so that they will be understood and, where relevant, acted upon. As Chomsky says there are all kinds of reasons why one would wish to express propositions : to direct future action, to transfer information, to display knowledge and so on. So the proposition can take on a range of illocutionary values. There is now a good deal of literature about the conditions that have to be met for illocutionary acts to be achieved and on the interpretative procedures which are engaged in the actual achievement. I am not now concerned with that ' interpersonal' aspect of the communicative function, but in the other, the 'textual' aspect, again leaving considerations of their inter-relationship aside for the present. What I want to consider is how, in written discourse, the writer gets his propositional meaning across, and how the reader takes it in. The problem for the writer is that he has to convey his propositions without the benefit of overt interaction which enables conversationalists to negotiate meanings by direct confrontation. This means that he has to anticipate

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

sequence, I think, of recurrent reconciliations o f these private and sociai functions. Halliday, in common with most linguists represents grammatical rules as inhabiting stable systems in peaceful harmony. But the very fact that they have to serve different purposes creates conditions of internal strife. In describing the language system one can represent different options as con­ verging neatly into a unified product, but in using the system there are con­ flicts to be resolved in the process. And these resolutions, these reconciliations of private and social requirements (a feature of human life in general of course and not confined only to the use of language) are the source of language change and of language development in acquisition. Slobin has relevant remarks to make on this matter. He refers to 'four basic ground rules which a communicative system must adhere to' . These are: (a) be clear, (b) be humanly processible in ongoing time, (c) be quick and easy and (d) be expressive. He goes on to say:

H . G. W I D D O W S O N

239

1. 2. 3. 4.

We expand the meaning conveyed by the cues into the nearest equivalent in text terms, according to our best understanding of it . We expand and make explicit the referents of pronouns to other utterances and events in other time frames. We introduce factual material that i s presented before and after this utterance, sometimes from widely separated parts of the interview. We make explicit some of the shared knowledge between par­ ticipants. (Labov and Fanshel l 977:49)

Such procedures are especially necessary for the analyst, concerned as he is with a third person rendering, when the discourse under consideration is very close to its conceptual source. This would be so when the shared knowledge and assumptions of the participants were such as to make conveyancing a relatively straightforward business, when there would be little need to make calculations and provisions for the establishment of common ground. In general, it will follow that the closer the correspondence of conceptual worlds,

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

possible reactions by in effect enacting the roles of both first and second person participants. He is engaged in a covert interaction, shifting roles in Socratic fashion to pose and respond to questions like 'What do you mean by that? ' 'So What?' 'Can you give an example? ' and so on. All the time he must provide for the possible lack of convergence of shared knowledge: of the world, of social conventions, of the language itself. So the propositional content he wants to transmit is elaborated to service what the writer judges to be an effective communicative- interaction. Of course, this interactive en­ deavour does not only serve to facilitate the conveyance of propositions planned, as it were, in advance. The very interactive process involves a con­ tinual shift of perspective which will frequently yield new propositions, new conceptual connections not originally thought of. The writer's covert and non­ reciprocal interaction not only provides for the conveyance of ideas but also helps to generate them. (For further discussion see Widdowson 1978, 1979b .) The writer of course typically only records the first person participant's contribution to this interaction. This partial record of the discourse is written text . Turning now to the reader, his task is to derive a discourse from the text. The extent to which this derivation will reconstitute the writer's discourse will depend on how far he corresponds in actuality to the interlocutor the writer has presupposed. He may not need, or may not want, to follow the course of interaction-so painstakingly plotted by the-writer on his behalf. He can take short cuts according to the state of his knowledge, or according to his purpose in reading. The writer's recording of discourse as text and the reader's derivation of discourse from text can, I think, be related to the process of expansion, as discussed in Labov and Fanshel, on the one hand, and to the process of reduction , as discussed by Van Dijk on the other. And these processes can, in their tum, be referred to the two principal language functions I have been talking about: the conceptual and the communicative. For Labov and Fanshel, expansion is a device for analysing spoken discourse. It consists of the following procedures:

240

C O N C E P T U A L A N D CO M M U N I C A T I V E F U N C T I O N S

the smaller will be the communicative effort called for, and the greater the task of the analyst in textualizing the discourse to make it interpretable.

. . . . the expansion itself is often a help to our understanding and plays a crucial role in the analysis of interaction. But the expansion can also be somewhat deceptive, since there is an interactive component of over-explicitness, which throws many of the actions into a wrong light. . . . Expansions magnify the strains and tensions in the social fabric and will produce distorted interpretation unless we remember that the expansion loses the important dimensions of backgrounding, which subordinates one form of social interaction to an­ other. . . . Psychotherapists at the agency being studied expressed their ap­ preciation for the insights gained, but remarked that this kind of analysis makes the therapeutic session seem like a type of 'warfare' , and makes the relationships with patients seem much more abrasive than they actually are. (Labov and Fanshel l 977:5 1 ) O n the other hand, with respect t o the second example, care-taker and child discourse, Brown seems to imply an equivalence between participant and analyst expansion procedures:

It is not necessary to rely on parents to provide glosses; the researcher can do it himself. Indeed, researchers cannot help doing it. The adult mind receiving a telegraphic utterance in a given context quite automatically expands it into an appropriate sentence. (Brown 1 976: 1 33) It seems clear that expansion as a participant 'member' strategy will take on different functions in different kinds of spoken discourse and will accordingly vary in its approximation to the analyst's use of it as a methodological device. I want to suggest that in written discourse the writer relies very heavily on expansion procedures to provide his conceptual meanings with the necessary communicative conveyance, since he is operating under conditions at the zero end of the co-operation scale. These conditions of non-reciprocal interaction lead to a convergence of participant and analyst perspectives . The writer is a participant in that he is enacting a discourse with an assumed and absent in­ terlocutor but he is at the same time detached from immediate involvement and so he can, in third person analyst fashion, put himself at a remove from the interaction. Expansion requires close attention to surface structure so that it is fashioned in such a way as to ensure the effective conveyance of information. Reduction,

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Procedures for expansion are not, however, the prerogative of the analyst. They must be available to participants too as a communicative resource in conversation. Again, there will be occasions when they have, as such, a particularly important role to play. This will be the case when one participant in an interaction does not, for one reason or another, have the inclination or capacity for communicative elaboration. In these circumstances, the other participant becomes the custodian of the co-operative principle and has to provide expansion to sustain the interaction. One example of such uneven distribution of communicative effort might be the discourse of psychotherapist and patient; another that of care-taker and child. The familiar question now arises as to the relationship between the analyst's procedures for expansion and those of the participant. With respect to the first example, Labov and Fanshel point to the danger of disparity:

H . G. W I D D O W S O N

241

on the other hand, is a device for directing attention to the salient features of information, for stripping the discourse of its communicative integuments to get to the conceptual gist. The demonstration of such a procedure applied to written discourse appears in Van Dijk (1977) where it is represented as the recovery of semantic 'macro-structures' by the techniques of deletion, combination and generalization. The assumption seems to be that the analyst's reduction matches that of the participant. If the reader is regarded as taking a non-reciprocal role in interaction corresponding to that of the writer, such an assumption might seem reasonable. There is an important reservation to be made, however, which I will come to presently. Meanwhile, it is worth noting that reduction, like expansion, is a common resource in spoken interaction. Indeed both activities are described by Garfinkel and Sacks undeF the general ' heading of ' formulations' :

This quotation appears in Heritage and Watson (1 979). They are concerned with reductive formulations and they, too, specify three operations : in their case, preservation, deletion, transformation. They give the following example of a solicitor being interviewed on a radio programme : S:

I: S: 1:

S:

The inescapable facts are these, er in nineteen thirty two when he was er aged twenty three mister Harvey was er committed to Rampton hospital under something called the mental deficiency act nineteen thirteen which of course is a state that was swept away years ago and er he was com­ mitted as far as I can er find out on an order by a single magistrate er sitting I think in private. How long did he spend in Rampton? Well he was in er Rampton and Mosside hospital er alternatively er until nineteen sixty one That's the best part of thirty years (formulation) That's right.

Heritage and Watson comment as follows: In this example, the interviewer's formulating utterance: 'That's the best part of thirty years' exhibits these three properties. Specifically, it preserves the length of time Mr. Harvey was in hospital whilst simultaneously deleting such in­

formation as: the names of the hospitals involved, the Act of Parliament under which Mr. Harvey was committed, what subsequently happened to the Act, the circumstances of his commital and so on. At the same time, the interviewer's utterance transforms some of the information furnished to him (i.e., that Mr. Harvey entered hospital in 1 932 and left in 1961) and re-presents this in­ formation as the outcome of an arithmetical operation: 'That's the best part of thirty years. ' In furnishing the formulation, the interviewer re-describes or re­ references parts of the information already delivered to him, thus preserving them 'in other words. ' (Heritage and Watson 1 979: 1 30)

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

A member may treat some part of the conversation as an occasion to describe that conversation, to explain it, or characterize it, or explicate, or translate, or summarize, or furnish the gist of it, or take note of its accordance with the rules, or remark on its departure from rules. That is to say, a member may use some part of the conversation as an occasion to formulate the conversation . . . . We shall speak of conversationalists' practices of saying-in-so-many-words-what­ we-are-doing as formulating. (Garfinkel and Sacks 1 970:350,3 5 1 )

242

C O N C E P T U A L A N D CO M M U N I CA T I V E FUNCTIONS

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Although the properties o f formulation, as presented here, and those of reduction as presented in Van Dij k can be seen to have some correspondence (deletion is common to both), it is not clear how the interviewer's formulating utterance can be understood as the semantic macro-structure of this con­ versation. And here we return to the question of the equivalence of analyst and participant reductions. The critical point about the interviewer's formulation is that it is made to further some conversational purpose. He selects the in­ formation he wishes to present as of particular relevance; he does not abstract the whole. And the same is true of reader reductions of written discourse. The communicative conditions provide the reader with the opportunity to recover the conceptual macro-structures of the writer's intention but he may not wish to take advantage of it . My argument then (to reduce or formulate it to its essential conceptual content) is this . Written discourse operates by means of the same basic in­ teractive procedures as characterize spoken conversation but the absence of reciprocity calls for a different mode of exploitation. The writer is involved in a process of discourse enactment whereby conceptual content is expanded for conveyance and in the absence of an active interlocutor to negotiate the course of the interaction, his expansions will tend to match those recoverable by analysis . This discourse process is partially recorded as a textual product. The reader reconverts this product into a process and so derives a discourse from text . This discourse, however, is reduced and this reduction yields not the underlying macro-structure of the writer's original formulation (so far as this is recoverable by analysis) but whatever conceptual content corresponds with the reader's state of knowledge and his purpose in reading. One might say, in general terms, that in writing expansion provides the means whereby the conceptual function can come to terms with the communicative and in reading reduction provides the means whereby the communicative function can come to terms with the conceptual. Let me, finally, return to the general theme with which I began . The teaching of languages should, I would agree (and have often enough asserted) be concerned with communication. But we must take care not to define this too narrowly. Our aim must be to develop in learners a capacity for using language for both thinking and acting so that they can exploit its meaning potential in discourse. This is not a simple matter of learning how to express a selection of notions or perform a selection of illocutionary acts. It is, more fundamentally, a matter of learning strategies (like expansion and reduction as I have presented them here) for reconciling conceptual and communicative functions in the discourse process. I mentioned at the beginning the dangers of misrepresentation in simplifying ideas for practical application, and some might say that the present paper is itself a good illustration of this. Perhaps it might appear so. But my intention has been to try to correct distortion and discourage excess by presenting a more balanced model of language communication than is commonly promoted at present. Whether I have succeeded in my intentions is another matter. At all events (if I may end on a note of exhortation) the responsibility of applied linguists in this matter is clear: to mediate between the -theory of language communication and the practice of teaching it without misrepresenting the former or misleading the latter. A difficult task but if we

H . G . W I D DOWSON

243

do not achieve it we shall be discredited. And then another line from Sir Thomas Wyatt will become apposite: it is the opening line of the poem from which my first quotation was taken . It runs: 'They fle from me that sometyme did me seke' .

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

ANALYSE DE DI S COU R S E T P R A G M ATIQU E D E LA PAROLE DANS Q UELQ UES US AGES D'U NE DIDACTIQUE DES LANGUES D A N I E L CO STE

CREDIF, Ecole Normale Superieure de Samt-Cioud

A B S T R A CT

Mon propos sera bref et limite: montrer comment la didactique des langues, du moins en France, parait se referer ou ne pas se referer a la pragmatique de la parole et a !'analyse du discours; essayer d 'interpreter ensuite cette lecture proposee d ' une evolution en cours. En gros, le constat sera que, dans ses developpements recents, la didactique s'est revelee plus sensible a ce que pouvait lui apporter l'etude des actes de parole qu'au parti a tirer de !'analyse du discours. Le questionnement tendra a trouver dans les pesanteurs et les aleas de Ia didactique elle-meme, ainsi que dans le champ ou elle prend place, des raisons eventuelles de cette distinction et de cette distribution. Quelques remarques s'imposent pour eviter, je l'espere, certains malen­ tendus . Tout d'abord, il va de soi que le bilan initial ainsi que les essais d'in­ terpretation restent a dessein sommaires, presque caricaturaux et, en tout cas, transitoires . Il n'entre pas dans mon intention d'engager un examen historico­ epistemologique attentivement nuance; la visee est plutot de souligner des tendances majeures pour mieux pointer, peut-etre, des manques, des impasses ou des derives privilegiees dans le travail de la didactique. Il faut en outre insister sur la portee regionale des considerations qui suivent: si elles ont quelque pertinence, c'est d'abord pour le contexte de la didactique du fran�ais langue etrangere, vu de France, qu'il y a lieu de la rechercher. Les distinctions etablies-et, au premier chef, celle qui partage, dans le titre meme, analyse de discours et pragmatique de Ia parole m'ont semble pouvoir rendre compte de certains phenomenes notables 'localement' . I l est plus que probable qu'elles ne sont pas exactement les memes ailleurs et -

Appl1ed Linguistics, Vol. I,

No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

This paper discusses the way in which speech act theory and discourse analysis have been applied to practical pedagogy, with particular reference to the teaching of French as a foreign language. The point is made that there has been the tendency to relate these two areas of enquiry to different aspects of language teaching, with speech act theory being associated with the teaching of the production of separate spoken utterances in general contexts and discourse analysis with the teaching of the understanding of larger stretches of written language in specialized contexts. The author outlines an explanation for this state of affairs and points to the importance of understanding and resolving the ambiguities that it creates in language teaching methodology.

DAN I E L COSTE

245

que d 'autres clivages operatoires peuvent, au meme moment, se situer sur des axes differents pour d'autres langues et dans d'autres pays. Cette precaution prise, il va de soi que je ne m'interdirai pas toute allusion a des experiences etrangeres mais il doit etre clair que ces allusions n'impliqueront nullement que le constat et le questionnement aient valeur generique.

1.

U N E P O L A R I S A T I O N AN A L Y S E D E D I S C O U R S / P R A G M AT I Q U E D E

L A PAROLE E N D I DACTIQUE

1 . 1 Repartition sommaire

pragmatique de Ia parole oral expression evenement de parole perspective onomasiologique paradigmatique selection, adequation communication 'usuelle'

analyse du discours ecrit comprehension texte perspective semasiologique syntagmatique cohesion, coherence domaines specialises

Dire que Ia didactique a plutot tire l'acte de parole du cote de l'oral et le discours du cote de l'ecrit, au moins pour le domaine fran�ais, ne demanderait pas un tres gros effort d'illustration et je m 'abstiendrai d'aligner les exemples de cette derive. II suffit peut etre de rappeler que, quand Ia notion d'acte de parole s 'est-tres rapidement-repandue dans les milieux interesses par l'enseignement des langues, elle a surtout ete introduite a l'aide de cas ('demander O'heure) ' , 'demander (son chemin)' , 's'identifier, se presenter') dont les realisations etaient d'abord proposees et diversifiees a l'oral. C'est, me semble-t-il, ce qui se passe non seulement dans des textes de Wilkins (1 976) ou de Roulet (1976, 1977) destines a un large public, mais aussi pour nombre des publications issues du Conseil de l'Europe et relatives notamment a Ia construction de niveaux-seuils (Van Ek, 1 975; Coste et alii, 1 976). L'echange face a face entre deux interlocuteurs y est largement privilegie. De plus, l'accent est generalement mis sur Ia production de l'acte (de l'intention de communication vers les formes linguistiques), dans une perspective onomasiologique qui ouvre un eventail de possibilites, etant entendu qu'il appartient au locuteur de choisir, dans le paradigme a sa disposition, les formulations qu'il estime les plus adequates a son propos et a Ia situation. L'accent, de ce fait, se trouve porter sur l'expression plus que sur Ia com­ prehension et Ia visee enonciative voit ici l'acte du cote du producteur plus que du recepteur du message. Enfin, le gros des propositions pedagogiques-et

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

S'il en etait besoin, le tableau ci-dessous suffirait a mettre en evidence le choix simplificateur et le point de vue partiel qu'on a retenus. II ne s'agit pas d'in­ terroger le rapport theorique postulable ou a construire entre discours et acte de parole ni, a ce point, de rechercher des explications a ces apparentes polarisations; mais bien de marquer des regroupements qui ont pu se dessiner dans l'usage en didactique de notions, d'instruments, de problematiques ayant leur origine ailleurs.

246

D I S C O U R S ET P R A G M A T I Q U E DE LA P A R O L E

singulierement celles qui ont trait a !'elaboration d e documents d'en­ seignement-paraissent associer Ia pragmatique de Ia parole a Ia com­ munication quotidienne usuelle plus qu'a Ia frequentation de 'textes

1 .2 Commentaires Ce jeu de mises en regard servait a baliser a tres grands traits le champ. 11 appelle des commentaires multiples mais je m'en tiendrai a trois types de complements pour: a cette (a) avancer quelques premieres j ustifications possibles bipolarisation; (b) souligner qu'elle est d'autant plus symptomatique que d'autres facteurs auraient pu, notamment en France, contribuer a la reduire ou a l'eviter; (c) corriger le parallelisme egalitaire apparent dans le tableau en rappelant que !'analyse de discours est bien loin d ' avoir eu autant d'echo en didactique que, jusqu'a ce jour, la pragmatique de Ia parole.

1 .2 . 1 Explications possibles Les premieres justifications a proposer soot doubles et convergentes, relatives, d ' une part , a l'origine des notions et, d ' autre part, a la nature de Ia plupart des cursus d 'enseignement. On notera en effet : que la notion d'acte de parole et les pratiques d'analyse du discours n'ont pas la meme origine epistemologique: celles-ci soot at­ tribuables d 'abord a des linguistes debordant le cadre de la phrase (Harris , par ex.), celle Ia a des philosophes du langage s'interrogeant sur les jugements de verite et les rapports entre semantique et pragmatique (Austin , 1 962); que les linguistes ont d'abord travaille sur des corpus de textes ecrits pour reperer les fonctionnements discursifs dans des sequences observees alors que les philosophes du langage ont plutot fait porter leur reflexion sur des phrases efficaces dans la mesure ou elles soot hie et nunc prononcees par un sujet (' Je

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

specialises' . Sur tous ces points, pour simplistes et peremptoires que soient les affirmations qui precedent, il est probable que, globalement, le constat vaut pour les annees recentes, avec bien entendu quelques contre-exemples dont certains seront evoques plus loin. Meme panorama rapide en ce qui concerne l'analyse de discours: sans recourir a des etudes multiples et a des demonstrations detaillees, il n'est sans doute ·pas faux de dire que c'est avant tout a propos de la lecture de 'textes specialises' qu'il a ete fait appel, en didactique franc;aise, a la notion d'analyse de discours (Darot, 1 975, Beacco et Darot , 1 978, Lehmann et alii, 1 979); on s 'interesse alors surtout , pour l'acces a une comprehension ecrite, aux facteurs de cohesion syntagmatique et aux marques qui semasiologiquement per­ mettront au lecteur de reconstruire du sens. C'est une telle tendance a une polarisation double que le tableau propose plus haut entendait, dans son parallelisme, manifester brutalement . 11 ressort evidemment de cette opposition que !'articulation entre actes de parole et fonctionnement discursif n'est pas clairement envisagee ni meme vraiment recherchee par Ia didactique dans son ensemble (la encore, on evoquera plus loin quelques 'exceptions'). 11 serait facile de montrer que la plupart des ouvrages a orientation pedagogique-a tout le moins, j ' y insiste, ceux portant sur l'enseignement du franc;ais-ne temoignent pas d'un net souci d'etablir cette mise en relation.

DAN I E L COSTE

247

declare la seance ouverte' 'Je te baptise au nom du Pere . . . '); et, meme en s'etendant largement ensuite au dela de l'examen initial d ' une categorie de

Enseignement a des debutantf priorite a !'oral maitrise active d'une cawcite d 'expression . travail du dialogue en face a face . production de phrases

. situations de communication 'usuelle'

Enseignement a un 'niveau avance' . place croissante de I' ecrit . importance de Ia comprehension . contact avec les textes authentiques Gournaux, oeuvres litteraires) . interpretation, resume ou commentaire de sequences longues ou bien . orientation vers l'acces a des discours a orientation specialisee (professionnelle ou autre)

Les recoupements sont clairs entre ce tableau et le precedent. En bref, les contextes et les propos d'origine de l'analyse de discours et de l'etude pragmatique de Ia parole, aussi bien que Ia distribution assez classique des cursus d'enseignement-apprentissage d'une langue vivante, peuvent conjointement rendre compte du nichage ecologique privilegie de chacun de ces deux apports exterieurs, quand la didactique entend en faire usage. 1 .2.2 Lesjacteurs de mise en relation Ce partage net entre pragmatique de la parole et analyse du discours peut passer pour d'autant plus significatif que d'autres influences, representees notamment en France, auraient pu le freiner ou peser a l'inverse. J'en note la aussi de deux ordres, theorique et didactique. Grace aux travaux de Benveniste ( 1 966) et a l'influence non negligeable de Jakobson en France ( 1 963), il est net que la notion de discours a ete, en linguistique, chargee enonciativernent beaucoup plus qu'elle ne l'est dans les premiers modeles harrisiens; ainsi, chez Benveniste, Ia distinction entre discours et recit, le concept d'instance de discours, le travail sur les pronoms et les temps du verbe dans leur rapport possible avec le hie et nunc de l'enon­ ciation tendent a situer, dans une certaine mesure, Ia notion de discours du cote de l'oral et du sujet producteur ou, a tout le moins, a etablir des

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

performatifs, leur analyse s'est souvent faite, paradigmatiquement, en comparant non des enchainements d'enonces mais des phrases substituables ayant ou non, dans tel jeu de conditions, telle valeur efficace. A cet egard il y a evidemment une sorte de predilection de l'analyse de discours a se rattacher plutot a la partie droite du tableau dresse plus haut (ecrit, lecture syntagmatique, . . . ), alors que la notion d'acte trouve plus 'spontanement' place du cote gauche de ce meme tableau (oral, selection paradigmatique, etc.). Mais on notera aussi que, sur le plan pedagogique, une telle distribution peut etre considerablement renforcee par le fait que, dans son ensemble, Ia didactique reste marquee par un clivage des cursus qu'on pourrait resumer ainsi :

248

D I S C O U R S ET P R A G M A T I Q U E DE LA P A R O L E

1 .2.3 Le desequilibre defait Malgre ces possibilites de mise en relation plus etroite, le reequilibrage qui permettrait de ne pas simplement separer, en didactique, analyse de discours et pragmatique de Ia parole, ne s'est pas vraiment opere jusqu'a present, et Ia forte tendance a Ia bipolarisation semble l'avoir longtemps emporte. 1 Mais cela tient peut etre d'abord au fait (et a cet egard le tableau initial presenterait une symetrie trompeuse) que !'analyse de discours elle meme n'est pas en position aussi forte que Ia pragmatique langagiere dans Ia pedagogie d'aujourd'hui. Globalement, et surtout en France, !'analyse de discours n'a guere embraye efficacement, jusqu'a maintenant, sur Ia didactique, au, plus exactement, cette derniere n'a guere mobilise !'analyse de discours (sinon, recemment pour l'ecrit, avec des propositions comme celles, deja mentionnees, de Beacco-Darot, Darot, Portine, Lehmann et ali1). A bien des egards, il serait meme interessant de montrer que Ia notion d'acte de parole, et, (par contagion?), !'approche dite fonctionnelle-notionelle, envahissent l'ecrit meme et en viennent a concurrencer, sur son propre terrain de predilection, !'analyse de discours (ce qui apparait pour partie dans les travaux qu' on vient de rappeler et aussi dans Moirand 1 977, Colombier et alii 1 977, Coste et Martins-Baltar (dir.) 1977). Simple juxtaposition ou etape vers une necessaire integration?

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

passerelles et a mettre e n valeur des coherences. Complementairement, si l'on peut dire, les analyses de Ducrot ( 1 972- 1974), sur le rapport semantique/pragmatique, son attention aux presupposes et aux sous-entendus, entretiennent avec Ia tradition linguistique des relations plus etroites que ce n'est le cas pour les analyses d' Austin au d'autres. Et Ia encore, il y aurait matiere a nuancer fortement le tableau d'oppositions entre acte et discours . Du cote didactique, certaines autres mises en liaison auraient pu elles aussi contribuer a ne pas trap distancier au plan pedagogique actes de parole et analyse de discours. C'est le cas des travaux sur le discours rapporte faits au CREDIF (Gauvenet, 1 976) au le passage de !'oral a l'ecrit, etudie dans les comptes rendus journalistiques de conferences de presse, souligne !'im­ portance des verbes introducteurs de discours-objets ('il a declare que . . .' 'X a pretendu que . . . insiste sur le fait que . . . approuve Ia proposition faite par . . . ' etc .). Se trouve ainsi mise en valeur, a l'ecrit, Ia dimension discursive d'actes produits a !'oral et rapportes; mais se marque aussi tres nettement, par les prises de position et les interpretations que revelent les formulations du journaliste rapporteur, Ia mise en oeuvre a l'ecrit d'actes a propos d'autres actes dont on rend compte. Dans Ia continuite de ce premier exemple, on peut noter que Ia section 'actes de parole' dans Un niveau-seuil (Martins-Baltar, in Coste et alii, 1 976), si elle met !'accent, dans les formulations proposees, sur les realisations orales, presentees et illustrees comme lieu exemplaire de Ia parole individuelle, comporte aussi des references multiples a l'ecrit et propose une sous-section 'operations discursives' . au le fonctionnement meme du discours (forme et substance) fait !'objet d'un essai de typologie pragmatique qui ne se limite aucunement au niveau phrastique et s'applique en grande partie in­ differemment a !'oral et a l'ecrit (voir aussi Martins-Baltar et alii, 1 979).

DAN I E L COSTE

2.

249

LE REFUS RELATIF DE L ' ANALYSE D U DISCOU R S

La didactique persiste a avoir quelque peine a sortir du phrastique. Peut etre

a.

b.

le poids qu'a pu avoir une tradition de traitement plus lexicographique que discursif des corpus (ex.: fran�ais fondamenta/ et enquetes sur les vocabulaires scientifiques et techniques); Ia redecouverte indirecte d u discursif par l e biais du lexical, notamment dans le cas de Ia preparation du Dictionnaire contextuel de fran�ais pour Ia geologie (Descamps et alii, 1978), qui propose un travail considerable d'etiquetage semantique et discursif du fonctionnement en contexte de mots releves dans un corpus de textes de geologie.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

parce que Ia dichotomie langue-parole, les conceptions de l'apprentissage fondees sur l'exemple, continuent a peser tres lourd malgre les propos liberateurs qui , depuis peu, voudraient centrer !'action de formation sur le ' discours' de l'apprenant. Tout se passe comme si on avait toujours besoin d ' une 'matiere a enseigner' , coupee en tranches denombrables, itemisee et cataloguable. Or, de ce point de vue, le discursif est mal place: il ne se laisse pas aussi facilement mettre en listes que le lexique ou en arbres que les structures. A l'inverse, les actes de parole ou les notions, tels que normalises par certaines descriptions, se substituent sans peine aux contenus anciens. 11 n'est pas inutile de constater qu'on enregistre aujourd'hui en France un decalage symptomatique entre l'air du temps et les references auxquelles fait appel Ia didactique. Alors que, dans les annees 60, Ia notion de structure regnait en didactique comme partout, on releve aujourd'hui que Ia notion (ou a tout le moins le terme) de discours, omnipresent en histoire, epistemologie, philosophie . . . et linguistique, reste a l'etat de mention beaucoup plus discrete ou, en tout cas, de contagion plus tardive en didactique des langues . Cela alors meme que l'analyse de discours etait, d'une certaine maniere, 'disponible' depuis quefque temps (depuis plus longtemps, par le canal de Ia linguistique, que la pragmatique de Ia parole). S'il faut considerer comme assez revelatrices des revues comme /e Fran�ais dans le Monde ou les Etudes de linguistique appliquee, il apparaitra que, ces demieres annees, les sommaires n'y ont guere temoigne de Ia pertinence que pourrait avoir !'analyse de discours pour la didactique des langues; quant a !'analyse textuelle, elle y reste encore quasiment lettre morte (voir, en revanche, Languejran�aise n° 38, dirige par J. Peytard et M. Charolles). Dans le meme sens, il serait facile de relever que le Dictionnaire de didactique des Langues (Galisson et Coste, dir.), publie en 1 976, ne fait pas grand place au discours et a son analyse, alors meme que, dans des pays voisins, des manuels de linguistique appliquee avaient deja consacre d'importantes sections a ces problemes, (Allen et Corder, 1 973-75). 11 y a simplement maintien d'une problematique de Ia typologie des discours ('polemique' , 'didactique' , 'scientifique'), d'une opposition globale entre discours oral et discours ecrit, mais sans que ces categorisations renvoient effectivement-et pour cause-a des techniques ou a des resultats de !'analyse. L'explication des ces differences intra ou intemationales demanderait a coup siir des etudes et des developpements qui restent a faire. Je me con­ tenterai de signaler:

250

D I S C O U R S ET P R A G M A T I Q U E D E LA P A R O L E

c.

!'approche aussi des phenomenes d e discours par l e biais mor­ phosyntaxique (etude des phenomenes de nominalisation, des ar­ ticulations transphrastiques lexicalisees, des relations diaphoriques) .

plutot manifeste jusqu'a present, malgre quelques tentatives de rap­ prochement amiable, une disharmonie entre les termes parole et discours (il n'est d'ailleurs pas impossible que !'existence de deux termes, ayant chacun son histoire en linguistique fran�aise et fonctionnant l'un et !'autre dans certains jeux d' opposition avec langue, ait contribue a compliquer les rapports entre pragmatique de la parole et analyse de discours) . Les travaux du CRAPEL (Melanges) et les autres mouvements integratifs deja mentionnes n'ont pas change grand chose pour !'instant a une dynamique qui, massivement, est plutot allee dans un autre sens. Les pesanteurs de la didactique n'expliquent pas tout a cet egard puisque se manifeste par ailleurs sa capacite a 'recuperer' rapidement certains aspects de Ia pragmatique, tels ceux mobilises par les travaux du projet ' Langues vivantes' du Conseil de !'Europe. ' Recuperation' dans Ia mesure ou on peut craindre que des outils comme Un niveau-seui/, qui font une large place a une typologie des actes et a des listes de notions soient Ius, malgre les precautions prises , comme des tables de correspondance inventoriant les contenus souhaitables d'un enseignement fonctionnel-notionnel et servent ainsi, bien malgre eux (?), a favoriser une atornisation de la 'matiere a enseigner' et a maintenir a l' ecart les aspects discursifs de toute mise en oeuvre pragmatique du langage. Ce refus relatif du discursif, cette reduction qu'on en fait ou son rejet a des phases tardives de l'apprentissage, apparaissent au bout du compte, jusqu'a aujourd'hui, comme une tendance forte et durable qui , en matiere d'en­ seignement des langues, ne s'exerce sans doute pas seulement sur la scene fran�aise. Aux interpretations diverses deja proposees dans les pages qui precedent, d'autres a COUp SUr pourraient etre ajoutees, d'ordre sociologique

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Toutes entreprises dont les resultats ont ete tres riches et didactiquement 'rentables' mais dont on voit bien que, si elles operent un traitement non negligeable des problemes, elles insistent avant tout sur cette sorte de lexicalisation du discursif qui risque de laisser echapper la dimension textuelle, ne dit presque rien de !'organisation des contenus et bien peu du rapport avec la visee pragmatique, tout simplement parce que son propos est autre. Ceci dit, il est permis de se demander si, ailleurs qu'en France aussi , et meme dans les cas ou Ia linguistique appliquee s'est interessee plus tot et plus ac­ tivement a !'analyse de discours, la prise en compte de cette problematique par l'enseignement des langues n'a pas souvent conduit surtout a attirer !'attention sur les marques morphosyntaxiques et lexicales des articulations trans­ phrastiques (tous Ies indices proprement linguistiques de la cohesion) plutot qu'a une recherche des conditions et des realisations textuelles de la coherence. Si l'on admet que cette demiere approche de ! 'organisation du discours (qui bien evidemment ne rend pas l'autre moins indispensable) n'est guere possible que si !'entree linguistico-linguistique n'est pas Ia seule retenue, on retrouve­ sans surprise-le besoin d'une mise en relation du pragmatique avec le discursif, permettant de reconcilier, si l'on peut dire, parole et discours, cette patrie desunie en deuil de la langue. Or, comme ce qui precede a essaye de le montrer, la didactique du fran�ais a

D A N I E L COSTE

25 1

et institutionnel. J e m e contenterai d 'evoquer simplement ici, indirectement, quelques unes des 'justifications de rejet' qu'on peut entendre ici ou Ia: . Le discours ne peut etre produit et connu que si les eU:ments linguistiques qu'il comporte ont d'abord ete appris isolement (hors discours en quelque sorte) . . Les savoirs et savoir-faire tenant au discursif n'ont pas a etre appris en langue etrangere; ils soot transferables a partir de Ia langue matemelle. . S'il faut enseigner aussi le discursif et les contenus et conventions socio­ culturels qu'il vehicule et auxquels il repond, il n'y aura plus de temps pour

3.

L ' A V E N I R A U X S T R AT E G I E S?

Les decalages entre, d'une part, certaines preoccupations actuelles de la linguistique appliquee et, d'autre part, ! 'evolution de la didactique dans la lecture qu'elle fait des innovations et Ia portee qu'elle leur confere pour l'enseignement, paraissent done aujourd'hui tout a la fois considerables et legers. Considerables si l'on observe certaines inversions de signes: alors que la didactique (dans le secteur regional considere) associe plutot les actes de parole a l'oral et quelques aspects de l'analyse du discours au travail de l'ecrit, la linguistique appliquee, alentour, centre plutot l'analyse de discours sur !'in­ teraction orale et tend a introduire l'analyse textuelle pour un traitement privilegie de l'ecrit. Considerables aussi quand on note les reductions que subissent Ia pragmatique et le discours des qu'ils 'passent' dans des pratiques d'enseignement institutionnalisees. Mais legers en meme temps si l'on s'en tient a la vitesse de certains de ces passages et a !'impression assez generale d'une acceleration dans le processus par lequel la didactique absorbe et triture certains emprunts a d 'autres champs. Legers si on estime que le constat etabli et discute ici est revelateur d'une phase de transition et d'un bref temps de latence avant que ne s'etende ineluctablement un profond mouvement de renovation didactique qui puiserait, entre autres , aux meilleures sources de Ia linguistique appliquee! II n'est pas besoin de souligner que personne ne se fait beaucoup d'illusions sur les lendemains de l'enseignement des langues s'ils prenaient de nouveau simplement les voies d'un suivisme de la linguistique appliquee. Mais deux remarques d 'intention optimiste peuvent etre formulees en guise de con­ clusion. La premiere est que bien des experiences susceptibles de combkr les ecarts et les decalages ont ete ponctuellement engagees de par le monde. Ce qu'il est convenu d 'appeler 'approche communicative' rassemble aujourd'hui sur des memes terrains des individus d'horizons differents dont le travail evitera peut etre en partie les distorsions et les deviations que d'autres tentatives de

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

enseigner Ia langue . . De toute facon, a l'examen, ce n'est pas ca qui est evalue. Pour contradictoires qu'ils paraissent et pour implicites qti'ils restent souvent, ces arguments (dont aucun, il faut le souligner, n'est entierement ridicule, et qui pointent tous des problemes reels) se combinent plus qu'ils ne s'annulent et ne laissent flltrer des apports possibles de l'analyse du discours que ce qui ne tombe pas sous l'une ou l'autre de leurs massues. Autant dire peu de chose.

252

D I S C O U R S ET P R A G M AT I Q U E DE LA P A R O L E

NOTES 1 I I convient d e souligner une fois encore que le tableau est schematique et regional. Disons, pour faire vite, qu'il n'existe pas en fran� d'equivalent de Teaching Language as Communication . (Widdowson 1 978), c'est a dire d'une articulation pensee pour Ia didactique.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

renovation ont p u connaitre. Sans faire preuve d e trop d'angelisme, i l faut a tout le moins noter que la vieille dichotomie theorie-pratique a ete ici ou la repensee. La seconde est que, par ajustements successifs , s'agissant d'apprentissage des langues, tout se passe comme si la pratique de la theorie et la theorie de la pratique faisaient apparaitre certaines convergences, certaines focalisations des problematiques . Ce n'est sans doute pas tout a fait un hasard si la notion de strategie est de celles auxquelles beaucoup aujourd'hui font appel, a un moment par ailleurs oil !'interet pour les rhetoriques, !'etude de !'argumen­ tation (Vignaux 1 976), paraissent ouvrir de nouvelles pistes a !'analyse pragmatique de la production-interpretation de sens dans des discours par un sujet social; a un moment aussi oil les dimensions cognitives et sociales de l'apprentissage, ainsi que les modalites de !'action d'enseignement sont souvent pensees et presentees en termes d'operations et de strategies. La en­ core, tout en se gardant du fetichisme des mots-etendards et de l'oecumenisme rassembleur des faux-concepts, il est permis de se demander si on ne com­ mence pas enfin a toumer autour de quelques vrais pots.

S OME I MPLI C ATI ONS OF DI S C OURSE ANALYS I S FOR ESP METH ODOLOG Y J O H N M cH. S INCLAIR

University ofBirmingham

Applied Linguistics, Vol. I, No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Discourse Analysis is emerging as a body of theoretical concepts and descrip­ tive techniques that place centrally the interactive aspect of linguistic events. It is complementary to the tradition of formal descriptive linguistics and anchors its observations to accepted linguistic categories through realisation state­ ments (at least in respect of discourse verbalisations, which by definition are the primary objects of study). It also offers an explanatory dimension for those patterns which cannot be fully accounted for by non-interactive linguistics. An analogy is proposed for the relationship between the level of discourse and that of language form (syntax and lexis). It can reasonably be compared to that between form and phonology, in that each level is separately organised, and provides a limiting range of distinctions which are available, unordered, to realise categories at the next level. There are substantial traces of systematic correspondences between a category established at one level and its deployment in the next, such as might be suggested by reasonable principles of economy, but priority in description is given to the autonomy of patterns within a level, because of the particular range of linguistic phenomena which a level is established to describe. The realisation relation in English between syntactic and lexical affixes and their phonological realisations is one instance of this cross-level correspon­ dence. A similar one between discourse and syntax would be categories like question , command, statement and the clause types interrogative, imperative and declarative. The relationship of realisation as used here works in both directions between levels. Neither has priority, but with respect to one level, another carries the inventory of possible realisations, and its own internal organisation is not considered relevant. The phenomena which the level of discourse describes are those that arise when more than one participant is involved in creating linguistic structures, and where the activity is supposed to be purposeful. For convenience, these phenomena will be called multiple-source, and contrasted with single-source phenomena, such as are described by conventional syntax, phonology, etc. In the absence until recently of a level in which multiple-source phenomena could be comfortably described, attempts have been made to incorporate some of them within single-source linguistics, but not to the satisfaction of those in­ volved. For three quick examples, consider the domains of pragmatics, in­ tonation and ellipsis. The linguistic facts are readily available and hardly in dispute, but the attempts at explanation suggest that the descriptive apparatus lacks the necessary organisation to accomplish the task. Pragmatics concerns the achievement of interactive events, and thus cannot be observed in single-source models. Much of intonation is to do with the

254

I M P L I CA T I O N S O F D I S C O U R S E A N A L Y S I S

assumed status o f propositions in another participant, and is therefore a

a. b. c.

a set of propositions evidence of the writer's position evidence of the characteristics of the target reader .

The reader, in addition to understanding the propositions- at (a), must also

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

meaningless counterpoint on sentences. Ellipsis and kindred phenomena are only accounted for by the presumption of a separate language-understanding entity for whom full formal articulation is unnecessary. The assumption of purposefulness is a major characteristic of discourse analysis. The purpose of activity as seen from single-source linguistics is that of construction of meanings through arrangements of items following formal rules. The purpose of activity as seen from multiple-source linguistics is the achievement of outcomes. At least two language-using individuals contract to exploit their ability to construct meanings in such a way as to move from one state of awareness to another. Whatever may be the planning of individuals, the outcome is determined by the interaction. Outcomes are not a class of events that can be enumerated, and it becomes a central tenet of discourse theory that a finite sub-class of outcomes can be discerned which are recognised within the discourse. For the present argument, it is important to note this assertion, despite the lack so far of any serious attempts to enumerate the sub-class . Outcomes are not the same as 'What transpires' or 'What happens next ' . They are not causal chains and they are not sequences of events. They are constructs of discourse. Again using an analogy, if acts in discourse are similar to clauses in syntax, and moves in discourse are similar to sentences in syntax, then outcomes are combinations of sentences into paragraphs, of moves into exchanges, sequences and transactions. The written language is amenable to description in interactive terms, less obviously than the spoken, but no less valuably. Many teachers of English nowadays are under pressure to concentrate on reading comprehension, and so may find more direct interest in this area than in the study of oral interaction. Discourse analysis should not only be able to illuminate the study of written texts, but also provide explicit links between written and oral language, and between receptive and productive roles. The latter link is already implicit in the requirement of more than one participant, and the oral-written link will be developed later in this paper. First, it must be established that written language is sensitive to interaction, despite the appearance of much written communication. The participants have very distinct roles and are normally far apart in space, time or both. Dialogue is not to be found, normally, apart from the simulations in literature, some correspondence, and fringe areas like form-filling and crosswords . Nevertheless, a writer constructs for himself a target reader, often explicitly, and refers to this construct for decisions about the selection, ordering and presentation of material. Some texts make this more obvious than others, and in some the target reader is more clearly specified and identified than in others. In many he is directly addressed, at least in prefatory material. The writer also constantly indicates his own position vis-a-vis his material, and the reader is presented with a package which includes, roughly

JOHN McH. SI NCLAIR

255

ORAL

TEXT

WRITTEN �TEXT WTR (TR) w

---t AR

writer

=

AR

=

target reader actual reader

Much the same model holds for pre-recorded oral composition which is listened to in playback. A graphic experience of the need to create a target listener is available to anyone who encounters a telephone answering service. The priorities of multiple-source construction oblige us to elevate a distinction which, although available in single-source linguistics, is not nor­ mally considered of great importance. This is a distinction in directionality, and the terms are prospective and retrospective. It is a fundamental distinction in discourse simply because the forward, or prospective control of discourse construction is by negotiation of participants, whereas each participant in a turn has an opportunity to develop his personal messages out of what has gone before, through the creation of retrospective patterns. The frameworks of categories that will be introduced to account for these two types of organisation are quite different, and text instances are not required to map onto each other. That is to say, a text analysis may well mark the occurrence of retrospective items in the middle of prospective ones, or straddling the boundaries of them, and vice versa. They are two different kinds of phenomena. Prospective structures are concerned with control over what happens next. They are the attempts of an individual to pre-classify the next utterance of another, to negotiate speaking rights or to indicate desire or willingness to relinquish them. They are understood with reference to a finite network of possibilities shared by participants, organised in a hierarchy of units. They are thus similar in pattern to syntactic structures, and a model adapted from

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

recreate the position of the writer as presented, and must adjust his own in­ terpretive skills in respect of his perceptions of (c), the target reader. Those models of reading which recognise only or primarily (a) above seem not only to oversimplify the process, but to over-emphasise the linear decoding of text. This is indeed one technique available to a reader, but the dominating skill is the continuous adjustment of reactions as the whole package of (a), (b) and (c) unfold in the text. In some circumstances, for example, when the actual and the target readers' goals do not match well, the decoding of the text might be considered a last resort rather than the main activity. The reading process sketched above seems to fulfill our criteria for in­ teraction, in that there are at least two participants who are constantly reacting with each other, and, in each reading of a text, a set of outcomes . The dif­ ference, compared with face-to-face speaking, is that the exigencies of the medium oblige one of the participants to be only provisionally represented at the writing stage, thus complicating the process for both parties. A simple diagram may show the contrast.

256

I M P L I C A T I O N S OF D I S C O U R S E AN A L Y S I S

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Halliday's Categories of the Theory of Grammar i s used i n description. When a participant initiates an exchange with an elicitation, the next communicative event is classified according to its relationship to the initiation. A second participant may decide not to continue with the exchange as initiated, and has a range of challenges, sidesteps and so on whereby he replaces the initiation with one of his own. If he decides to make a response, then his move will be interpreted within the range specified by the initiation, and his tactical options lie between a minimal response, closely following the terms of the initiation, to a quick termination of the exchange followed by an initiation of his own. If he decides not to take this latter option, then he opens up an opportunity for the initiator or a third participant to produce a follow­ up to the I-R pair, or to proceed with a new exchange. At every stage, the prospective analysis is sensitive to what is going to follow. Sometimes a string of questions and answers, or statements and muttered acknowledgements may suggest that only initiations are prospective and not responses or follow-ups . But this is merely what happens in a variety of un­ balanced discourse, which obscures the options which are in principle available to each participant in each turn. Some discourse types may have the imbalance built into general constraints on participants, like classroom language, but the general statements about prospective structures are distinct from any one discourse type, and are similar in abstraction to syntactic statements. All moves in discourse, then, have a prospective function. The exercise of this function enables speakers to signal what course the interaction should take in order to fit in to the structure of discourse-the perspections are of elements of structure that may or may not be realised in the ensuing interaction. Retrospective patterning is quite different, so much so that it is possibly misleading to call it structure at all . It is interpretive, fitting an explanation on what has already occurred, and it seems to be mainly semantic in nature. Although some cohesive patterns are stylistic, in that they work by recalling previous structural patterns, the main focus of retrospection is upon the meaning of what has occurred rather than the form, or the actual words. A directional view of language sees this as quite natural. Where meaning is created by acts of choosing, as in syntax, then it is proper to highlight the range of choice and the organisation of the process, to study it prospectively as it unfolds. But once the process is over and a particular structure has been formed, our attention shifts to what it has contributed in semantic terms to the discourse. By then the way it was done is hardly relevant. Some of the meaning may have been created by nonverbal signals, and participants in a discourse often have little accurate memory of how the meaning got there. The student, coming along afterwards and having the advantage of the whole text and no constraints from real time, may be tempted to propose more clear structural links than the original participants perhaps were aware of. In the written language we all expect the retrospective links to be clearly indicated because the writer is not constrained by real time. The spoken and written forms of the language exhibit a major difference in directionality. In general, spoken language shows prominent prospective patterns with subsidiary retrospective ones, and the written language shows the opposite. A writer has no great need to struggle for attention on a minute-by-minute basis. His prospections concern matters like clarity and interest and there is no

JOHN McH . SINCLAIR

257

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

likelihood, in most writing, of anything preventing him fulfilling the ex­ pectations that he arouses. He is not overtly competitive. The speaker, on the other hand, may not have time to tie in what he wants to say with what has gone before; if he does not give priority to prospections, he may not get a chance to say anything at all. These are generalisations. Formal lectures will show similar patterns to writing, and informal correspondence can be very like conversation, and there will be many in-between types. The distinction, if there is a sharp one, is not between the two media of communication but whether or not the composition is limited by real-time considerations. Real time can be more or less of an influence but we do accept a convention in most written text that the author has left no trace of real-time influence in his work. The patterns of retrospection cover what we call topic, theme, content, subject-matter and notions. None of these can be accurately defined, because there is something ad hoc and locally interpretive about retrospective patterns in text. Some recall links may not be very strong, but may be brought into focus by a clearer pattern nearby-like sound patterns in poetry. The reader or listener is often aware merely of a semantic coherence running through the discourse, which can be named at any time as a topic or theme. On examination it may prove to have turned and twisted in all sorts of ways that participants were not conscious of, because each link was reasonable enough . If an argument comes to a surprising conclusion from believable premises, we tend to go over it again to find out what went wrong. If two people fmd themselves talking on a strange subject, they sometimes take pleasure in backtracking to find out how they got there. The analysis of retrospective patterning is not very far advanced, though a great deal of the patterns are already well described in cohesion studies and papers on information flow, and in the standard works on formal logic. After this brief account of some major features of discourse, we can con­ sider what relevance there may be to EFL methodology, and ESP in particular. With the current interest in notional and communicative teaching, discourse analysis might be able to offer a firmer theoretical and descriptive foundation for methods which rely heavily on intuitive descriptions. The studies should provide a large and organised array of facts which would serve at least as a checklist for a communicative syllabus. They should provide a packaged set of specimen texts to be used or adapted as models. They should provide a typology of conversations. They should give grounds for supporting certain classroom activities-like interactive games-rather than others . The structure of discourse is tactical, in large measure. It is capable of great subtlety, and breaks its own rules much more than we are accustomed to find in the syntax of sentences. We must therefore be cautious in putting it forward as a target for language teaching. It runs the ·risk of being taught in the same way as syntax, where form is substituted for skill. Whenever I present discourse to groups of language teachers and ask them to consider how it could be used in teaching the language, the almost unanimous response is to see it in terms of model dialogues, practising structured utterances, inputting language in its formal aspect, and moving from that to its functional. Or using a list of notions or functions as pegs on which to hang a miscellany of language forms. To revert to my analogy early on in this paper, that is like teaching syntax

258

I M P L I C A T I O N S OF D I S C O U R S E A N A L Y S I S

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

through phonology, o r worrying about students' phonological knowledge while practising syntactic skills. Before conversations could safely be put forward as models for learning, they would have to be processed by the teaching profession in order to be validated as good models. Many of the features of a discourse are only relevant to the original performance; many are of doubtful status from a pedagogical point of view-some kinds of hesitation or of interruption. The whole process would have to be evaluated before being taught. Then it must be remembered that interactional skill is the objective, and no amount of learning of a repertoire ensures successful outcomes, any more than grammar rules make for correct syntactic performance. The constructional rules must be acquired somehow or else the performances will not be suc­ cessful. If one feels that structural inputs do not mix with the teaching of communicative skills, how is it done? Do we leave everything to the learner, and just teach him good strategies? Do we have parallel or alternating courses? The syllabus for a traditional English course is set out in terms of the syntactic and lexical analyses that underlie the accepted view of language. A similar job could be done for discourse once the analytical frameworks are established, but it may not be felt to be desirable, because of the emphasis that it would place on the formal structuring. (It might, however, be very useful in monitoring and evaluating so long as it did not set up too marked a backwash effect.) The specification of a syllabus for discourse would rather be in terms of participants and outcomes, since these are the principal variables, and the essential difference from single-source organisation. For practical purposes, that means the exchange and above, and a lot of class time would be com­ mitted to interaction practice. It is important to be quite clear about this inference for methodology. The level of discourse is established in the first instance because of the need to deal with a kind of structuring which is not accessible with single-source descriptive apparatus. It is only in interaction that such structuring can be observed and practised because the sharing of control is not something that can be self­ generated. The teaching of discourse, then, does not involve simply a transfer of at­ tention to a different set of categories and rules, while the same sort of activity is practised around them. Nor is it j ust a change of focus from the structure of sentences to the structure of higher units like paragraphs, with extensions of familiar exercise types. It involves a completely different set of priorities and techniques from those to which we are accustomed. The process of interaction has to be created as part of the teaching activity, and everything else subor­ dinated to the development of language skills which are sensitive to discourse. As we shall see, this change around does not exclude the practice of language patterns in their non-interactive mode, in the long-established routines of · language teaching. The classroom becomes a Discourse Arena. Instead of just negotiating the basic progress of a lesson, the lesson becomes a series of negotiations, in pairs and groups , achieving things. This continuation preserves much of the original motivations of our ordinary social selves, and contrasts with simulations which are of a different character and produce different motivations.

JOHN McH. SINCLAIR

259

So much is misunderstood nowadays about concepts like real, genuine, authentic and so on that one must go carefully. But where the identity of a

The steps in procedure therefore would be: the model identifies an im­ portant skill area, expressed as types of exchange and/or sequences of ex-

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

participant does not change, nor does his physical setting, and he is not asked to imagine anything at all, then one can assume that the natural com­ municative drives and skills are available for exploitation . The teacher will not b e able t o remain i n the roles o f dominating participant and residual speaker if the students are going to practise a wide range of oral fluency skills. The teacher becomes much more of a manager, contriving the emergence of the sort of talk that is required, controlling the situatio!lal variables so that the discourse produced exercises the skills that are currently the focus of attention. As with any other movement in teaching, there are serious implications for the training of teachers in any attempt to teach language as discourse . The profession as a whole will have t o devise and learn a whole variety o f new techniques, controlling the information flow, the linguistic and social conventions, prescribing the interactional formats for a variety of com­ municative exercises. Some ESP work already approaches this kind of development, because of the orientation provided by the notion of purpose specified in terms relating to the special area of study or training. Where language teaching is very closely associated with a vocational or similar course, · and the purpose-ohhe language learning is very ·clearly the improvement of efficiency in some area of work or study, the evaluation of tlie language teaching and learning may be indirect, expressed through the overall im­ provement of an individual. But it is possible to specify purposes and examine outcomes within the language behaviour, using an adequate model of analysis. The profession will require such a model in order to extricate itself from over­ dependence on the detail of other work or study areas, to convert the output of needs analyses into syllabi, and to lay the foundations for a proper professional stance for a language teacher, emerging after the ESP upheaval and the communicative teaching movement. As an outline example of the relationship between model and method, let us consider the categories of exchange, move and act in discourse. The Exchange is the minimum unit of interaction, where more than one participant is in­ volved and there is a specific outcome. The teaching of how to make ex­ changes, and all higher units and broader patterns, involves interactional practice and example. This is presumably the major focus of communicative teaching. Below exchanges there is no discernible communication. The Move is the contribution of one participant to an exchange. It is the construct of one individual, but it is designed to be used interactively, and can often be quite elaborate. Students learning how to discourse in English would have to spend a lot of time building up moves for various purposes, and ex­ ploring the options open to them. Such work, however, would be quite meaningless without a context of exchanges and outcomes, because the principles on which the moves are constructed are interactional. But if students are brought to feel the need for an extensive repertoire of moves, and find that the acquisition of them pays off in subsequent communicative tasks, they may learn them in a fairly natural fashion.

260

I M P L I C A T I O N S OF D I S C O U R S E A N A L Y S I S

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

changes. The teacher devises interaction activity to practise and develop the skill, starting perhaps with an exploratory session to see how competent the students are, how much they can transfer from the native language, and so on. If need arises, the teacher changes the focus down one rank to Move, and teaches a repertoire of valuable move types. Much current communicative teaching activity, and many recent proposals, concern the rank of move, but without its communicative context . That is perhaps the closest that one can get to discourse, while keeping language structure as the main object of study. However a student has acquired the basic syntactic pattern of the language, it is vulnerable to teaching of a communicative kind. Attention shifts off the structural detail and, so to speak, distracts the student, who might well be able, under test conditions to perform constructional tasks immacu­ lately. Obviously communicative fluency requires a combination of both accurate construction and adequate interactive ability, and so teachers must be prepared for some remedial work at least on the structure of the language. For some students the structural teaching may not be remedial. They may be learning the syntax as they go along, feeling their way towards the structure of the language through the interactive work, and using structural sessions as opportunities to confirm and organise the patterns to which they were ap­ proximating. They will learn their syntax at the rank of Act. The Act, as the smallest unit, has no structure at the level of discourse, just as a morpheme has no syntactic structure by definition. If we are con­ centrating on the construction of acts we are working in the realm of syntax and lexis, and not in discourse at all. The main value of the approach to syntax through acts is the link with moves and beyond, which gives an overall coherence to the teaching. Structural and communicative teaching are seen as complementary to each other, with the communicative framing the structural. This seems a safer method than one which implies much more direct correspondence between syntax and discourse, and uses the rank of move almost exclusively. From the point of view of discourse, the written medium is more complex than the spoken, since real-time interaction is bypassed and has to be imagined. A similar sophistication arises with tape recorded materials for listening, and combinations of these like Reading-while-Listening. The classroom can be used in many ways to teach these further skills, and to use some features of the written language in oral teaching. The pedagogical value of building written text through interaction is frequently attested, and the techniques are capable of much more develop­ ment, so that the target reader can be directly built in to the text. In reading we are beginning to see ESP material which consistently prescribes a purpose for the reader/student which is clearly distinct from the assumed purposes of the target reader, and thus leads to practice in reconciling the two readers. A fuller understanding of the processes would lead to a much richer variety of pur­ poseful material. The requirement of retrospective consistency in written text, discussed above, is handled prescriptively by time-honoured techniques. The develop­ ment of analytical systems is allowing a descriptive dimension to arise, and this

J O H N M c H . S I NC L A I R

261

should help a great deal, particularly in teaching the mastery of specialised

forms of writing in English . Oral skills are sometimes worth transferring directly into the written medium, so that the interaction process can be slowed down and considered carefully. This is particularly helpful in ESL when there is an indigenous spoken variety of English with local syntactic features that the teaching is intended to counterbalance. The tendency of the written form to be closer to an international standard can be used in transfer. The teaching of the ritual end of spoken language, the courtesies and the conventions, can be facilitated by a written language stage. All these suggestions maintain the basic premises of this paper-that ap­ propriate discourse can be created and controlled in the classroom; and used as the basis for communicative language teaching.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

DIS COURSE AS S ELF-EXPRE S SION­ ON T H E REDUCED PERSONALITY OF T H E SECOND-LANGUAGE LEARNER PETER HARDER

University of Copenhagen 1 . I NT R O D U C T I O N

2.

T H EO R E T I C A L P R E L I M I N A R I E S

Ever since the heyday o f the standard theory, the accepted notion o f what a language is has been growing more and more blurred round the edges. The general trend, which is perhaps most evident in the development of the work of George Lakoff ( 1 97 1 , 1 973, Lakoff and Thompson 1 975) has been to take account of an increasing number of phenomena which were previously con­ sidered extra-linguistic, in one's theory of language. While I think this development has been theoretically inavoidable, in that dilemmas within the theory itself could not be resolved without taking a wider context into con­ sideration, I do not think that any way of handling these dilemmas has yet provided the necessary clarification. The situation is still very much as Widdowson described it in 1 973: linguistic forms and the use of these forms are confused, and no satisfactory solution of the problem is within sight. I do not propose to present one here, though I do intend to make some of the distinctions I believe necessary in solving the problem. The reasons for wanting to go into this discussion are not only academic . From the applied perspective the question raises itself a s the problem of what it is we teach when we teach a language. The problem of what exactly it is you are doing, when you are teaching a language will inevitably be raised, if not by your own subconscious mind, then by the pupils in the form of questions such as: why do we have to discuss this subject? Why should I say something if I don't have anything to say? After all, we're here to learn the language, aren't we? What I shall try to do then, in this part of the paper, is to make as rigid a

Applied LinguistiCS, Vol.

I, No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

There are two sources for the present paper: one rather theoretical, and one very practical. From a theoretical point of view, I have been interested in the question of what precisely a language is, of where the boundaries are of such an object in relation to the context that it is always, in practice, inextricably interwoven with. From a practical point of view, in my experience as a teacher and learner of English I have been confronted with a number of behavioral mechanisms that seemed to be inseparable from the situation in which one uses a foreign language, which I should like to place somewhere in the picture of language learning. Since their place seem to be within the area of discourse analysis, I have taken this opportunity to present a hypothesis concerning the place of these phenomena in relation to language and learning, suggesting some adjustments in our view of the situation and activity of the language learner in the light of this hypothesis.

PETER HARDER

263

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

distinction between linguistic and patterns of action as possible-in order that I may show, in the next part, just how close the interdependence between the two phenomena are. It has often been said (cf e.g., Giles and Smith 1 979:54) that one should not make a hard and fast distinction between the means of saying something, and the message conveyed through these means; but the point that is expressed in such statements is better expressed by saying that our choice of means crucially influences the contextual import of what we say. The distinction is necessary precisely because we want to explore the relationship. In defming language, it is customary to begin by making a distinction between system and actualization , and although few would now accept either langue/parole or competence/performance without serious modification, I think most people would probably accept that it is very difficult to do without some version of this dichotomy. The version I have in mind would call for a description of, as one half of the dichotomy, all the linguistic conventions the speaker can invoke in a given speech community, comprising the potential that is at the disposal of the speaker. This would take the form, in an individual speaker, of a communicative competence that enabled the speaker to use all the linguistic items available for just the purpose it was recognized to serve among the speakers of the community. In order to use this 'linguistic com­ municative competence', the speaker would have to possess other kinds of competence, too; but analytically, the knowledge (i.e. ,_knowledge-how) that concerns linguistic items can be kept separate from the rest. Aetua/isation of this potential would then consist in invoking these conventions for the purpose of concrete communicative action. Linguistic performance thus requires two types of decision of the speaker: deciding what course of action to take, i.e., forming communicative intentions, and selecting the appropriate linguistic items for the encoding of these intentions. Both courses of action and linguistic items form patterns, since neither conventions nor human behaviour is random, but the patterns of action reflect the way in which the individual leads his life, and the patterns of language reflect the structure of the language that is used to encode the course of action. How does discourse analysis fit into this picture? Sinclair and Coulthard's ( 1 975) concepts have come into existence in the course of working with a series of texts, and therefore they cannot immediately be placed in relation to this sketch. Sinclair and Coulthard make clear that they do not conceive of discourse as a type of supra-sentential grammar, proceeding to place it as another level of description; its relationship to grammar is then compared with the relationship between grammar and phonology. Above discourse in this hierarchy comes non-linguistic organisation (p. 24). Discourse, accordingly, emerges as the last stop before we cross the border to non-language. However, if discourse is to be linked to grammar and phonology, it must be seen as organisation not only in the body of texts, but also on the other side of the dichotomy, as part of the speaker's know-how. This is not only a theoretical necessity ; again the applied purpose requires clarification: if language includes a level of organisation called discourse, this must presumably be taught-and we cannot teach language as a body of analysed texts. What the learner needs to know is not something that exists in terminated instances of com­ munication-he needs to learn it as know-how, to be invoked when he needs it . Can we transform Sinclair and Coulthard's discourse categories into aspects of

264

D I SCOU R S E A S S E LF-E X P R ES S I O N

linguistic communicative competence, as outlined above? If we were to do that, we would have to show that just as the learner needs to know the sounds of a language and combine them into words with the ap­ propriate lexical meanings, and just as he would have to know the gram­ matical relationships that obtained between different words and signal them through appropriate . inflection and word order, the learner would have to know the discourse categories in which sentences belonged. This, I suspect, is a fairly common way of thinking of the matter-but I think there are reasonably convincing arguments against it. Knowing words, affixes, and phonemes means knowing how they are combined according to the rules of the language. If you speak the language without knowing the rules, you will give yourself

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

away as a foreigner, as someone who does not know the linguistic conventions of the community. Does this apply to Sinclair and Coulthard's discourse categories? Can we envisage a combination of two moves which would yield a non-English utterance? I find this implausible. Within the framework of the book, a wrong combination of moves might give the speaker away as a non-pupil or a non­ teacher (as anyone who has done teacher training will know)-in exactly the same way as if he mismanaged the organisation at the level Sinclair and Coulthard call non-linguistic . Therefore, for my purpose, discourse must be seen as dealing with patterns of action rather than patterns of language. The patterns that are established in the book are a description of ongoing action, and we recognize the structure through comparing it with other examples of the same type of interaction-regardless of whether the participants by ex­ treme ingenuity found a way of never using the same words or grammatical structures for the same discourse contribution type twice in a row. If it is objected that a combination of moves might well betray a speaker as non­ native, because no native would yank such moves together, I would then suggest that it is the foreign habits ofaction that are revealed. This is a pattern in the lives o f the participants, not in the medium of communication they use. I am not accusing Sinclair and Coulthard of overlooking the action aspect of discourse; their awareness of this aspect is abundantly clear. However, the point of view from which it is all a question of levels of organisation makes it tempting, in an applied perspective, to look for errors or organisation as if they could be rolled together in one category, as we shall see later. Therefore I think it essential to be aware of the distinction between that type of organisation which is imposed by the code in which one tries to render one's contribution, and that type of organisation which it is up to the speaker to impose on his own behaviour. I would like to add some remarks about the reasons why Sinclair and Coulthard call discourse a linguistic level. As I understand it, this has to do with the existence of linguistic signals to mark the various discourse categories, such as 'well' or 'right' for ' framing' etc. However, from the point of view of linguistic potential, this should be seen as resulting from lexical features at­ tached to the words in question. Knowing the word 'right' entails knowing that it can be used to signal that the speaker considers the previous topic closed; the awareness o f how that functions in class is part of the teacher's professional competence. Similarly with other elements that are o ften used to argue for structural relationships of a linguistic nature between sentences (or utter-

PETER H A R D ER

265

ances). Knowing the meaning of definite noun-phrases, for instance means knowing that they are used to refer to things . What things you have to find out in each individual situation, whether they are introduced by the preceding discourse or are part of the pre-defined situation.

3.

I L L U S T R AT I O N A N D Q U A L I F I C A T I O N

which it may be desirable for them to know. For instance, since a democratic society presupposes the possibility of debating important decisions before they are taken, it is to be expected that one of the aims of the educational system will be to enable the pupils to master the pattern of action known as arguing. In the Danish system, this is part o f the aims of primary as well as secondary school. This means that the school tries to influence the type of intentions the pupils will form in a certain type of situation, creating a more complex and delicately organised pattern of action than would otherwise have occurred. It is the social behaviour of the pupils that is being influenced here. Similarly with that activity which traditionally occupies a large part of written mother­ tongue education, essay-writing. If pupils make a mess of it, I have argued in Harder ( 1 979) that it is because that type of social behaviour is meaningless to them, only secondarily because of linguistic deficiencies. Conversely, in foreign language teaching the primary aim will be to get the learners to master the set of linguistic conventions that can be invoked in the foreign community. Although various patterns of interaction will, of course, enter into foreign language-teaching (cf. , Sinclair and Coulthard for a description of some of them), they are not the essential thing; typically one will presuppose the ability of the pupils to form appropriate intentions and focus on their coding ability (though this may not be the most productive way of tackling the problems of language teaching). However, having been made, this distinction is now in need of qualification . In a number of respects action-planning is closely dependent on knowledge of conventions . This is true to the extent that certain communicative intentions can only be achieved through specific linguistic items . Extreme cases of this are ritualized patterns of communication, such as greetings : as described by Halliday ( 1 973, p .97) choice between different types of greetings is equivalent to choice between different words. In this category we also find Austin's in­ stitutionalized performatives, such as ' I name this ship " Liberte" ' . It is worth stressing, however, that all lexical items probably have some element of action­ specificness; if you want to refer to a certain class of things, there will generally only be a very limited number of words that are appropriate for the purpose in the context in which you need them. The notion of idiomatic language can also be seen in this context: when a phrase is 'correct, but non-

'

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Having made this distinction between action (including discourse) and language-as-such, we can use it to clarify the aims of different kinds of language teaching. In mother-tongue teaching we do not primarily aim to teach people language-as-such, since we do not have to teach people their own language. (In cases where there is a marked difference between the standard language used in the school system and the pupil's vernacular this does not apply-but then it is doubtful whether in fact it is the pupil's 'own' language that is being taught.) We do, however, wish to teach them certain types of discourse, patterns of linguistic action with which they are unfamiliar, and

D I S COURSE A S S E L F- E X P R ESS I O N

266

English' , it means that for the purpose at hand, there are some recognized expressions that one should use. Because of this dependence, extending one's action potential inevitably extends one's language. But even where the action performed and the linguistic conventions invoked are less obviously tied to each other, the connexion in practice is clear. In those cases where com­ munication is less ritual, the act performed through the use of linguistic convention will not be exhaustively described by enumeration of the con­ ventional content. In using the conventions , we therefore have to calculate their value in relation to the context-otherwise we will be unable to predict what act we are performing. Therefore learning (and teaching) languages will always be a matter of discourse, of using linguistic conventions to fit into the ongoing interaction. The medium itself is not accessible in any other way. D I SCOU R S E AND LANGUAGE LEARNING

4. 1 Linguistic patterns and the learner role. I have now tried to show how discourse patterns should be understood as reflecting the speaker's personal choice of actions in concrete situations, rather than as aspects of some broadly defined linguistic competence. In the following I shall attempt to demonstrate how this way of viewing discourse contributions as reflecting the speaker's social identity may add some shades to the picture of the learner that can be extracted from some of the literature on language learning. As we have seen, it is natural to view foreign language learning as having primarily to do with coding, i.e., learning linguistic conventions, so that it is not surprising that research interest has tended to concentrate on linguistic patterns. However, in the case of spontaneous learning, it is clear that linguistic patterns must be acquired through the discourse in which the learner is involved. Typically, this is the only assumption about discourse: that it should be there-so that the learner or perhaps his language acquisition device could have some material from which to derive and refine hypotheses con­ cerning structure. Nor is there any reason to doubt that to a certain extent all messages and interactions are grist to the mill of grammatical learning. However there are also examples of learning in which there are clear discourse implications of the learning process. Corder cites the following dialogue as an exemplification of the hypothesis-forming activity of the learner (the distinction between mother tongue and foreign language can be ignored in this context): Mother: Child: Mother: Child: Mother: Child: Mother: Child:

Billy had his egg cut up for him at breakfast Yes, I showeds him You what? I showed him You showed him? I seed him Oh, you saw him Yes, I saw him

Corder ( 1 967, p. I l l ) describes this as a case where the child tries out three hypotheses concerning the proper form of the word he's looking for-but clearly this is not a case of learning going on as an undercurrent to whatever

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

4.

PETER HARDER

267

discourse might be going on. The situation clearly changes in character from whatever it was to a learning situation, where the three middle utterances of the child serve the purpose of clearing up the linguistic difficulty. This requires the child to accept the role of learner-instead of, e.g . , activating the 'pop go weasel' effect (cf. , Brown & Bellugi 1 964), opting out of the learner role in order to get on to something which was less of a fuss and bother. The role of learner is a discourse slot into which one may choose to fit (or which one may want to avoid) based on what kind of interaction one would like to go on and the role one would like to have in it. Thus the 'grammar-acquirer' is at the mercy of the 'discourse participant' . Some important factors which may in­ fluence the outcome will be discussed below (in the section entitled 'The compleat learner').

4.2 Communication strategies: Ieamer r,eduction Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

However, though interest in acquisition of grammar still has a large share of the interest in language learning, the outward trend in linguistics is also evident in applied linguistics. The justification for emphasizing the speaker's responsibility for some patterns, while other patterns are facts about the medium he uses, is thus not that awareness of the social aspect of language is ·lacking, but rather that the social facts tend to be described more or less analogously to grammatical facts. Thus the tendency has been for the ex­ panded perspective to be understood as looking for what 'non-nativeness may mean beyond the realm of grammatical correctness and into the murky realm of appropriateness' (Tarone, Cohen and Dumas 1 976, also quoted by Kasper 1 979) or looking for 'covert errors' (Varadi 1 973) as opposed to the hitherto investigated overt errors. This perspective on the learner fits in with the purpose of giving a 'learner profile', describing all those things which can be seen as characteristic of learners. One of the aspects which have been added to this 'learner profile' in recent years is 'communication strategies' (Selinker 1 972, Tarone, Cohen and Dumas 1 976). This concept is defined by Tarone, Cohen and Dumas as 'a systematic attempt by the learner to express or decode meaning in the target language, in situations where the appropriate target language rules have not been formed' (p. 78). In other words, all ways of getting round a problem in which you find yourself because of insufficient command of the target language. Among these we find 'epenthesis' , which refers to the learner strategy of inserting schwa vowels to break up 'un­ pronouncable' consonant clusters. This is placed alongside 'avoidance' , which subclassifies into e.g., 'topic avoidance' : not talking about a topic for which you do not have the words. My purpose here is to emphasize the difference between these two types of procedure. The first is a way of getting round problems with the linguistic medium with as little damage to your com­ municative purpose as possible. The second is a way of letting linguistic problems prevent you from performing actions you might otherwise wish to perform. Within the metaphor of strategy, one might compare the first with a commander who attacks in spite of deficient artillery, and the second with one who retreats across the river. If we emphasize those aspects of linguistic behaviour which express the speaker's personal choice of action, the latter is not merely an instance of 'learner deviation ' , it is an instance of a learner's role in events being reduced. The learner is not free to define his place in the

268

D I S C O U R S E AS S E L F- E X P R E S S I O N

4.3 The Compleat Learner Hatch ( 1 978) makes a point of studying learner-native dialogues as discourse, e.g . , as interaction, rather than linguistic material to be interpreted as in­ dicative of various stages of gramm atical learning. Her position is that learn­ ing, instead of being parasitic on discourse, is dependent on discourse structure, i . e . , on the pattern of action in such dialogues. Quoting Scallon ( 1 977) she shows how this might be the case with simple subject-subject­ predicate sentence structure evolving out of the action pattern of topic nomination, later followed up by comments on the nominated topic. Having exemplified how the structures of learner-native dialogues can be seen as promoting language learning, she proceeds (although cautiously) to recom­ mend a number of discourse choices which should be taught to the learner to facilitate language learning. This is admirable in that learner contributions to conversation are understood as actions reflecting the speaker's choice in a situation rendered difficult by limited resources; and she uses this un­ derstanding to point out a way for others to follow. However, like many others, she overlooks the consequences of following this advice, if we view the

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

ongoing interaction as he would like; he has to accept a role which is less desirable than he could ordinarily achieve. This is an inevitable consequence of the point mentioned above: that free choice of action depends on free access to convention. This is especially relevant in case of people with an integrative motivation, since an instrumentally motivated person may have quite limited ambitions anyway. But most people will probably be aware of the un­ comfortable feeling of having to choose between rather undesirable alter­ natives, when acting in a foreign language environment. Most readers will recall how being a foreigner entails not understanding jokes and therefore having the choice between sitting tight or being the simple­ ton who asks for the explanation while the others move uneasily on their chairs; of having so little redundancy that one requires repetition even of utterances the linguistic items of which one is familiar with, thus assuming the unenviable status of the deaf person in conversations; of saying 'yes' where one would have liked to say 'you can't win them all' or 'phew ' ; reduction is thus far from being a matter of saying quantitatively less. As examples of those things that are characteristic of learner reduction I can refer to Kasper ( 1 979) who shows that politeness markers are scarce in learner speech; Varadi's examples (Varadi 1 973) show that only the skeleton remained of stories retold in the target language; in other words, the learner is a coarse and primitive character from an interactional point of view. To those problems are added, of course, the role that is already defined for you once you have given yourself away as a foreigner. Since people, through speaking with foreigners, have more or less the experience of them outlined above, according to a well-known psychological mechanism they adjust their own behaviour and their interpretation of the foreigner's contributions accordingly, so that even if you do succeed in finding words for your clever remarks, you are likely to be politely overheard. A foreigner is not permitted to go beyond a certain limited repertoire; if he starts swearing fluently, for instance, he is unlikely to achieve the conventional communicative effect, i .e. , underlining the serious objections he has against the situation in question.

PETER HARDER

269

matter from a purely interactional point of view. Because she is mainly concerned with the problem of how the speaker best moves through the learning process, she does not think of the problem of planning one's actions in a way which is consistent with one's need for maintaining an acceptable role in the foreign language group. Discourse, as it were, is in danger of being chained to the chariot wheels of learning. Some of her recommendations are:

(Learners should) practice saying huh, echoing parts of sentences they do understand in order to get the rest of it recycled again, pardon me, excuse me, I don 't understand etc. They should be told to use uh-uh-ah-ah or whatever fillers they can to show the native speaker they really are trying. If the learner gets to recycle the same topic several times with the same or different native speakers, he will then have the vocabulary and know the possible questions that will be asked. When he's got that much, he can recycle the topic again with another person and pay attention to his syntax and morphology.

As far as I can see, one gets the picture of a very well-defmed social role, when one imagines the learner: assiduously saying huh whenever there's a pause, always repeating bits of the previous utterance, blocking out interruptions by saying uh-uh , while he tries to find the word beginning the same topic all over again immediately he's struggled through it (in order, later on, to continue with syntax· and morphology), sticking like glue to unfortunate natives who said hello, etc. The picture that emerges is that of an utter pest. And this the learner, unless he's an unusually callous or charming person, is likely to be acutely aware of. This creates a dilemma for all language learners who are trying to practise their target language in a native environment: either they do what's good for learning and forget all about those criteria that are ordinarily decisive for one's pattern of action, or they prefer generally to remain silent, whenever they feel they can't act the way they would like. Most learners will probably, in deciding what to say (if anything) have a sort of cut-off point for the reduction they will tolerate, below which silence is preferable. Instead of seeing silence as the extreme point on the scale of message reduction, it can also be seen as the alternative to it. This dilemma, naturally, does not invalidate Hatch's advice; it should, however, be understood not merely as a list of things to do , but as a type of role which goes with the position of learner. Those who feel they are making fools of themselves and that they are exploiting their native speaker friends in an unacceptable way, may feel relieved of the bad conscience, at least partly, if they are told that this is not just their individual dilemma: in order to be a wit in a foreign language you have to go through the stage of being a half-wit­ there is no other way. If the problem is not addressed explicitly, learners may just be aware of it as a constant internal resistance against opening their mouths. What can an awareness of this dilemma do for us in understanding the process of language learning? There are two points on which it fits in with already established learning patterns. Among the reasons for the speed with which children acquire new languages in a natural environment this problem

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

. . . he should be taught not to give up in any contact he has with a native speaker. (Hatch 1978 : 1 34.)

270

D I SCO U R S E AS S E L F- E X P R E S S I O N

might offer one. Anybody who has lived with a child in the house will know that there are virtually no limits to how unreasonably children will behave in order to remain in the ongoing interaction. There may, in other words, be no cut-off point below which children prefer to drop out rather than accept a role as pest or simpleton. Also, children are used to being in 'client' position, having to ask for everything and depending on others for help in all problems. This means, essentially, that children are likely to follow a good deal of Hatch's advice.

4.4 Interlanguage as independent system orfragment

French as erroneous English. However much one may sympathise with the aim o f reforming teachers' attitudes to errors, this is hardly a tenable view of the learner's situation. Within this framework, two reasons can be given for revising this picture. If language-as-such is a matter of knowing linguistic conventions, the learner does not have his own language; the conventions he is trying to invoke are those of the target language community. This might not matter if the learner's language, as Corder says, could be viewed as an independent dialect within the target language community. But languages (and dialects) are customarily assumed to be equivalent with respect to expressive capacity; certainly no Frenchman (to use Corder's �xample) would be very happy if he was told that French was less capable of expression than English. If, however, we asked interlanguage speakers if they felt they could express all they wanted in the language that Corder calls their own, I think the answer would be emphatically no. Without wishing to discard the gains of viewing learner language on its own terms, I think that view must include the understanding o f it as fragmentary. Furthermore, Corder makes a point of talking about learner language as a system, in contrast to teacher talk, as a mode of discourse in which the teacher makes use of his resources in a particular way. If we understand the linguistic resources of the learner as fragmentary, rather than constituting a 'whole' language, we can complete the picture by suggesting that learner performance should also be understood as a particular kind of talk, the most characteristic feature of which is that where his own language is incomplete, he tries to eke it out with a number of means which have been touched on above. The variability and permeability of the system could thus be reinterpreted (consider also Faerch 1 979) as a consequence of the kind of talk in which the learner constantly feels the need to say more than his resources permit.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

The last point I wish to make in this connexion concerns the understanding of the status of learner language. For a number of years researchers, with Pit Corder as one of the champions of the cause, have been at pains to emphasize the positive aspects of the linguistic performance of learners . One of the consequences of this tendency has been to reinterpret differences between learner language and target language as achievements rather than deficiencies: they are (in most cases) signs that something has been learnt rather than blemishes. It is one of the major points in Corder ( 1 978) that the learner's language should have the status of a language in its own right, with its own criteria of gramrnaticality and everything-to drive his point home, Corder compares the practice of calling learner-specific features errors with viewing

SOME PROBLEM S CONCERNI NG TH E EVALUATION OF FOREIGN LANGUAGE CLAS SROOM DI S COURS E W I L L I S J. E D M O N D S O N

Ruhr-Universitiit Bochum Teacher: Mary, would you try to take the part of Dr. Headache? Mary: No INTRODUCTION

1 . D I S C O U R S E A N A L Y S I S A N D L A N G U A G E T E AC H I N G The potential for confusion over the use o f the term 'discourse' or 'discourse analysis' is considerable. Widdowson's phrase 'the use of sentences in com­ bination' (Widdowson 1 973, 66 ff.) is a useful starting point in coming to grips with the term. The phrase may be seen to imply two oppositions, firstly be­ tween sentential and suprasentential units (the sentence v. sentences in combination), and secondly between language units viewed as pertaining to a self-contained linguistic system (elements of which may or may not be viewed as linking with external states of affairs via reference, truth-value, propositional content and so on), and language units viewed as belonging to a complex of situational and contextual factors generally subsumed under the term pragmatic. Following these two binary oppositions, we may (with considerable oversimplification) establish the following perspectives on the study of language:

Applied Linguistics, Vol. I, No. 3

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

I wish to discuss i n this paper some aspects of communicative foreign language teaching which imply or require an evaluation of different types of discourse activity. Evaluation is implied for example in the choice of teaching items, in the choice of teaching presentation, and in the assessment of learner per­ formance. In syntax or semantics we may use notions such as grammaticality, acceptability or truth-value as evaluative criteria, however fuzzy-edged these notions might be: in pragmatics I take it that the notion of 'appropriateness' is the best that we can offer. It may be worthwhile to try to investigate how this notion might be interpreted and used with respect to foreign language classroom discourse. In the first part of the paper the term 'discourse' will be interpreted for my purpose, and some recent and ongoing tendencies in FL teaching and research which reflect such a discourse perspective will be sketched. Secondly, in the light of the peculiar nature of classroom discourse, the question will be posed 'appropriate to what? ' , and some possible answers to the question will be discussed. In the third section I shall introduce some broader issues concerning the evaluation of communicative acts in general, and then consider these issues with respect to classroom talk conducted in the foreign language. In the fourth and final section some non-startling conclusions will be drawn.

272

F O R E I G N LANGUAGE C L ASSROOM D I SC O U R S E

Sentence grammars Text grammars Speech act theory Discourse analysis

(-suprasentential, -use) ( + suprasentential, -use) ( -suprasentential, + use) ( + suprasentential, + use)

1

This seems neat enough-too neat perhaps-but fails to reveal a further distinction relevant for my purposes, namely that between the spoken and the written mode of language use. For my purposes in this paper I shall consider discourse analysis as concerned with spoken, conversational language activity. This restriction seems warranted because

If we accept this focus on spoken discourse, then in considering a discourse perspective we are viewing language as a means of social interaction, and unless this aspect of the use of language is present we are not strictly handling language as communication. Thus it is not enough to equate a discourse perspective with an interest in speech acts or 'functions' in the sense for example of Candlin et al (1976). 3 The work of and following Austin and Searle on a theory of speech acts has had and continues to have a considerable impact on foreign language teaching, and is highly important for the concerns of discourse analysis. However before we can talk of a discourse perspective on language, or of communicative language teaching, we need to consider both the illocutionary and interactional work that is simultaneously occurring when communication is taking place. Communication does not simply consist of isolated illocutionary acts sequenced in time. I propose that a communicative act be seen as both having an illocutionary and an interactional, or 'discourse' value or function. The distinction is made in the model used by Sinclair and Coulthard 1975 , where 'grammatical' and 'situational' categories are dis­ tinguished, roughly as is done in speech act theory via the direct/indirect illo­ cutionary force distinction, and then further 'discourse' categories are set up, of which the 'situational' categories are the exponents. I interpret Sinclair and Coulthard • s situational categories as similar to illocutionary acts, and their discourse categories as interactional acts, which are structured relative to one another, this type of structure being precisely what Goffman, Sacks, Schegloff, Jefferson and others seek to discover in the analysis of conversation. 4 This social aspect of talk is an important aspect of discourse analysis, and finds its reflection in some recent or current trends in foreign language teaching and research . For example (a) with respect to the goals of language teaching, a greater concentration

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

(i) this may with some justification be seen as the basic or most essential use of language, (ii) the main thrust of discourse-oriented research has dealt with this type of data-most markedly in the work of the ethnomethodologists, (iii) the central impact of discourse analysis on language teaching seems to me to have been its influence on the teaching of the spoken language as a means of communication, and (iv) the relations holding between a written 'text' or 'discourse ' , and spoken discourse are little understood, such that it is difficult to consider both types of data inside one and the same theoretical framework.2

W I L L I S J . EDMONDSON

273

2.

' L A N G U A G E' v. 'TE A C H I N G' I N L AN G U A G E T E A C H I N G

The type or types o f discourse involving the use o f the foreign language which occur in the classroom are clearly one factor determining what is learnt there. We tend to assume with every justification that one profitable way to learn to communicate in English is to practise communicating in English. However the question of appropriateness arises, as we need to ask whether the type of discourse by means of which practice is intended to take place, is in fact the type of discourse which it is intended be practised. Hence we need to ask whether different teaching procedures and strategies are appropriate to the communicative goals of foreign language teaching. Widdowson notes for example as follows: . . . the teacher in the early stages of an English course might hold up a pen, point to it and say: This is a pen. Here we have a correct English sentence. It is an instance of correct usage. But is it an instance of appropriate use? (Widdowson 1978:7)

Widdowson's conclusion is that it is not, as what is meant is something like This is called a pen in English i.e., 'naming' is going on here, not 'identi­ fication'. If, however, Widdowson goes on to argue, barometers had been discussed in a chemistry lesson, and the teacher holds one up and says This is a barometer, then the sentence 'takes on a natural function in the situation' . But note that the chemistry teacher might presumably hold up a barometer of which his class has never heard, and use the same sentence, before going on to characterise the object in question. Here too the sentence is presumably used 'appropriately', but it would appear 'naming' is going on, not 'identi­ fication' . In other words, This is a pen may well be 'natural' or 'appropriate' -

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

on 'communicative competence' , communicative 'skills' or 'abilities' is noticeable, (b) with respect to the content of courses, and materials provided in text­ books, we may discern a greater use of 'authentic' spoken language materials, and of dialogue material, (c) with respect to the sequencing of course materials, we are familiar with the notion of a situational progression, following on from the 'notional' or functional sequencing familiar from the work of Wilkins, and also illustrated in Un niveau-seuil, Coste et al l976, 5 (d) in foreign language classroom research, interest in the interactional analysis of learning-teaching events has increased, and richer analytic schemata have been developed reflecting a discourse perspective on observed classroom behaviour, {e) research into the classroom and school as social institutions, and the implications for the nature of the learning and teaching that takes place in them has flourished. 6 The above listing is not intended to imply a cause-effect relationship be­ tween discourse analysis and language teaching. It would be false to claim that the suggested discernible trends are the result of research findings in discourse analysis . To do so further implies a naive interpretation of the relations holding between 'pure' and 'applied' linguistics, which cannot be sub­ stantiated in fact.

274

F O RE I G N LANGUAGE CLASSROOM DISCOURSE

in the situation, given that the situation is a pedagogic one. W e need to be clear as to what grounds we have for evaluating various foreign language teaching

A: Okay, I'll do it tomorrow 8: Fine, thanks

to say that A did not really promise to do it, even if C knows that A in fact has no intention of doing the deed in question. To say this is in fact a nonsense-A has entered into a commitment and his or anybody else's intentional states do not alter the fact. The sincerity criterion with respect to classroom discourse

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

procedures as 'appropriate', and other pedagogic settings do not provide a relevant comparative norm, as classroom discourse of the type to be found in a chemistry lesson is not necessarily the type of discourse one wishes to practise in the foreign language lesson. Note here that the discourse structures discerned by Sinclair and Coulthard ( 1 975) are evidenced both in foreign language lessons and in other school subjects. 7 In short, as has often been pointed out, • there is a tension in classroom discourse between what is appropriate with respect to the teaching goals being pursued (target discourse), and what is appropriate with respect to the teaching setting (pedagogic discourse). In the light of this distinction I have tried to argue elsewhere (Edmondson 1978b) that systems developed for the analysis of classroom interaction may be self-defeating when it comes to interpreting the results, if the categories used in the analysis have relevance only to learning­ teaching events, and that an alternative would be to analyse the target language, and then seek to discover what is different in classroom discourse, and where the crucial differences lie. The problem is precisely that of prac­ tising target discourse inside a pedagogic setting, where what is 'appropriate' may in fact be some other kind of discourse. 9 There is then a tension between discourse which is appropriate to pedagogic goals, and discourse which is appropriate to the pedagogic setting. One way out o f this dilemma is to propose that if in the classroom setting the learner can be led to express his own communicative needs and interests, then the resultant discourse is appropriate both with respect to target situations and with respect to the pedagogic setting. I have elsewhere described this as the sincerity criterion (in Edmondson and House 1 979). Inside the conceptual framework proposed here, the sincerity requirement may be interpreted as claiming that appropriateness with respect to the intentional state of the speaker entails appropriateness with respect to the teaching goals-i.e., if the learner is ex­ pressing what he feels and needs to say, then the resultant discourse is necessarily 'genuine' , and implicitly 'appropriate' in terms of the teaching goals. 10 This sincerity criterion seems to me to be mistaken, and I shall try and show why. If we go back to Searle, we find that the conditions held to be necessary before a speech act of a particular kind-say a Promise-can be said to have occurred involve intentional states: this is clearly the case with the sincerity condition for example. From the perspective of discourse analysis however, this is to put the cart before the horse. Note that with Grice for example we assume sincerity via co-operative principles, such that the illocutionary value of what is said is derived in part from the assumption of sincerity. Thus it does not make sense for C, reporting on a dialogue sequence finishing

W I L L I S J . E DMONDSON

275

also assigns intentional states a role which they are not capable of carrying in the business of conversation, and is thus unrealistic and misplaced. Firstly, intentional states are not open to inspection. Secondly, the needs, drives and interests of the learner at the time of learning may not match the goals of the course of study being pursued, are non-predictable, and are uhlikely to be a constant for a specific group of learners. Thirdly, the classroom imposes such restrictions as a social setting that the amount of language activity that can be 'genuinely' practised in this setting is severely restricted. Fourthly, the sincerity criterion does not on its own discriminate between classroom-specific and target discourse. Consider the following extract from an English lesson, in which the teacher is practising the word prefer: 1 1 Do you prefer cornflakes with or without sugar? I prefer cornflakes with sugar I prefer cornflakes without sugar I don't like cornflakes

We tend to applaud the third pupil, but have in fact no means of knowing whether he is telling the truth or not : the first or second pupil may be being just as sincere or insincere-we cannot tell. Part of the reason the third response is striking is in fact the assumption that the learner is deliberately flouting a discourse pattern established by the teacher-;-an assumption that the learner resents beirig a learner-i.e., carrying this social role. But it is evident from the total sequence that it is pedagogic and not target discourse that is occurring here. Again, we may claim with every justification that, for example, the initiation of repair work and the performance of various non-co-operative communicative acts we may gloss as 'contradictory' are valuable com­ municative skills for a learner, but are under-represented in classroom discourse relative to other settings (see e.g., Somig 1 977, von Unwerth and Buschmann 1 980). A conclusion drawn by a proponent of the sincerity criterion would be that the conventions holding for classroom discourse be changed, such that learners do freely initiate repair work 12 and do 'contradict' their teachers . The conclusion does not remedy the problem however, as the learner will perform the desired communicative act only if he disagrees, or fails to understand. The motivation for changing the nature of the classroom as a social setting cannot be simply derived from an analysis of the problems of teaching a foreign language as a means of communication in this setting. The above argumentation is not to be seen as denying the importance of catering for learner interests and experience in the choice of the content of teaching programmes and materials , but claims that 'sincerity' itself does not suffice as regards the appropriateness of the discourse that results to com­ municative teaching goals . The only point of any discourse is to produce outcome: in classroom discourse, the desired outcome on the part of at least one participant involved-the teacher-is we may say learning. The discourse that occurs in classroom settings is necessarily affected by this. 1 3 3 . T H E E V A L U A T I O N O F I N D I V I D U A L C O M M U N I C AT I V E A C T S

If we accept that spoken discourse is social activity, then in evaluating an actual communicative act, we are evaluating both the utterance as a social act

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

T : P1 : P2: P3:

276

F O R E I G N L A N G U A G E C LA S S R O O M D I S C O U R S E

(a) the utterance is consistent with/inconsistent with hearer expectation, (b) the utterance is consistent with/contrary to that which the hearer desires to hear-in the case of a responding communicative act this is equivalent to saying that the utterance is consistent with/contrary to the perlocutionary intent of the hearer's preceding communicative act, (c) the utterance is consistent with/inconsistent with the behaviour patterns implied by socially-sanctioned norms of authority, (d) the utterance is consistent/inconsistent with what the evaluator interprets as

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

and the speaker as a social member. 1 4 It follows that the teacher i n deciding which forms of discourse to teach and practise is determining in part which forms of social action the learner will be (in theory at least) equipped to perform. Further in evaluating learner discourse, the teacher may be implicitly evaluating the learner as a social member. What an individual says and does in an ongoing discourse is determined by his social competence, as much as by his 'linguistic' or 'communicative' competence. u By which social norms is the teacher to ' evaluate-indeed is the teacher justified in imposing social evaluations on the learner? Take the case of a specialist English course for doctors, and let us assume that the doctor-patient interview is an important element of the desired ter­ minal behaviour. It is a non-trivial question to ask how far in producing such a course the teacher or researcher seeks to make the learner a 'successful' doctor in terms of his interpersonal relations with his patients . If the learner is to practise in a country where a doctor's income is proportionate to the number of patients he can attract, we may be asking how far the teacher is involved in teaching the learner how to make his practice a profitable one. A more concrete case is provided by the following: imagine conducting an oral examination to test a student's 'proficiency' or 'communicative skills' in spoken English. The student however responds to every question by saying 'I don't know really' , ' It depends I suppose', or the like. How does the examiner assess this student? 1 6 These two illustrations are marked cases: I am suggesting that the same sort of issue is raised implicitly in for example the teaching of 'greetings' -often found early in communicative-oriented teaching courses. Note that what may be established in sociolinguistic research are probabilistic rules, based on a macro-survey, and such rules have therefore only a probabilistic predictive value. If I greet a group of colleagues, some will return the greeting, others will not. The returned greeting may be in the form of a ritualised expression in some instances, in others it might be best described as a grunt, in other in­ stances it might fairly be described as idiosyncratic, maybe involving the use of an esoteric literary reference, the use of a foreign language, and so on. It is clear that large-scale research into native-speaker usage does not in itself resolve any evaluative issues, as what might at best result is a probabilistic spectrum detailing who greets whom when and how. On what criteria does the teacher make use of such empirical data? (Unfortunately, the question as of now ought to be posed as how would the teacher use such empirical data if he had them). 1 7 When communicative aCtivity i s evaluated, the following criteria are amongst those which might be employed:

W I L L IS J . EDMON DSON

277

social norms of politeness (as formulated in the 'tact-principle' in Leech 1977, for example), (e) the utterance is consistent/inconsistent with the evaluator's beliefs as to the speaker's beliefs, state of mind, or feelings. This is the sincerity criterion once more, reformulated with respect to non-pedagogic discourse. Note that this criterion tends to be invoked to counter a non-positive evaluation made according to some other criterion ('Well at least he speaks his mind', 'You know where you are with him' and so on).

·

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

These criteria may converge, but need not, though it would seem likely that criteria (a) and (c) would most commonly lead to the same evaluation. Further, these criteria are clearly subjective, 11 and may be exercised both with respect to which communicative act is performed in a particular case, and with respect to the linguistic (and other) means through which it is performed (its 'modality' in the sense of House and Kasper 1978). On the first dimension we might evaluate a communication as being a 'turn-down' of a request, as opposed to an 'agree' for example, while on the second, we might evaluate the realization of a tum-down, as a choice from a set of possible realizations including for example no/sorry . . . II would normally, but . . . /can 't do I'm afraid/l'd love to I really would but . . . Iget stuffed/do it yourself and so on. The learner as a social member brings to the classroom his own social values. It is unlikely that these are a constant for a group of learners, and it is unlikely that for any individual learner these values remain a constant for other than a short period of time relative for example to the nine-year period of English study which is commonly undergone by a learner completing his secondary education in a German Gymnasium. 19 The teacher as a social member also brings his social values to the learning-teaching situation. The confrontation of different social values in a foreign language classroom is further complicated by the fact that there is some evidence that even for two language communities as closely related as are the English and the German, there may be different social norms with regard to communicative activity operant in the two language communities (House, 1979, House and Kasper, 1 978). The problems of evaluation are then considerable and complex. One major issue is how far the teacher is professionally obliged or expected to uphold the social values of the institution and of the society in which he teaches, and who is to decide exactly what those values are. This is a political issue. Again, how far is the learner to be expected to conform with social norms assumed to operate for the foreign language community, assuming these can be established? There is here a danger of the promulgation of national stereotypes, whereby, for example, English people not only like queuing and invariably eat porridge for breakfast (which we have known for some time), but are also invariably polite, and preface almost every remark with the ex­ pression oh (this we have learnt more recently). The notion that the learner in his communicative behaviour should conform to this stereotype would seem even more dangerous. That the learner's own real or assumed social values be taken as the evaluative criterion has been argued against in 2 above.

278

F O R E I G N L ANGUAGE C L ASSROOM D I SCOURSE

4. C O N C L U S I O N S Much o f what I have said has been negative i n tone: I have also raised some issues which I have not attempted to answer, nor in some cases do I feel competent to do so . 20 Two general conclusions I would draw concerning the discourse by means of which the foreign language is used as a means of communication in the classroom setting are as follows:

(a) A wide spectrum of communicative activities and of ways of performing them should be taught (not necessarily for active control) and practised in communicative language discourse, and this will largely be done within the restrictions im osed by the pedagogic setting via simulations of various kinds, role-play, pupil-pupil discourse, and other techniques. Thus the practice of target discourse will often be embedded within a pedagogic discourse. Any teacher evaluation is then addressed to the pupil as learner, and not to the pupil as social actant. (b) Side by side with this, the social values implicit in �poken discourse, and affecting issues of face, tact, social convention and social authority need to be brought out in the practised target discourse. Alternative realizations of com­ municative acts may be put forward and compared evaluatively . This reinforces Candlin's ( 1 975) point concerning a cognitive element in communicative language teaching, and links with suggestions concerning the usefulness of the analysis of register and the potential usefulness of translation made in Wid­ dowson 1 974, House 1977.

p

NOTES 1 On the text/discourse distinction made here, see Widdowson 1973, Sandulescu 1976, Edmondson 1 978a. 1 But see Gray 1977, Edmondson (forthcoming), and the papers by Sinclair and Widdowson in this volume, for some suggestions. 1 A discourse perspective is of course often implicit in such work. For example Candlin in discussing doctor-patient interviews talks of limiting the range of speech 'functions' , and of in­ dicating 'function patterns', noting that in 'oral discourse, speech functions co-occur and entail each other in predictable ways' (Candlin 1975:80). ' A shift from a speech act perspective to a discourse perspective is most readily achieved of course via the notion of speech act sequences, a sequence forming some higher order unit-an example of this approach is the work of Ehlich and Rehbein-e.g., Rehbein and Ehlich 1976, Rehbein 1977. ' In itself a 'notional' syllabus does not reflect what is specific to a discourse perspective on language use, as the notional units are dealt with in isolation one from another. Similarly, Leech and Svartvik (1975) is not in the terms being used here a communicative grammar . ' To document all these alleged tendencies fully would require a plethora of references likely to exceed this article in length. The literature cited elsewhere in this paper perhaps suffices. ' Students in fact sometimes assume that the data in Sinclair and Coulthard provide a model of 'real' FL teaching. The implicit reasoning would seem to be: 'We wish to •teach natural native-

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

What we require in order to follow such teaching procedures is more empirical findings from discourse analysis concerning the social values implicit in spoken discourse, such that these may be taught, and such that subjective, intuitive, or dogmatic evaluations of classroom discourse be replaced by empirically-based descriptions of what is involved in speaking a foreign language.

279

W I L L I S J . E D M O N DSON

speaker discourse in the FL classroom. Here we have documented natural native-speak er discourse in the FL classr oom. Therefore we may model our teaching on this data' . The reasoning is o f course false, but the unfortunate conclusion

seems

t o be implied i n the following, taken from a

brief review of Sinclair and Coulthard 1 975: 'FOI den deutschen Anglisten liegt ein weiterer, gar nicht beabsichtigter Anreiz vor: das Buch enthalt ca. SO Seiten gegliederten Textes von tat­ sachlichem

"class-room

discourse" ,

eine

Fundgrube

fiir die

Sprache

des

Lehrers

im

Englischunterricht. ' (GOtz, 1978).

See for example Edmondson,

1

9

1978b, 1 978c and the references therein.

The question-answer technique as a teaching strategy for exatnple does not necessarily lead the

learner to practise any skill other than that of answering such pedagogic 'questions'. Edmondson 1 979). I t i s difficult t o accept therefore conclusions such as the following:

.

(See e.g . ,

. .

'urn

die

Schuler auf das Fiihren von freien Gespra.chen besser vorzubereiten, miissen sie sprachlich reaktiv und initiativ tatig sein. Deshalb muB die Frage-Technik in situativen Obungen einen bedeutend grOBeren Raum einnehmen' (KOnig 1979, p. 1 9). 10

A certain atnount of simplification is inevitable here, but the position I have sketched is

exatnple, and in part underpins the stress on 'emancipatory' language teaching-see e.g., Maas 1 974. This does not imply of course that to argue against the validity of the sincerity criterion is necessarily to argue against the validity of the notion of emancipation.

1 1 The data is taken from Walmsley 1978. Similar data are cited in Krumm 1978, Edmondson

1978b. 12

With repair-work, the main point must be

how

it is initiated: I would guess that learners

initiate repair-work quite frequently-the trouble is tl!ey do this most commonly by maintaining silence, breaking eye-contact, and so on. '1

Jakobovits and Gordon 1974, p. 80: ' . . . we shall call "natural" conditions of acquisition

those situations in which the indiVIdual is exposed to social interactional settings that exclude the learning of language as one of its recogrused and legitimate functions-all other conditions of acquisition are artificial' . " In Labov's research into varieties of American English for exatnple (e.g., Labov 1 972a, pp. 201-240) it is clear that the question as to the status of a non-standard dialect or variety as a means of communication necessarily involves the question as to the status of users of that dialect or variety as social members. " I have tried elsewhere to sketch a distinction between 'communicative' and 'social' com­ petence (Edmondson 1978b and forthcoming). The distinction is not a totally sharp one, but communicative competence concerns the ability to encode and decode 'central' communicative acts, and an awareness of their potential sequencing in discourse. Social competence concerns the use to which these abilities are put, and may involve the use of 'supportive' communicative acts, and the use of conversational strategy. Some of us are 'better' at this than others.

16 This example is derived from an anecdote supplied by Werner Hullen: in the actual case

encountered by Hilllen he later discovered that the examinee was in fact a native-speaker of English. " There are clearly some issues concerning language use which cannot be resolved by surveys. Klein ( 1 977, 1 979) for exatnple has established that a majority of people when asked for directions give false information. To describe this as a norm for German speakers to be adopted by learners of German is clearly absurd: it is perhaps instructive however to ask why this conclusion is absurd. 11

For example, the 'natural' thing to say may also be the 'wrong' thing to say. An illustrative

anecdote occurs in Russell's

A utobiography,

Vol. II, p. 1 0 1 : note that Russell provides this

account as evidence that Wittgenstein 'was not always easy to fit into a social occasion ' . " With intensive courses, and with adult learners, the point here i s less valid.

10

There are of course several issues associated with the discussion of evaluation I have not gone

into here at all-one such is the notion of 'pragmatic error' in the study of interlanguages. e.g . , Kasper 1978, 1 979a.

See

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

strongly reflected in the work of Piepho (e.g., Piepho 1974, 1 979 centrally pp. 1 20- 1 23), for

280

F. O R E I G N L A N G UAGE C L A S S R O O M D I S C O U R S E

REFERENCES Allen , J. P . B . and Corder, S . P . 1 973- 1 975. The Edinburgh Course in Applied Linguistics. London: Oxford University Press. Volumes 1-3 . Anscombre, J .-C. and Ducrot, 0. 1976. 'L'argumentation dans la langue', Langages 42. Anscombre, J .-C. 1 977 'La problematique de l 'illocutoire derive' , Langage et Societe 2. Austin, J . L . 1962. How to do Things with Words. Oxford: Oxford University Press, 1962. Trad. 1 970 Quand dire, c'estfaire Paris: Le Seuil. Bally, C. 1943. ' Syntaxe de la modalite explicite' , CFS 3 3- 1 3 . Bausch, K . R . (ed .) 1979. Beitriige zur didaktischen Grammatik. Kronberg: Skriptor.

Beacco, J. C . and Dacrot, M . 1978. Pour lire les sciences sociales: une analyse de discours. Paris: BELC (multigr .) Benveniste, E. 1 966. Problemes de linguistique generale I. Paris: Gallimard, (1 974). Benveniste,

E.

1 974.

' Structure

des relations d 'auxiliariti:' ,

in Problemes de

linguistique generale II Paris: Gallimard. Brazil, D . and Coulthard, M. 1 977. Exchange Structure. University of Birmingham: Monographs on Discourse, No. 5. Brown, R . and Bellugi, U . 1964. 'Three Processes in the Child's Acquisition of Syntax'. Harvard Educ. Review 34. Brown, R. 1976. A First Language. Harmondsworth: Penguin Books. Brown, P. and Levinson, S. 1978. 'Universals in Language Usage: Politeness Phenomena' . in Goody, 1 978. Candlin, C. N. 1 975. 'Some Metalinguistic Problems in Communicative Language Learning' , in Corder and Rou1et 1975. Candlin, C . N. Leather, J. H. and Bruton, C. J. 1976. ' Doctors in Casualty: Applying Communicative Competence to Components of Specialist Course Design'. In IRAL 14. Chomsky, N . 1957 . Syntactic Structures. Mouton: The Hague. Colombier, P. et a/ 1 976. 'Niveau avance, travailleurs migrants: Ia recherche d'un emploi ' . Saint-Cloud: CREDIF. (multigr.) Cole, P . and Morgan, J. L . (eds.) 1975. Syntax and Semantics. Vol. 3 Speech Acts. New York: Academic Press. Corder, S. P. Books.

1967. Introducing Applied Linguistics. Harmondsworth: Penguin

Corder, S. P. and Roulet E. (eds.) 1 972. Theoretical Linguistic Models in Applied Linguistics. Brussels: AIMAV. Paris : Didier. Corder, S. P. and Roulet, E. (eds.) 1975. Some Implications of Linguistic Theory for Applied Linguistics. Brussels : AlMA V. Paris: Didier. Corder, S. P. 1 978. Learner Language and Teacher Talk. A udio V"ESUOI Language Journal, Vol. 1 6 No. 1 .

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Bazell, C. E . , Catford, J. C., Halliday, M. A. K. and Robins, R. H. (eds .) 1 966. In Memory of J. R. Firth. London: Longmans .

W I L L I S J . E D MO NDSO N

281

Coste, D., Courtillon, J. , Ferenczi , F., Martins-Baltar, M., Papo, E. et Roulet, E. 1976. Un niveau-seuil. Strasbourg: Conseil de !'Europe. Coulthard, M. 1 977. An Introduction to Discourse Analysis. London: Longman. Culioli, A. 1979. 'Why teach how to learn what is best learnt untaught? ' . Cahiers Charles V No. 1 Universite de Paris VII. Darot, M . 1975. 'Discours mathematique et discours didactique' . Paris: BELC (multigr.) Descamps, J. L. 1979. 'Contribution a !'analyse des discours fonctionnels (pedagogie des langues de specialite et lexicographie contextuelle.)' Saint-Ooud: CREDIF (multigr.) Descamps, J. L. et a/ 1 978. Dictionnaire contextuel de francais pour Ia geologie. Paris: Didier,

Ducrot, 0 . 1972. Dire et ne pas dire. principes de semantique linguistique. Paris: Hermann. Ducrot, 0. 1974. Lapreuve et le dire. Paris: Reperes-Mame. Ducrot , 0. 1 977. ' Illocutoire et performatif, Linguistique et Semiologie Ducrot, 0. 1979. 'Les lois du discours' , Languefrancaise, 42. Duncan, S. 1972. 'Some signals and rules for taking speaking turns in conversation' in Journal of Personality and Social Psychology 23.2. Duncan, S . 1973. 'Toward a grammar for dyadic conversations ' . Semiotica 9 Edmondson, W. J. 1978a. 'A note on " Pragmatic Connectives' " , in lnterlanguage

Studies Bulletin. Edmondson, W. J . 1 978b. 'Worlds within Worlds-Problems in the Description of Teacher-Learner Interaction ' , paper read at the 5th AILA Congress, Montreal 1978. (To appear in Congress Proceedings.) Edmondson, W. J. 1 978c. ' Context and Stress-the "Old" and the "New" in Foreign Language Teaching' ' , in Anglistik und Englischunterricht 6. Edmondson, W. J . 1979. 'Funktionen von Fragen im Fremdsprachenunterricht', in Heuer et al 1919. Edmondson, W. J . (forthcoming) Spoken Discourse: a Mode/for Analysis. Edmondson, W. J. and House, J. 1979. 'Konzeption einer didaktischen Inter­ aktionsgrammatik ' , in Bausch 1979. Faerch , C. 1979. ' Describing Interlanguage through Interaction Problems of Systematicity and Permeability. ' •Working Papers on Bilingualism 19. Firth, J. R. 1957. Papers in Linguistics 1934-1951. Oxford: Oxford University Press. Fraser, B. 1 975. 'Hedged Performatives ' , in Cole and Morgan 1 975. Galisson, R. et Coste, D. Hachette.

(eds.)

1976. Dictionnaire de didactique des langues. Paris:

Garfinkel, H. and Sacks, H. 1970. 'On formal structures of practical actions' in McKinney and Tiryakian 1 970.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Diller, A.-M. 1977. ' Le conditionnel, marqueur de derivation illocutoire' , Semantikos Z/ 1 .

282

F O R E I G N LANGUAGE C LA S SROOM D I S C O U R S E

Gauvenet, H . 1 976 . Pedagogie du discours rapporte, Paris: Didier, Collection VIC. Giles, H . and Smith, P . 1 979. 'Accommodation Theory' in Giles and St. Clair 1 979. Giles, H. and St. Clair, R. 1 979. Language and Social Psychology. Oxford: Blackwell. Goffman, E. 1 97 1 . Relations in Public: Harmondswonh: Penguin Books.

microstudies of the public order.

Goffman, E. 1973. La mise en scene de Ia vie quotidienne. Paris: Minuit. Goody, E . N. (ed.) 1978. Questions and Politeness. Cambridge: Cambridge University Press. Gotz, D. 1978. Review of Sinclair and Coulthard 1 975 in Die Neueren Sprachen. Gray, B. 1917. The Grammatical Foundations of Rhetoric. The Hague: Mouton.

Gremmo , M .-J., Holec, H. and Riley, P. 1977. ' Interactional structure: the role of role' in Melanges Pedagogiques. CRAPEL: Universite de Nancy I I . Gremmo, M . J . Holec, H . and Riley, P . 1978. 'Taking the initiative: some pedagogical applications of discourse analysis' in Melanges Pedagogiques CRAPEL: Universite de Nancy I I . Grice, H . P. 1 975. ' Logic and Conversation' in Cole and Morgan 1975. Gutknecht, C . 1977. (Hrsg) Grund begriffe und Hauptstromungen der Linguistik. Hamburg: Hoffman und Campe Verlag . Halliday, M . A. K. 1 96 1 . 'Categories of the Theory of Grammar' , Word, 1 1 . Halliday, M . A. K . 1 968. 'Notes o n Transitivity and Theme i n English: Part 3 ' Journal of Linguistics Vol. 4 No. 2 . Halliday, M . A. K . 1978. ' Language Structure and Language Function' i n Lyons 1 970. Halliday, M . A. K. 1973 . Explorations in the Functions of Language. London: Edward Arnold. Harder, P. 1 979. 'Tekstpragmatik' Nydanske Studier 1 0- 1 1 Copenhagen 1979. Hatch, E. M. 1978. 'Discourse Analysis' in Hatch, 1978. Hatch, E . M . (ed.) 1978. Second Language Acquisition. Rowley, Mass . : Newbury House. Heringer, J. T. 1 972. 'Some Grammatical Correlates of Felicity Conditions and Presuppositions ' . Working Papers in Linguistics I I . Heritage, J . C . and Watson, D . R . 1979. 'Formulating as conversational objects' in Psathas 1979. Heuer, H . et: al (eds .) 1979. Dortmunder Diskussionen zur Fremdsprachendidaktik. Dortmund: Lensing. Holec, H . 1 975. ' L'approche macro-linguistique du fonctionnement des langues et ses implications pi:dagogiques: role du visuel' in Melanges Pedagogiques, CRAPEL: Universite de Nancy I I . Holec, H . , Gremmo, M . - J . and Riley, P . (forthcoming) 'Prolegomenes a une description de Ia structure des echanges communicatifs directs ' . Paper read at the

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Greene, J. 1 972. Psycholinguistics: Chomsky and Psychology. Harmondsworth: Penguin Books.

283

W I L L I S J . E D M O N DSON

5th International Congress of Applied Linguistics, Montreal, Canada. To appear in IRAL. House, J. 1 977. A Modelfor Translation Quality Assessment. Ttibingen: Narr. House, J. 1 979. ' lnteraktionsnormen in deutschen und englischen Alltagsdialogen' , in: Linguistische Berichte 59. Householder, F. W. V. (ed.) 1 972. Syntactic theory I, Penguin Modern Linguistics Readings. Harmondsworth: Penguin Books. Hutchinson, T. 1 978. 'An analysis of the effect on discourse structure of a visual display' . In Practical Papers in English Language Education : University of Lancaster. Jacobs, R. A. and Rosenbaum, P. S. (eds .) 1 970. Readings in English Trans­ formational Grammar. Waltham, Mass: Blaisdell.

Jupp, T. C . , Hodlin, S. Heddesheimer, C. et Lagarde, J. P. 1 978. Apprentissage linguistique et communication: methodologie pour un enseignement fonctionnel. (Traduit et adapte de: Jupp, T. C. et Hodlin, S. 1 975. Industrial English. Paris: C. L. E. International.) Kasper, G . 1 978. 'Pragmatische Defizite im Englischen deutscher Lerner'. In Kongressberichte der 8. Jahrestagung der Gesellschaft fur Angewandte Linguistik, 1 1 . Stuttgart: Hochschul-Verlag. Kasper, G. 1 979. 'Communication Strategies: Modality Reduction' . Paper presented at the 17th Interaction Conference on Polish-English Contrastive Linguistics, Boszkowo. Kasper, G. 1 979. 'Errors in Speech Act Realization and Use of Gambits' . In Canadian Modern Language Review 35. Keenan, E. L. (ed.) 1 975. Formal Semantics of Natural Language. Cambridge: Cambridge University Press. Kendon, A. 1 967. 'Some functions of gaze direction in social interaction ' . Acta Psychologica, 24. Klein, W. 1977. ' Wegausklinfte' . Max Planck-lnstitut Nijmegen. (To appear in Gutschriftfur Literaturwissenschaft und Linguistik.) Klein, W. 1979. Local Deixis in Route Directions. L.A.U.T. paper no. 41, Trier. Konig, E. 1 979. Vorchlage zur situativen Unterricht gestaltung. sprachenunterricht 23 .

In Fremd­

Krumm, H .-J. 1978. 'Lehrerverhalten im Hinblick auf Lernerverhalten: Entwicklungsgemll.sser Fremdsprachenunterricht?'. In Kongressberichte der 8 . Jahrestagung der Gesellschaft fUr Angewandte Linguistik 1 pp. 29-42. Stuttgart: Hochschul-Verlag. Labov, W. 1 972. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, W. 1972. Language in the Inner City. Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press .

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Jakobson, R. 1963. Essais de linguistique generate. Paris: Editions de Minuit. Jakobovits, L. A. and Gordon, B. 1 974. The Context of Foreign Language Teaching. Rowley, Mass: Newbury House.

284

F O R E I G N L A N G U A G E C L AS S ROOM DISCOURSE

Labov, W . and Fanshel, D . 1977 . Therapeutic Discourse: Psychology as Conversation. New York: Academic Press. Lakoff, G. 1 97 1 . 'On Generative Semantics' . In Steinberg and Jakobovits 197 1 . Lakoff, G. 1 973. ' Pragmatics in Natural Logic' . In Keenan 1975. Lakoff,

G.

and Thompson,

H.

(eds.)

1 975. Introducing Cognitive Grammar.

Proceedings of the First Annual Meeting of the Berkely Linguistic Society, Berkeley

1 975. Lakoff, R. 1 977. 'What You Can Do with Words: Politeness, Pragmatics and Per­ formatives ' . In Rogers et a1 1911. Leech, G. N . 1977. Language and Tact. L.A. U.T. Paper No. 46, Trier. Leech, G. N. and Svartvik, J. 1975 . A Communicative Grammar of English. London: Lehmann, D . , Catalan, R . , Moirand, S. and Marlet, F. 1979. Lire en francais les sciences economiques et sociales. Paris: Didier. Lyons, J. 1 966. ' Firth 's theory of meaning' . In Bazell et al 1966. Lyons, J. (ed .) 1 979. New Horizons in Linguistics. Harmondsworth: Penguin Books McKinney, J. C. and Tiryakian, E. A. Appleton-Century-Croft.

(eds.) 1970. Theoretical Sociology.

New York:

McRae, A. 1 977. 'Comprehension: the psycholinguistic view' . In The Teaching of Comprehension. London: The British Council. U. 1 974. Argumente fiir die Emanzipation von Sprachstudium Sprachunte"icht. Frankfurt: M . Athenaum Fischer.

Maas,

und

Maingueneau, D. 1976. Initiation aux methodes de /'analyse du discours. Paris: Hachette Universite.

De /'enonce a l'enonciation: une approche des jonctions intonatives. Paris: Didier, Collection VIC.

Martins-Baltar, M. 1977.

Martins-Baltar, M. et al 1979. L 'ecrit les ecrits: problemes d'analyse et consideration

didactiques. Strasbourg: Conseil de !'Europe. Moirand, S. 1977. 'Analyse de textes ecrits et apprentissage "grammatical" ' . Etudes

de linguistique appliquee, no. 25. Moirand, S . 1 978. 'Les textes aussi sont d es images ' . Le Francais dans Le Monde, no.

137. Morgan, J . L. 1 978. ' Two Types of Convention in Indirect Speech Acts' . I n Cole and Morgan 1 978. Muller B. D . (ed.) 1 980 Konjrontative Semantik. Weil der Stadt: Lexica-Verlag. Murphy, D . and Candlin C. N. 1976. Engineering Discourse and Listening Com­ prehension ; K.A.A.U. Project in Listening Comprehension. First annual report. Department of Linguistics and Modem English Language: University of Lancaster. Nickel, G. (ed.) 1976. Proceedings of the Fourth International Congress of Applied Linguistics, Volumes 1-3. Stuttgart: Hochschul-Verlag. Oldfield, C. and Marshall , J. C. (eds.) 1968. Language. Harmondsworth: Penguin Books.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Longman.

W I L L I S J . EDMONDSON

285

Olson, D. R. and Filby, N. 1 972. 'On the Comprehension of Active and Passive Sentences' in Cognitive Psychology 3 . Parret, H . (ed.) 1974. Discussing Language with Noam Chomsky. Mouton: The Hague. Perret, D. 1 974, 'Les verbes "pouvoir" et "vouloir" dans les enonces de propositions ' . Langue Francoise 2 1 . Peytard, J . et Charolles, M . (dir.) 1 978. ' Enseignement d u recit et coherence du texte' .

Langue Francoise, 3 8 . Piepho, H.-E. 1 974. Kommunikative Kompetenz als Ubergeordnetes Lernziel im Englischunterricht. Dornburg-Frickhofen: Frankonius. Piepho, H.-E. 1 979. Kommunikative Didaktik des Englischunterrichts. Limburg: Frankonius.

Psathas , G. (ed.) 1979. Everyday Languages: Studies in Ethnomethology. New York: Irvington. Rehbein, J . and Ehlich, K. 1 976. 'On Effective Reasoning' . In Nickel 1976, Volume 1 . Rehbein, J . 1 977. Komplexes Handeln. Stuttgart : Metzler. Riley, P. 1976. 'Discursive and communicative functions of non-verbal com­ munication ' . In Melanges Pedagogiques, CRAPEL: Universite de Nancy I I . Riley, P . 1 979. 'Towards a contrastive pragmalinguistics ' . Papers and Studies in Contrastive Linguistics, Vol. 10 Poman, Poland, Arlington, Virginia. Rogers, A . , Wall, B. and Murphy, J. (eds.) 1977. Proceedings of the Texas Conference on Performatives, Presuppositions and lmplicatures. Washington, D.C.: Centre for Applied Linguistics. Ross, J. R. 1970. 'On declarative sentences' in Jacobs and Rosenbaum 1 970. Roulet, E. 1 976. 'L'apport des sciences du langage a la diversification des methodes d'enseignement des langues secondes en fonction des caracteristiques des publics vises' . Etudes de linguistique appliquee, 2 1 . Roulet, E . 1 977. Un niveau-seuii-Presentation et guide d'emploi. Strasbourg: Conseil de l'Europe. Roulet, E. 1 977. 'Etude des realisations directes et indirectes de l'acte d'offre en francais parle'. Studi Italiani di Linguistica Teorica ed Applicata 6. Roulet, E. 1978. 'Essai de classement syntaxique et semantique des verbes poten­ tiellement performatifs du francais' . Cahiers de Linguistique 8. Roulet, E. 1980. 'Modalite et illocution-pouvoir et devoir dans les actes de permission et de requete' . Communications 32, 2 1 6-239. Roulet, E. and Holec, H. (eds .) 1976. L'enseignement de la competence de com­ munication en langues secondes. Bulletin CILA no special 24. Roussel, F. 1 974. 'The Modulation of Discursive Functions. Melanges CRAPEL 41-52. Sandulescu , C. G. 1976. 'Theory and Practice in Analysing Discourse'. In Nickel 1 976, Volume 1 .

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Portine, H. 1978. 'Apprendre a argumenter. Analyse de discours et didactique des langues ' . Paris: BELC (multigr .)

286

FOREIGN LANGUAGE CLASSROOM DISCOURSE

Scollon, R . T . 1 974. One Child's Language from One to Two. Ph.D. Dissertation, University of Hawaii. Searle, J. R . 1 975 . ' Indirect Speech Acts'. In Cole and Morgan 1 975. Selinker, L . 1 972. 'Interlanguage' IRAL 10.3. Sinclair, J . McH and Coulthard, R. M . Towards an Analysis of Discourse. Oxford: Oxford University Press. Slobin, D. I . 1 975. 'The more it changes . . . on understanding language by watching it move through time'. Papers and Reports on Child Language Development No. 10. Sornig, K . 1977. 'Disagreement and Contradiction as co·mmu,.nicative Acts' . In

Journal of Pragmatics.

Stevens, P. 1 974. Patterns in Nature. Harmondsworth: Penguin Books. Tarone, E . , Cohen, A. and Dumas, G. 1 976. 'A Closer Look at some Interlanguage Terminology: A Framework for Communication Strategies ' . Working Papers in Bilingualism 9. von Unwerth , H. J. and Buschmann, U. 1 980. 'Konfliktive Sprechakte und Lehrwerk­ text' in Muller 1 980. 1977. Text and Context: Explorations in the Semantics and Pragmatics ofDiscourse. London: Longman.

Van Dijk , T. A.

Van Ek, J. A. 1 975 . The Threshold Level. Strasbourg: Conseil de !'Europe. Varadi, T. 1 97 3 . 'Strategies of target language learner communication: message­ adjustment ' . Paper presented at the VI Conference of the Romanian Linguistics Project in Timisoara, 28/29 May. Vignaux, G. 1 973 . ' Le discours argumente ecrit'. Communications, no. 20. Vignaux, G. 1976. L 'argumentation. Essai d 'une logique discursive. Geneve: Droz. Vigner, G. 1 97 5. Ecrire pour convaincre. Paris: Hachette. Walmsley, J. B. 1 978. 'Cloud-cuckoo-land' or: Feedback as the central component in foreign language teaching ' . Paper read at the 5th AILA Congress, Montreal 1978. Wason, P . C. 1 965 . 'The contexts of plausible denial ' . In Journal of Verbal Behaviour and Verbal Learning Vol. 4 reprinted in Oldfield and Marshall 1 968. Widdowson, H . G. 1 973. 'Directions in the Teaching of Discourse ' . In Corder and Roulet 1 97 3 and in Widdowson 1 979a. Widdowson , H. G. 1974. 'The Deep Structure of Discourse and the use of Trans­ lation'. In Corder and Roulet 1974, and in Widdowson 1 979a. Widdowson, H. G. 1977. 'Approaches to Discourse' . In Gutknecht 1977, and in Widdowson 1 979a. Widdowson, H. G. 1 978. Teaching Language as Communication. Oxford: Oxford University Press. Widdowson , H . G. 1 978. 'The realization of rules in written discourse' . Paper read at the British Council/Goethe Institut Seminar. Paris October 1978 mimeo.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011

Steinberg, D . D . and Jakobovits, L. A. (eds.) 1 97 1 . Semantics: An Interdisciplinary Reader in Philosophy, Linguistics and Psychology. Cambridge: Cambridge University Press.

WILLIS J . EDMONDSON Widdowson, H. G.

287

1 979a. Explorations in Applied Linguistics Oxford: Oxford

University Press. Widdowson, H. G. 1 979b. 'New starts and different kinds of failure ' . Paper read at the CCTE Conference Ottawa May 1 979 mimeo. Wilkins, D. A. 1 976. Notional Syllabuses. Oxford: Oxford University Press. Wunderlich, D. 1 977. Studien zur Sprechakttheorie. Frankfurt: Surh kamp.

Downloaded from applij.oxfordjournals.org by guest on January 1, 2011