Beruflich Dokumente
Kultur Dokumente
AUTHOR ""
SUBJECT ""
KEYWORDS ""
WIDTH "150"
VOFFSET "4">
Edited by
Susan L. Eerdmans
University of Bologna
Carlo L. Prevignano
University of Bologna
Paul J. Thibault
University of Venice and Lingnan University, Hong Kong
Preface vii
Chapter 1
Presenting John J. Gumperz 1
Aldo di Luzio
Chapter 2
A discussion with John J. Gumperz 7
Carlo L. Prevignano and Aldo di Luzio
Chapter 3
Contextualizing “contextualization cues” 31
Stephen C. Levinson
Chapter 4
Contextualization and social meaning-making practices 41
Paul J. Thibault
Chapter 5
On Gumperz and the minims of interaction 63
Carlo L. Prevignano
Chapter 6
A commentary on a discussion with John J. Gumperz 79
Afzal Ballim
Chapter 7
A review of John J. Gumperz’s current contributions to
Interactional Sociolinguistics 85
Susan L. Eerdmans
Chapter 8
Response essay 105
John J. Gumperz
vi Contents
Chapter 9
Body dynamics, social meaning-making, and scale heterogeneity:
Re-considering contextualization cues and language as mixed-mode
semiosis 127
Paul J. Thibault
Chapter 10
Continuing the discussion with John J. Gumperz 149
Carlo L. Prevignano and Paul J. Thibault
Index 165
Contents vii
Preface
Luzio is the center piece of this volume. It provides a forum in which Gumperz
elaborates on and explains many of the central tenets of his approach. In
particular, the discussion examines the ethnographic basis of Gumperz’s re-
search in sociolinguistics, the need to distinguish linguistic forms from the
communicative practices in which these are embedded and in which they have
their meaning. This is a constant theme throughout this discussion and it is
refracted through an interesting series of reXections on intercultural commu-
nication, social action, communicative conventions, and contextualization.
Levinson’s essay draws on the author’s early association with Gumperz as
a Ph.D. student in Berkeley. Levinson develops an internalist approach to
contextualization cues, arguing that utterances specify their own contexts of
interpretation in often very implicit ways, through the use of a range of verbal
and non-verbal resources. Context in this reading is not something which is
externally imposed on utterances. Levinson’s essay shows how the explicit
propositional meaning of utterances is often signiWcantly modiWed by their
cross-modal connections with non-verbal and more implicit, topological mo-
dalities of semiosis. This further suggests that many of the implicit aspects of
human communication cut across apparent surface distinctions between
semiotic modalities, and that this may lead one to make overhasty and
superWcial distinctions between various meaning-making resources.
Thibault’s essay is complementary to Levinson’s in this sense. Thibault
argues for an approach which suggests a way of cutting across these apparent
‘surface’ distinctions in an attempt to develop a conceptually uniWed ap-
proach to the question of communicative — meaning-making — practices,
irrespective of the semiotic resources that are deployed. To achieve this, he
argues for a three-way distinction between indexical, intertextual, and meta-
textual meaning-making practices. This distinction is proposed as a means
for analyzing the ways in which agents access and co-ordinate their deploy-
ments of semiotic resources in the activity structure-types and discourse
genres of a given culture.
Prevignano’s contribution sheds light on a diVerent aspect of this problem
with reference to what he deWnes as the “minims of interaction”. In this, he
contrasts Gumperz with Grice, Leech, and other pragmatists, whom Pre-
vignano characterizes as “maximists”. In particular, Prevignano examines
some “interaction minims”, which participants deploy in order to signal to
each other what they are doing in the course of the interaction. Importantly,
such interaction minims are embedded in activity structures and entail inter-
Preface ix
pretive principles that are historical rather than universal in character. The
analyst’s interpretation, insofar as it is a hermeneutical reconstruction of the
participants’ interpretation, is historical in two fundamental senses. First,
participants are embedded in a history of their applications and interpre-
tations of each others’ interactional minims. Secondly, the analyst, too, is
embedded in a history of ethnographic and analytical reconstruction and
understanding. This history itself acts on and aVects our understanding as
both participants in and analysts of the object of study — human interaction.
Ballim approaches the same problem of interpreting utterances in con-
text, though from a diVerent perspective — that of computer-mediated com-
munication from the point of view of applied computational linguistics. A
central problem for Ballim is how utterances are interpreted in the light of the
knowledge, beliefs, desires, and intentions of discourse participants. Conse-
quently, he is interested in how discourse participants create models of each
other in the course of interaction. Ballim indicates that the overall system
forms top-down constraints on the ways in which participants make infer-
ences and reason about each other in the course of interaction. His discussion
is informed by his research on the use of language as a communicative me-
dium between humans and computers.
Eerdmans provides a critical account of two exemplar case studies con-
ducted by Gumperz in the area of intercultural communication. She shows
how participants involved in inter-ethnic communication interpret and nego-
tiate their interlocutors’ contributions by embedding what they see and hear
in interpretative frames which may conXict with or misconstrue those of the
interlocutor. Interactional sociolinguistics, with its hermeneutic basis, can
provide a powerful tool for explaining such failures on the basis of its ex-
panded understanding of the interaction between the internal dynamics of
agents — their structures of understanding and inferencing — and the larger-
scale dynamics of the unfolding activity-structure, along with its relations to
the wider social and cultural contexts at play in the interaction. In this way, we
begin to see some meaning-based answers as to why interactants act and
behave in the ways they do, and how misunderstanding and communication
failure occur. Eerdmans concludes her discussion with some comments on the
usefulness of interactional sociolinguistics for both teachers and learners of a
second language. In particular, she draws attention to the practical and theo-
retical insights which a critical awareness of the impact of diVerent cultural
backgrounds can have on teaching and learning contexts.
x Preface
In his Response essay, Gumperz ranges over his career trajectory as socio-
linguist before clarifying and expanding on a number of points raised in the
individual contributions as he responds to these. In particular, he makes
further pertinent observations on how his current thinking about language
and interaction builds on his earlier ethnographic Weld studies. In so doing, he
also clariWes the ways in which linguistic and cultural diversity and sociocul-
tural boundaries are represented in and shape linguistic interaction.
In a second essay, Thibault gives voice to some further developments of
the theoretical issues raised by the individual contributors. In particular,
Thibault returns to the notion of action and interaction as a unifying principle
for the analysis of the multimodal co-deployment of semiotic resources. He
argues for the embeddedness of action in higher-scalar ecosocial environ-
ments which interact with the lower-scalar embodied dynamics of the agents
who participate in discursive interaction. He also explores the relationship
between the notions of fully-coded message content and contextualization
cues, as discussed by both Gumperz and Levinson. In particular, he suggests
how these notions can be brought together in a more theoretically-uniWed
framework by considering the typological and topological dimensions of
language as a form of mixed-mode semiosis.
A Wnal discussion between Gumperz, Prevignano and Thibault rounds oV
the discussions which have taken place throughout this volume. In this way,
Professor Gumperz provides further clariWcation of a number of key issues in
interactional sociolinguistics, the role of inferential processes in interpreters’
understandings of each other’s meanings in interactional events, and the
importance of these insights for understanding the contribution of cross-
cultural factors in communicatively-diverse environments. Gumperz also
provides further comments on the centrality of ideological processes in hu-
man interaction, as well as sharing with his readers his views on a number of
other approaches with similar aims to interactional sociolinguistics. A short
bio-bibliographical note detailing the most salient moments in Professor
Gumperz’s career concludes the volume.
Susan L. Eerdmans
Carlo L. Prevignano
Paul J. Thibault
xii Contents
Presenting John J. Gumperz 1
Chapter 1
Aldo di Luzio
nity did not reXect either caste diVerentiation or other social factors, such as,
for example, income, educational level, or social status. On the contrary, and
as a decisive factor for the distribution of the linguistic variables, norms were
operating which governed and constrained the social quality of interpersonal
communication. In his analysis of code-switching between Hindi and Punjabi
in Delhi, he made it clear that change of code represents a style which depends
on social norms and may express solidarity and group membership. Accentu-
ating the fact that code-switching is not only a grammatical but also and
especially a discursive process, he showed that this could not be analyzed with
traditional linguistic methods. His analysis of the mutual adaptation of Kan-
nada (a Dravidian language), Marathi and Urdu (two Indo-Aryan languages)
in interactions among their speakers in a multilingual community in the
region of Mahrastra represents a milestone for the analysis of creolization and
language convergence studies. In his analysis, he made clear that under the
inXuence of very frequent code-switching and consequent simpliWcation and
reciprocal assimilation, especially at the syntactic level, the local dialects of the
three typologically very diVerent languages become creole-like varieties of the
same language. They present a similar linear structure and can be reciprocally
translated word for word. What in earlier structuralist and psycholinguistic
approaches was analyzed as interference is now interpreted by Gumperz as
socially and communicatively aVected convergence.
Some very important results of Gumperz’s earlier research are that (i)
diVusion and limitation of linguistic variables are determined and displayed
in interaction; (ii) speakers’ perceptions or deWnitions of language equiva-
lence or diversity do not depend on genetic aYliation; (iii) speech commu-
nities are not linguistically homogeneous; (iv) linguistic variation and
alternation are not random or arbitrary, but communicatively functional and
meaningful. These earlier achievements already evidenced new directions in
sociolinguistics. Gumperz showed how multilingualism or multidialectalism,
as well as style variation, with their dialectical and dialogical aspects, represent
a prototypical paradigm or condition of human communication.
Gumperz’s conception, which strongly contrasts with the traditional and
structuralist one of idealized and homogeneous languages and speech com-
munities, is strictly bound to his new concepts of speech community, lan-
guage repertoire and sociolinguistic co-occurrence rules. Members of the
same speech community do not necessarily speak the same language variety.
The only requirement is that the speakers have in common a linguistic reper-
Presenting John J. Gumperz 3
toire, that is, the whole of the linguistic resources and communicative strate-
gies at their disposal and the rules for using them in communication. These
are conceived as co-occurrence rules or implications, which determine at the
paradigmatic and the syntagmatic levels the choice of units from the reper-
toire in building utterances; they also deWne the stylistic proWle of the discur-
sive units according to the communicative goals of the interaction. In this way,
the apories derived from the dichotomy between langue and parole, compe-
tence and performance, are resolved.
For Gumperz as well as for Hymes, with whom he collaborated in the
1960s in order to deWne some aspects of his approach to the ethnography of
communication, linguistic activity is always situated in a dialogical or interac-
tional context and is revealed within the ethnographic context of the speech
community of which the interactants are members. However, Gumperz tends
to consider the communicative units in which communication is enacted as
types of speech activities or communicative genres rather than speech events.
The interactional sociolinguistics which Gumperz developed in the 1970s
implies a new hermeneutical orientation. It is critically opposed to correla-
tional methods because these are external to discourse and cannot explain the
relation between linguistic facts and social structure. Interactional socio-
linguistics may be seen as a sociolinguistic analysis of dialogue and of the
social meanings displayed in the dialogical interaction against the background
of the ethnography of the sociocultural context of the interactants. Regulari-
ties of linguistic interactions are no longer considered as reXecting inde-
pendent social norms; on the contrary, these norms are considered as
interactively (re)produced in the communicative situation. The task of the
interactional sociolinguistic analysis of determinate communicative situa-
tions is to reconstruct the interactive methods by which speakers use discur-
sive means from their sociolinguistic knowledge (for example, language
choice, code-switching and stylistic variation according to determined co-
occurrence rules) for the production and interpretation of the social meaning
intended. An illuminating example of an interactional sociolinguistic recon-
struction of how, in a determinate sociocultural situation, discourse is struc-
tured and social meaning is produced, is already oVered by Gumperz’s study
of code-switching in the small speech community of Hemnesberget in north-
ern Norway.
A very important element of interactional sociolinguistics is a generalized
theory of contextualization and interpretation in communicative interactions.
4 Language and interaction
Chapter 2
Carlo L. Prevignano: I’d like to begin by asking you about your current
research projects.
John J. Gumperz: My interests are both theoretical and applied. Currently I’m
engaged in empirical research on classroom interaction with eight- and nine-
year-old Spanish/English bilinguals. I use methods of conversational and
discourse analysis to formulate hypotheses about the indirect reasoning pro-
cesses students employ in problem-solving. These processes are for the most
part indirect in that they rely on implicit, taken-for-granted presuppositions
to convey information that adults would expect to be overtly lexicalized. By
formulating explicit, veriWable, analytical assumptions about what these pro-
cesses are, and about how children of that age rely on them and talk about
them with their peers, we hope to help teachers gain better insights into their
students’ ways of dealing with the learning task. Ultimately, we intend to make
training Wlms illustrating such reasoning processes, using audiovisual tech-
niques similar to those employed in the B.B.C. Crosstalk Wlms (Gumperz et al.
1979) to reveal the possible linguistic causes of misunderstandings in intercul-
tural communication.
As with my earlier research, such empirical studies are directly related to
my theoretical interest in showing how detailed analyses of communicative
practices can illuminate basic issues in social theory. I’m about to begin
working on a book which I hope will bring out the import of sociolin-
guistically-oriented pragmatic analysis for our understanding of social pro-
cess. Theorists of many persuasions argue that the social environments in
which we live and act are dialogically constituted. The main question I pose is:
how does verbal communication aVect such dialogic processes and to what
extent do these processes depend on shared linguistic knowledge?
My perspective on verbal communication is grounded in earlier studies
8 Language and interaction
ments. But since there was no one else with him on the program, I inferred
that by the way he contextualized his talk, he was indirectly — without putting
it ‘on record’ — implicating the second part, “I hope you have enjoyed listen-
ing to me” (see Gumperz 1996). A second, somewhat more complex example,
comes from my analysis of the cross-examination transcript of the victim in a
rape trial. Counsel: “You knew at the time that the defendant was interested in
you, didn’t you?”. Victim: “He asked me how I’d been... just stuV like that” (see
Gumperz 1995a).
In both cases, I had to search my memory of past communicative experi-
ence to construct a likely scenario or narrative plot that might suggest pos-
sible interpretations. My initial hypothesis in example one conXicted with
what I knew about the radio program, and this triggered a search for a
diVerent, more plausible, scenario. In the second case, I relied on what I
knew about cross-examinations as adversarial proceedings, where the attor-
ney attempts to expose weaknesses in the defendant’s testimony. But while
these general facts tell us something about participants’ motives in their
choice of verbal strategy, we need to turn to what they actually said to under-
stand what they intended to convey. By the words he chose, and by the way
he contextualized his talk, the attorney raised the possibility that defendant
and victim had had a prior relationship. The victim’s move, on the other
hand, positioned as it is immediately after the attorney’s question, implicitly
argues for a diVerent scenario, one where the two were merely casual ac-
quaintances. In this way, she sought to deny and in a sense ward oV the
questioner’s potential attack on her testimony.
I use the term “activity type” or “activity” to refer to the above type of
constructs or “envisionments”, to borrow Fillmore’s term. My claim is that all
interpretation rests on such constructs. Activities are an aspect of GoVmanian
frames and are subject to constant change in the course of the exchange. That
is, they do not apply to events as wholes, they apply to each component move.
I argue that ultimately all interpretation at the level of discursive practice relies
on these constructs.
This view of understanding has some similarity to Fillmore’s notion of
“scene” that he discusses in his work on the semantics of understanding. But
whereas Fillmore is concerned with physical settings, I take more of a social
perspective. I see activities as evoking the actions of actors engaged in strategi-
cally formulating and positioning their moves in order to accomplish com-
municative ends in real-life encounters. In so doing, they rely on their
14 Language and interaction
J.J.G.: Not exactly. Our assumptions about relevant activities determine our
view of how the interaction coheres.
A.d.L.: By “type of activity”, do you mean “genre”?
J.J.G.: Not quite. I use the term “genre” for another level of pragmatic analysis,
the ideological level. For me genre is not an analytical category, because, as I
understand it, genre, at least in the Bakhtinian sense of the term, is better
treated as an analyst’s, or for that matter also a lay-person’s, concept for
referring to or labeling texts or speech exchanges.
A.d.L.: A mode of speaking about something.
J.J.G.: It’s not just a “mode” of speaking. When we speak of moral discourse as
a speech genre, for example, we are making an ideologically-charged
metapragmatic assessment. That is, we are engaging in a form of talk about
talk. As I said above, I use the notion of activity to refer to conversationalists’
and analysts’ hypotheses about what cognitive processes are involved in un-
derstanding, hypotheses that can then be validated by methods such as those I
have outlined.
C.L.P.: Could you add some remarks about your idea of linguistic convention
and typiWcation?
J.J.G.: Yes, I don’t want to use the term “convention” in the linguist’s sense of
grammatical convention. I use convention as a general term in the lay sense, as
the outcome of a process of typiWcation.
A.d.L.: In the Schutzian sense?
J.J.G.: In the Schutzian sense. For example, Liliana Ionescu-Ruxandoiu this
morning was talking about “pre-sequences”.2 She said certain expressions can
serve as pre-sequences. This reminded me of a Weld-work experience in a
South Asian village where I lived for almost two years. People would come up
to me in the street and say, “Dinner is ready, come on over.” Once, when I
seemed hesitant, the speaker said, “Don’t you want to come? We’ve made
dinner, why don’t you join us?”. I still hesitated. I didn’t know what to do.
Finally, thinking I had perhaps actually been invited and had forgotten about
it and might oVend them if I didn’t come, I went along. Well, it turned out that
I was...
C.L.P.: Invited?
16 Language and interaction
J.J.G.: I was not invited, no. It was just a greeting. An invitation is a very
solemn matter in the village. To invite someone, you send an emissary, a
young man in your household, to give the invitation. And even then, one
doesn’t go to the host’s house until a second emissary comes to call and asks
the guest to come over. This means that a special meal has been prepared and
is now ready. So that to say, “Dinner is ready”, to someone in the street, simply
counts as a form of greeting.
C.L.P.: This is a convention, the outcome of a typiWcation within a commu-
nity.
J.J.G.: One Wnds similar usages all over South Asia. In some parts of the
continent they say, “Have you eaten yet?”, as a way of conveying something
like, “How are you?”. Such conventions arose over time as outcomes of cultur-
ally-speciWc processes of typiWcation. They may reXect a time when people in
farming communities did not have everything they wanted to eat.
C.L.P.: So when you are confronted with this kind of utterance, if you come
from another community, like myself, you take it as an invitation.
J.J.G.: Exactly, as I did. In fact, once I went, and they had to prepare the food.
And I don’t know whether they had enough to feed me. They were not poor
people — it was a wealthy village — but still, you know, the women had to get
to work and make some more food.
C.L.P.: To come back to your notion of contextualization cues. You said that
you are not interested in the propositional content, but in the indexical
function in the Peircian sense, versus the symbolic, is that right?
J.J.G.: Yes, yes.
C.L.P.: So you see cues as orienting people to...
J.J.G.: Certain interpretations.
C.L.P.: And what about the orienting relationship? You have an addressee
who has to be oriented to, who has to be instructed via cues...
J.J.G.: That’s right.
C.L.P.: However, there are diVerent competences, repertoires, but let us as-
sume that you have two people with just the same repertoire or competence.
Faced with the same cue, they are supposed to be oriented in the same way, in
A discussion with John J. Gumperz 17
the sense that they are supposed to reach the same cued element ...
J.J.G.: The same inferences. Be able to draw similar inferences, yes.
C.L.P.: But this is not always the case. Given these two people belonging to the
same community, with the same repertoire of cues, are we sure that there are
no cases of ambiguity, contrary to the one-to-one relationship?
J.J.G.: There always are some ambiguities.
C.L.P.: There always are, okay.
J.J.G.: There usually are. Ambiguities always exist. This is why I emphasize
that the basic issue is not whether or not people understand factual informa-
tion, but whether or not participants in an encounter are able to attune to each
other’s interpretive processes. For example, ambiguities and misunderstand-
ings always occur, but you need to be able to repair them. And since conversa-
tional repairs must, for reasons inherent to conversing, rely in large part on
indirectness, repairs always make unusually high demands in the way of
shared inferences. In the cross-examination example I gave before, it is not
that the victim and the attorney were unable to draw similar inferences. On
the contrary, given the way jury trials work, they chose intentionally to rely on
shared interpretive conventions to convey conXicting accounts of what might
have happened.
C.L.P.: During the IADA Round Table, Edda Weigand spoke about coming to
an understanding in dialogue, a sort of Habermasian idea. However, you are
never sure you have reached a total agreement, a convergence of inferences, of
schemata and so on. So misunderstanding is a human condition, and you are
alone, in a certain sense, with your cues, facing the other person. And this is
the reason why I see individuals as a bit like solipsistic entities. They just send
and get cues, and cues about cues, trying, via them, to instruct the other’s
inferences or schema use.
J.J.G.: But we are not really talking about individuals agreeing on what some-
thing means. Nor is it so easy to say when people have the same communica-
tive background. Let me give another example. One of the very best of my
former students, a Nigerian, Niyi Akinnaso, was born in a Nigerian village. He
was the Wrst literate person in the community. His father, a local chief, was
secretary of the farmers’ cooperative and he used to keep accounts using a
local mechanical device, something like an abacus. As a child, Niyi used to
18 Language and interaction
keep records for his father. He went to missionary school in the region and
ultimately to the University, became an instructor there and then came to
Berkeley in the U.S. for his Ph.D. He now has a professorship in the United
States and specializes in issues of literacy.
Niyi and I now move in the same circles and are part of the same academic
community. I talk with him about a range of academic and other issues in
ways that I cannot with many others in the U.S., and would certainly never be
able to talk to his father or anyone else in Nigeria, although Niyi can. Through
participation in similar “networks of relationships” over time, we have been
socialized into similar network-speciWc communicative practices. Although
our backgrounds are about as diVerent as they could be, we share certain
communicative conventions and interpretive practices. It is long-term expo-
sure to similar communicative experience in institutionalized networks of
relationship and not language or community membership as such that lies at
the root of shared culture and shared inferential practices. In most people’s
lives, community membership is of course directly linked to participation in
such networks of relationships, but in our post-industrial worlds, it is less and
less possible to take this for granted.
Apart from cultural sharing in the face of diVerences in background, there
are now more and more cases of people brought up in what by ordinary
criteria counts as the same community, but in whose case surface similarity of
language and background hides deep underlying diVerences.
To go back to what I said about shared interpretations. It is not merely a
question of what something ‘means’. Ultimately, agreement on speciWc inter-
pretation presupposes the ability to negotiate repairs and re-negotiate misun-
derstandings, agree on how parts of an argument cohere, follow thematic
shifts and shifts in presuppositions, that is, share indexical conventions. The
more basic issue is to show how these tasks are accomplished. And it is for this
reason that my analysis puts so much stress on contextualization processes.
C.L.P.: You start from the idea of “metamessages” à la Bateson or Watzlawick,
and you seem to give great importance to these repair/adjustment procedures
via metacues, cues concerning other cues, or metapragmatic cues.
J.J.G.: Actually, the Wrst paper I did on this topic was a paper I gave in Urbino.
C.L.P.: It was published in 1974.
J.J.G.: In the 1974 series, yes (Gumperz 1974). That was my Wrst, but I didn’t
have the theory then. But now...
A discussion with John J. Gumperz 19
C.L.P.: Two other questions. What do you think about what could be called
“corpus pragmatics”? And the second question, what do you think about the
notion of “pragmatic creativity”?
J.J.G.: Corpus pragmatics?
C.L.P.: When talking about corpus linguistics, I have in mind the rather recent
wave of electronic corpora.
J.J.G.: But you mean ...?
C.L.P.: Sinclair, for example, John Sinclair’s work.
J.J.G.: John Sinclair, yes, and the Lund people.
22 Language and interaction
C.L.P.: For example, yes. But the problem, from your perspective, is how to
record the contexts as well...
J.J.G.: That’s right, exactly.
C.L.P.: That is the question.
J.J.G.: And also, it is not the corpus as such that is the problem. It is a question
of context, as well as of more basic questions of analysis. Take transcription.
We have as yet no generally agreed-upon, universally applicable system of
transcription. Not even something like the international phonetic alphabet.
And transcription, as recent work has shown, must always be related to
analysis. For conversation, the so-called Gail JeVerson transcription system
has come to be widely used throughout the world. But in my terms, it neglects
communicatively signiWcant prosodic and paralinguistic aspects of speech. It
cannot, for example, account for the interpretive import of phonetic variabil-
ity and so on.
There is also no agreement on methods of analysis. Quantitative methods
are widely used, but in the absence of agreement on what is signiWcant and
countable, quantitative methods are limited in value. Therefore, I currently do
not Wnd it productive for the issues that concern me. There has to be a division
of labor.
C.L.P.: Sure. It’s a problem I’ve found, too. When I have a number of cases
before me, I am faced with a set of pairs of text and context. But this is much
more than a corpus of texts only. Instead, when you have an electronic corpus,
you have a set of utterances only, you can enumerate their constituents, their
co-occurrence, and so on.
J.J.G.: There are quite sophisticated methods of computer-based scanning and
retrieval that work on raw data. But we need agreement on what to look for
and what the goals of the analysis are. Most discourse analysts work with
lexicon and clause level grammar and are concerned with issues of structure.
C.L.P.: Okay, it’s possible to carry out all kinds of search procedures, but given
this possibility, in your opinion, what’s the sense of these procedures with
regard to the problem of context?
J.J.G.: Such analyses cannot account for the eVect of context on interpretation.
You always need a prior analysis of context. You need independent, ethno-
A discussion with John J. Gumperz 23
C.L.P.: Okay, this is one case. You were able to reconstruct the dynamics of
this case...
J.J.G.: Only because I happened to have a graduate student with me who was
familiar with the relevant conventions. But one or two instances are only
suYcient for initial hypotheses. After that I had to go on to Wnd other in-
stances of the same phenomenon in order to construct a more general picture.
That is the way ethnography works. It is a matter of observation, hypothesis
formation, hypothesis testing on the basis of further information and then
critiques from locals who are familiar with the situation, and so on.
C.L.P.: Yes. But how can I continue? For example, I have tried to examine a set
of already published examples, already studied by other people.
J.J.G.: That’s very hard.
C.L.P.: And the problem is that you can reach diVerent conclusions, diVerent
hypotheses, even if you are still attached to the same presuppositions and
assumptions, because you can attend to or focus on diVerent subsets of cues
within the same overall set of cues.
J.J.G.: One solution is to have a local assistant who works with you on the
analysis and who will tell you what it is in what he/she hears or perceives that
leads to the interpretation. That gives you information at two levels, content
and form. The more one works with such interpretive analyses, the more
native-like one’s interpretations can become.
C.L.P.: But you can reach diVerent conclusions, given the same inputs or cues,
even if you are a native interpreter. So I think that the enterprise of prag-
maticians is a sort of game whose rules still have to be determined.
J.J.G.: It’s only a game if you share the rules, otherwise it can’t be a game. It’s
the kind of game ethnographers must learn to play. And many of them fail to
do it. The trick is to learn to ask the right kind, i.e. productive, questions.
Questions that lead to answers that give one the feeling that progress is being
made even though many problems remain.
C.L.P.: In your opinion, what kind of generalizations can we reach, starting
from these cases? Because they are single cases.
J.J.G.: Ultimately, large-scale quantitative studies are necessary. But current
studies of this kind tend to work with many unexplicated assumptions. Good,
A discussion with John J. Gumperz 25
occupations. In a way, the graduate student was being creative in calling his
fellows’ attention to a problem they all face.
C.L.P.: A last question. In your career, have you ever tried to describe a frame,
a script, or something like that?
J.J.G.: Bateson likens framing in communication to framing in paintings.
GoVman talks about how frames aVect the way we perceive an interaction and
how they change in the course of the interaction. For me framing is another
way of talking about the activity level presuppositions that aVect interpreta-
tions at any one point in an exchange. In my studies of South Asians living in
England, I have shown that frequently they perceive and frame encounters
quite diVerently from their native-speaking interlocutors. In one particularly
striking case, an interviewee seemed quite unreasonably to be denying or
contradicting information that the interviewer told him, even though only the
interviewer had Wrst-hand knowledge of the topic. It took us a long time to
discover that the interviewee was simply pleading for sympathy or under-
standing. Based on these and a series of other miscommunications, I was able
to argue that many South Asian speakers have ways of framing formal inter-
view-like encounters which are quite diVerent from those Westerners are
familiar with. A “frame”, as I use the term, can be described as a class of related
presuppositions that guide conduct in certain situations and that are directly
related to choice of topic and verbal strategies. I gave an example of this in my
paper in Mutualities in Dialogue (Gumperz 1995a).
C.L.P.: You said (in Gumperz 1995b) that Foppa, one of the editors of Mutu-
alities in Dialogue, asked you something about your paper.
J.J.G.: He asked me a question, one that I frequently hear: “How do you know
if people understand each other? When do I know that my interpretation is
the right one?”. And I said, “Well, only through context do you have the right
interpretation, and you’ll reach an agreement, but only as the result of a
dialogical process.”
C.L.P.: How can one study the cognitive aspects of this process?
J.J.G.: This is actually what I will talk about tomorrow.3 I will take two
examples. There’s a rape case, a re-analysis of Paul Drew’s example from a
rape case, and I’ll talk about some of the grammar that’s involved there. I’ll talk
about what each person has to know in order to do what they’re doing. And
28 Language and interaction
then I’ll take my example from a job interview to show again that they did not
share that knowledge, that therefore they couldn’t understand, and the more
they tried, the more things went wrong.
C.L.P.: That’s very interesting.
A.d.L.: Just to go back to the notion of genre for a minute: it is not very clear to
me. Is it a metapragmatic category?
J.J.G.: We use the term “genre” to label, that is, to refer to people’s modes of
talking. In a way genre is similar to what some people call speech style. But
genres are always associated with speciWc ideological values. As Hanks and
others have pointed out, by using speciWc genres, we allude to the values that
are associated with them.
A.d.L.: Yes, that’s what I thought.
J.J.G.: But the only thing is that I talk from the speciWc standpoint of some-
body who has either heard about it or analyzed it, and I am aware of what
other people think about it.
A.d.L.: I understand it as a sort of art of organizing activity in a special way, as
Luckmann says, “to solve a communicative problem”.
J.J.G.: Yes, that’s right. We organize activity to solve a communicative prob-
lem, but I would limit it. I would say that we do that in an activity, but genre is
the way we talk about it, and the way it gets transmitted and passed on. And in
fact I would say, when we talk about dialect, for example, Konstanz dialect
versus Kreuzlinger dialect, we’re making a judgement. That is, a judgement at
the level of genre. And what I would argue is that any attempt to delineate the
notion of dialect and style empirically, in terms of some kind of Saussurean or
empirical analysis, is not going to be very successful. You don’t need agree-
ment on the linguistic facts of what a genre is, as long as there’s some sharing
of certain key elements that would be the prototypical judgements that we’re
making, not a systematic analysis, that’s the point.
Notes
1. This discussion took place on March 31, 1995, at the Department of Linguistic and
Oriental Studies, University of Bologna, Italy, on the occasion of the 1995 Bologna Round
A discussion with John J. Gumperz 29
Table on Dialogue Analysis: Units, Relations and Strategies beyond the Sentence, organized
by the International Association for Dialogue Analysis (IADA) in honour of its President,
Sorin Stati. Transcribed and edited by Susan Eerdmans. Revised by John J. Gumperz and
received February 1996; revised for the present edition February 2002 [Editors’ note].
2. Gumperz is referring to the paper which Ionescu-Ruxandoiu Wrst presented at the 1995
IADA Round Table as “Pre-sequences as a conversational strategy”, now published as
Ionescu-Ruxandoiu (1997) [Editors’ note].
3. Gumperz concluded the 1995 IADA Round Table with a seminar on his work in
interactional sociolinguistics (see Eerdmans, this volume) [Editors’ note].
References
Chapter 3
Stephen C. Levinson
choice can partially determine situation — and this play between the presup-
posing and creative aspects of linguistic choice (to employ Silverstein’s termi-
nology) still dominates his work.
This background will also help to explain what may otherwise be taken as
a demerit: Gumperz’s analyses of conversations have nothing of the theoreti-
cal cleanliness to be found e.g. in conversational analysis. His tools are eclec-
tic, and the toolbox cluttered, on the one hand with pragmatic notions like
implicature, speech acts, frames, activities, cues, indices, and the like, and on
the other with sociological notions like network, ethnicity, gatekeepers, habi-
tus and so on. He is trying to depict processes that still defy understanding
with the best tools that come to hand from whatever school of analysis. And he
is trying to connect levels of analysis, from macro to micro, rather than to
develop an isolated level of conversational analysis. This is the real diVerence,
it seems to me, between his approach and e.g. conversation analysis, rather
perhaps than the one he oVers in the interview (in terms of his own preoccu-
pations with “situated on-line interpretation”; elsewhere he oVers a character-
ization of his own work in terms of the “social import of the Wne details of
verbal communication”, which seems more accurate).
With that as background, let me turn to the subject of Gumperz’s notion
of contextualization cues, which seems to me the central innovation in his
analysis of discourse. First, some remarks on the motivation. At the time he
was developing these ideas, I was lucky enough to be a graduate student in the
Language Behavior Research Lab at Berkeley where he worked. In Berkeley at
that time there was a rare and wonderful conXuence of ideas from diVerent
disciplines concerning the study of meaning — in philosophy, Grice and
Searle were expounding the ideas about implicature and speech acts now
associated with them, Fillmore was preoccupied with indexicality in language,
Kay with its sociological import, Robin LakoV with contextual meaning, and
George LakoV was attempting to wrap it all up in a uniWed theory of generative
semantics. It was an era of optimistic open-mindedness in which, for example,
Harvey Sacks could give an extended series of lectures to the Linguistics
Institute of the Linguistic Society of America. In Gumperz’s lab, we were taping
whatever we could get access to, and taping also in Weld locations around the
world, and we were trying to apply these theoretical ideas to the analysis of
actual snippets of conversation. It wasn’t easy. There was a yawning gulf
between what, on a simple-minded analysis, ‘the words meant’ and what we
took the participants to be self-evidently doing with their words. The gulf was
Contextualizing “contextualization cues” 33
“The King of Tonga will begin his state visit on Monday.” Some have argued
that this is as it were an abuse of the conventions governing the use of deWnite
descriptions which forces the interlocutor to “accommodate” the utterance by
interpreting it as if it were mutual knowledge that there is a King of Tonga. But
in fact there are many other (non-presuppositional) devices that work in this
way — conventional implicatures associated with expressions like ‘but’ or
‘even’ force the creative construction of a context for the interpretation of the
utterances they occur in. “Even Harry will come” projects a context in which it
is assumed that there is a ranking of people in terms of their likelihood of
coming, and Harry is very low on that scale. Or suppose I say, “What are you
doing tonight?” — this projects some kind of invitation or request sequence in
the oYng.
So the idea that utterances can carry their contexts with them, that is, the
set of assumptions necessary to unpack their interpretation, is not as outland-
ish as it may seem at Wrst sight. The paradoxical quality of the idea of a
contextualization cue is as much due to our wanting to hang on to the simple
information theoretic view of what a context is. But still, there is a puzzle: how
does it work? Here both the phenomenon and the theorists part company.
Perhaps we can distinguish diVerent species of context-invoking aspects of
utterances:
This is, I take it, the Gumperzian notion, in which the term “cue” denotes an
encoded or conventional reminder, like a knot in a handkerchief, where the
content of the memo is inferentially determined. Thus the “cue” cannot be
said to encode or directly invoke the interpretative background, it’s simply
a nudge to the inferential process. Moreover, the interpretative process is
guided more by a series of nudges now in one direction and now in another —
thus “cues” come as complex assemblages where the result of the whole
assemblage cannot be equated with the inferential results that each part alone
might have. The interpretive process may be guided by general pragmatic
principles of a Gricean sort, and thus be in many ways universal in character;
but the “cues” are anything but universal, indeed tending toward sub-cultural
diVerentiation. Hence the Gumperzian perspective on communication: at
once potentially possible across cultural divides and inevitably thwarted by
cultural nuances.
Further insights into how contextualization cues work, and their place in
an overall pragmatic scheme, may be found by turning to the second issue, the
relation between “foreground” and “background” in message structure. By
these terms I mean something entirely pretheoretical, the opposition between
central message content, coded propositional information, and peripheral,
more loosely associated and less clearly formulatable information, a sort of
informational penumbra. The opposition has aspects at diVerent levels: form,
content, and cognitive saliency (Table 1).
36 Language and interaction
Contextualization cues
Formal properties
1. a tendency towards non-segmentable, prosodic, paralinguistic, or kinesic
features;
2. if cued in lexico-syntax, then by lexical alternate (register) or minor gram-
matical class (e.g. particles);
3. any one clear function associated with a whole cluster of disparate features
(cf. Silverstein’s discontinuous feature), such features often being cross-
Contextualizing “contextualization cues” 37
Content properties
1. “out of awareness” background features; they are context-invocative, and
cannot therefore be easily directly responded to;
2. non-propositional content, e.g. aVectual, rhetorical, social or metalin-
guistic;
3. tendency to invoke holistic bodies of assumptions (contextual “frames”),
which then play a role in the interpretation of the utterance — e.g. help
select reference, clarify rhetorical structure, indicate illocutionary force,
etc.;
4. content not really coded, but “cued” — i.e. reliant on large dose of inferen-
tial reconstruction; thus the inferred content of the same cues can be
diVerent in diVerent utterances.
The reasons why Gumperz has made this notion central to his analysis of
discourse should now I think be clear. It is because he is interested in the
relation between the micro- and the macro-sociolinguistic that contextua-
lization cues have a special interest. First, their “out of awareness” features,
coupled with their essentially arbitrary but loose association with formal cues,
mean they can only be learnt by rich exposure to a communicative tradition, a
deep immersion in social networks. This takes us back to the earlier Gumperz
with his interest in social groups and their networks. Second, in the workings
of contextualization cues can be found the springs of dialect and group diVer-
entiation: we pseudo-speciate by slow degrees that start here in subtle mis-
communications. In Labov’s view, in contrast, sociolinguistic diVerences are
little carriers of prestige or stigma. In Gumperz’s view, the smallest formal
diVerences may carry with them a chasm of incomprehension, because con-
textualization cues invoke the essential interpretive background for the fore-
grounded message. This is where the barriers Wrst come up, later to grow into
the saplings of dialects or the oaks of languages. This is also the last place for
the barriers to come down, as demonstrated in his analyses of the mini-
tragedies — the failed job application, the lost welfare beneWt — which have
been a favourite theme of his recent work, and which can be seen to be an
almost inevitable outcome of diVerentiated socialization. Gumperz is Ameri-
canized enough to put forward here an optimistic message about the educabil-
ity of the gatekeepers in multi-ethnic societies; others will Wnd here a deeply
38 Language and interaction
Notes
1. My own response, with Penelope Brown, was to develop a theory of politeness that we
hoped would help to bridge the gap between the said and the unsaid. Other students of his,
like Susan Gal and Deborah Tannen, have developed their own responses.
2. One of the clearest expositions may be found in Gumperz (1992).
3. There are all sorts of things wrong with this, but that is another issue.
Contextualizing “contextualization cues” 39
References
Blom, Jan-Petter and Gumperz John J. 1972. “Social meaning in linguistic structure: Code-
switching in Norway”. In Directions in Sociolinguistics: The Ethnography of Communi-
cation, J.J. Gumperz and D. Hymes (eds.), 407–434. New York: Holt, Rinehart and
Winston.
Gumperz, John J. 1971. Language in Social Groups. A.S. Dil (ed.). Stanford, CA: Stanford
University Press.
Gumperz, John J. 1992. “Contextualization and understanding”. In Rethinking Context:
Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 229–252.
Cambridge: Cambridge University Press.
Gumperz, John J. and Levinson, Stephen C. (eds.). 1996. Rethinking Linguistic Relativity.
Cambridge: Cambridge University Press.
Silverstein, Michael. 1981. The Limits of Awareness. Sociolinguistic Working Paper 84.
Austin, Texas: Southwest Educational Development Laboratory.
40 Language and interaction
Contextualization and social meaning-making practices 41
Chapter 4
Contextualization and
social meaning-making practices
Paul J. Thibault
1. Preliminary considerations
(Prevignano and di Luzio, this volume: 22–23). This underscores the “se-
mantic importance of context” (Prevignano and di Luzio, this volume: 9).
The notion of “communicative practice” refers, then, to the real-time pro-
cesses whereby participants construe semantic signiWcance in the unfolding
speech event. Gumperz emphasizes how the participants — speakers, listen-
ers, audience members — actively and jointly construct the speech event.
Meaning-making is fundamentally dialogic in character. He criticizes the
structuralist tradition of phonological and syntactical analysis in so far as
this has failed to provide the analytical tools and conceptual frameworks
necessary for the analysis of “everyday talk” (Prevignano and di Luzio, this
volume: 8–9). In pointing out that “Saussurean phonological and grammati-
cal structures, deWned in terms of Wnite sets of oppositions” (Prevignano and
di Luzio, this volume: 9), cannot by themselves provide an account of com-
municative practice, Gumperz pays somewhat less attention to the impor-
tance of accounting for the semiotic resource systems which members
deploy in their enactment of particular communicative events. Gumperz has
written of the ways in which “the localized choice of one or the other of the
paradigmatic set suggests diVerent implicatures in terms of which the utter-
ance at hand is to be interpreted” (1992:46; di Luzio, this volume). However,
there is no detailed speciWcation of these paradigmatic resource systems and
their relations to actual instances. I shall now discuss this question.
How may the gap between the linguistic and the social in both linguistics
and social theory be overcome? In order to answer this, the relationship
between the semiotic resource systems of the community and its communi-
cative practices requires rethinking. This is especially so if we wish to avoid
being trapped by the abstract and de-contextualized formalisms in terms of
which language has been characterized as a closed and self-regulating system
of abstract forms.
If the detailed analysis of communicative practices is to “illuminate basic
issues in social theory” (Prevignano and di Luzio, this volume: 7) — that is, if
it is to show, for instance, how social ideologies are constructed and main-
tained in and through speciWc meaning-making practices, then we also need
an account of what it is possible to mean in a given community. In other
Contextualization and social meaning-making practices 43
ways (see above). There is no reason in principle why, for example, the
gestural semiotic modality, intonation, or depiction cannot be described as
systems of diVerences, or as systems of possible form-meaning relations, seen
as the resources for making context-dependent meanings in these modalities.
These may have proved intractable to the analytical methods and procedures
of traditional (structuralist) linguistic analysis. However, there is no reason
why a system of gestural diVerences, say, cannot be postulated so as to specify
the possibilities for meaning-making in that semiotic modality.
In any case, the assumption that lexical items can be speciWed on the basis
of Wxed form-meaning relations is itself questionable. From the point of view
of the system, semiotic forms of all kinds — linguistic, gestural, and so on —
have a meaning potential (Halliday 1978). In fully contextualized discourse,
there is a great deal of semantic variation in the use of the ‘same’ lexical item
according to factors which may relate to genre, for example, or to the semantic
relations that may be made between items in speciWc contexts. The system of
diVerences, seen as abstracted from speciWc contexts, is variably deployed
according to speciWc contextual requirements.
The second problem concerns the folk-linguistic view in our (western
European) cultures and their historical oVshoots that linguistic forms are only
explicitly communicative for insiders if they can be glossed as having a Wxed,
de-contextualized propositional meaning which functions to “refer to” desig-
nated states of aVairs in the real-world (Silverstein 1987).4
Now, from the point of view of the outsider-analyst of social meaning-
making, all semiotic forms are both meaningful and communicative, irrespec-
tive of whether they are explicitly communicative for insider-participants or
not. Once we drop our folk-theoretical Wction that some classes of lexicogram-
matical forms, prototypically nouns, may be assigned a Wxed referential mean-
ing, irrespective of context, then it should become clearer that, in the context
of a speciWc speech event, it makes no sense to uphold a distinction between
context-independent symbolic signs and context-dependent indexical ones. If
all semiotic modalities may, from the system point of view, be speciWed as
systems of diVerences, abstracted from speciWc contexts, then it is no less
important to show how it is through the co-deployment of these resources
that participants jointly enact the speech event. It is generally accepted that
intonation, say, can be analyzed as a system of potential meaning-making
diVerences which systematically contrast with each other, viz. rising, falling,
falling-rising, rising-falling tone, and so on. Abstracted from speciWc dis-
48 Language and interaction
course contexts, these can only have a meaning potential rather than a fully
contextualized discourse meaning.5
On the other hand, all semiotic forms are indexical when deployed in
some context. Further, the resources for indexing a particular form in a
contextually speciWc way may themselves be speciWed systemically. For ex-
ample, this is so of the often very complex systems of deictic categories in all
languages. In English, for example, the systemically speciWable diVerences
between ‘the’, ‘a’, ‘this’, ‘that’, and so on, specify how the Head noun in the
nominal group is grounded in relation to some material or other feature of the
situation which is made semiotically salient in the interaction. If I say, for
instance, “Look at that man over there”, to index a particular man in the
spatio-temporal purview of the speaker and listener, then I am using a selec-
tion from the system of deictic categories in English — ‘that’ — to index the
perceived object as an instantiation of the type-category speciWed by the Head
noun ‘man’ in the nominal group in question. Furthermore, I am doing so in
ways which index the man in question as a semiotically salient feature of the
interaction. The man is made a relevant part of the discourse context. From
the system perspective, the noun ‘man’ construes a type-category which is
internal to the grammar and semantics of the English language. As such, it
does not refer to any speciWc ‘real-world’ man. It is only when it is grounded
by the resources of deixis that the noun speciWes particular states of aVairs,
objects, events, and so on as instantiations of the type-category. Further, there
may be degrees of conformity to the criterial features speciWed by the type-
category (Langacker 1987:68–71).
A noun, as a grammatical class item, speciWes a type-category of a seman-
tic ‘thing’ which belongs to a given language system. Its indexical grounding
with respect to a given material or imagined object, event, etc. means that the
perceived or Wctive phenomenon so indexed through the resources of the
nominal group is construed as an instantiation of the type-category. There is
no direct word-object relation to start with. Nouns do not simply refer to
objects, states of aVairs, etc. in a direct, unmediated way. Rather, they specify
an experiential category which is internal to the language system. The
diVerence between signs with experiential content (e.g. nouns) and those with
none (e.g intonation) is one of degree. It is not an absolute distinction. In
discourse, all signs function in one way or another to index the wider situ-
ational context. Both lexicogrammar and intonation can be analytically re-
constructed as systems of contrasting form-meaning diVerences. In discourse,
Contextualization and social meaning-making practices 49
what is important are the cross-coupling patterns between (i) the diverse
semiotic modalities in operation, and (ii) these and the material world in the
enactment of the wider situational context (Lemke 1993:250; Thibault
1996:342). Informally, some examples are:
1. (nominal group / perceived object or event) // situation)
2. (lexicogrammar / intonation) // aVective-attitudinal orientation of speaker)
3. (phoneme / morpheme category) // semantic concept)
These examples illustrate, albeit informally, how local cross-coupling patterns
between, for example, a given nominal group and some perceived phenom-
enon in the material world, or between some choice in lexicogrammar and
intonation, combine to index some aspect of the wider context which the
combining of the two modalities helps to constitute. In each case, there is a
Wrst-order contextual redundancy pattern between the selections from the
two modalities. This is represented by the single slash, viz. ‘/’. Yet, in each case
it is necessary to talk about a second-order relation which speciWes the wider
context in which this particular combination has the meaning it does. This is
represented by the double slash.
But notice that the third of these three examples refers to the language
system in the narrow sense. The same basic principle applies here too. Thus,
the co-patterning of a particular phoneme category with a morpheme is a
Wrst-order contextual relation on the basis of which a second-order semantic
concept is construed in the language in question.
In each case we see an important principle at work. This is the principle of
‘what goes with what else’ in the enacting of the wider situational context in
which a particular combination has its meaning. An index was deWned by
Peirce as a sign which is “physically connected” to some object (Nöth
1990:123–4). If indexicality is centrally concerned with the making and speci-
fying of contextual relations, as Gumperz’s work shows, then this deWnition is
no longer very satisfactory for a theory of discourse. The general problem of
“indexicality” has to do with how particular cross-couplings among diVerent
semiotic modalities and between these and selected aspects of the phenomenal
world instantiate a local discourse context by specifying in various ways the
parameters of the interaction itself.
50 Language and interaction
that these, too, always function to index some aspect of the relevant contextual
parameters whereby a particular communicative event is enacted.
an important early attempt to explore the diversity of IMMP’s, and the rela-
tions of alliance, conXict, co-optation, and so on among these in particular
texts and occasions of discourse (Lemke 1988b; Thibault 1989).
As Gumperz’s examples show, IMMP’s entail more abstract, or higher-
order, relations of relations of relations ... in Bateson’s sense. They are neither
systems of ready-made text-types nor recipes showing how to construct par-
ticular types of texts. Rather, they provide an interpretative resource for
construing culturally salient patterns of meaning in speciWc instances. In
neither case is it a simple question of whether the speciWc instance matches the
intertextual pattern or not. Rather, discourse speciWc meanings are inter-
preted on the basis of whether and to what extent they conform to or contrast
with a given intertextual pattern. IMMP’s are neither homogeneous nor con-
sistent across the diverse practices that constitute a given community. That is
why they are said to be heteroglossic in Bakhtin’s sense.
The notions of “background information” and “inferential procedure”
risk both de-semioticizing what are in fact the systems of IMMP’s in a com-
munity and also reifying their essentially social nature as if these were proper-
ties of individual mind. Neither of these notions adequately speciWes the local
(instantial) resources whereby discourse participants reconstruct the relevant
IMMP’s as they deploy these in their own meaning-making practices. The
notion of IMMP’s better enables us to reconstruct in our analyses the diverse
ways in which members indicate which intertextual relations are relevant, and
how these relate to which other occasions of discourse (Lemke 1985:276).
Neither “background information” nor “inferential procedure” is ade-
quately grounded in a theory of the semiotic forms and their modes of deploy-
ment in discourse to bring about the conceptually uniWed account — one
which can reconstruct the links between resource systems and social mean-
ing-making practices — to make this possible. The idea of a ‘background’
certainly has the merit of suggesting the diVuse, taken-for-granted, and out-
of-focus assumptions that characterize the lifeworld or the habitus of values,
practices and orientations that inform the speciWc discursive event (see also
Levinson, this volume). As it stands, however, I remain unconvinced that it is
able to provide us with the means for analytically reconstructing the links
between the “macro” and the “micro” perspectives in a conceptually uniWed
way (see Thibault 1991:229–40 for further discussion).
56 Language and interaction
Gumperz also refers to Bakhtin’s notion of speech genres. This is the meta-
discursive dimension of social meaning-making. For Gumperz, genre is “an
ideologically charged metapragmatic assessment”, to echo a Silversteinian
turn of phrase, or “forms of talk about talk”, in the GoVmanesque parlance.
From this point of view, discourse participants make interpretative as-
sumptions about the meaning of their own and others’ contributions to the
ongoing discourse event. They self-reXexively interpret to varying degrees of
explicitness and awareness the meaning of particular discourse events or parts
of these, relative to some social viewpoint. To do so, members deploy the
meaning-making resources of the system so as to assign signiWcance to their
own and others’ local instantiations of these same resources. Social meaning-
making practices may be said to be metadiscursive in this sense.
Whereas IMMP’s connect the patterns in the speciWc instance to still
wider patterns of meaning in a given culture, metadiscursive meaning-mak-
ing practices (MMMP’s) serve to articulate the links between the local, or
occasion-speciWc, meaning selections in some speciWc communicative event
to the paradigmatically-organized contrast sets of context-types in relation to
which the former have their meaning. That is, MMMP’s link the speciWc
instance or meaning choice to the metasystem of possible alternative choices
from which it was selected. This may be so to varying degrees of completeness
and explicitness. MMMP’s bring into focus the meaning of the particular act
by providing a connection between this and the higher-order metasystem of
possibilities which gave rise to it.
However, it is not enough simply to say that metapragmatic assessments
involve “talk about talk”. There are many diVerent ways in which members
talk about talk. As analysts, we need to be able to specify which forms of talk
about talk occur, and in which context-types. What are the speciWcally ‘meta’
relations which specify the relations between the particular act, the context in
which it occurs, the always social viewpoints of the participants in the dis-
course, and the metasystem — the contrast sets — which any occasion of
metadiscourse always invokes? Whenever we engage in talk about talk, we
always invoke or reconstruct, to varying degrees of completeness, the meta-
system in terms of which the given act has its meaning. MMMP’s make
explicit in some way the fact that the metasystem is always immanent in
speciWc occasions of discourse. Thus, MMMP’s interpret or gloss the meaning
Contextualization and social meaning-making practices 57
of the instance from the global standpoint of the metasystem. In making the
link between local occasion and system, participants show how they have their
own socially-constrained metatheories — folk or otherwise — concerning the
ways in which occasion-speciWc acts of meaning are relatable to the global
system of possibilities which engendered them.
I. Indexicality is concerned with the ways in which selections from the global
resource systems of the various semiotic modalities cross-classify and inte-
grate both with each other and with selected features of the material-phenom-
enal world so as to enact a local, occasion-speciWc, discourse event;
III. Metadiscursivity is concerned with the ways in which the global system of
possible sets of alternatives — the metasystem — is made locally salient in
some interaction. It does so by providing the means by which participants
construct a self-reXexive connection between the metasystem and some dis-
course event. This means that discourse participants draw on this system in
the process of assigning meaning to some discourse event. They do so by
explicitly focussing the metasystem onto the immediate interpretative prob-
lem. In this way, the relevance of the global possibilities of the metasystem is
made an explicit object of metasemiotic consciousness, relative to some par-
ticipant viewpoint.
each other in the real-time unfolding of the discourse. This depends on the
ways in which interactants integrate all of the diverse contextualization cues
into their sense of the overall activity which is taking place (Gumperz
1992:44). Rather than trying to interpret the other semiotic modalities from
the exclusive standpoint of linguistic analysis, the task should be to develop
ways of grounding the multimodal nature of such events in a functional
semantic approach to action as a whole. The three-way distinction I made
above attempts to provide a suYciently generalized and integrated framework
for analyzing the local and the global dimensions of social meaning-making.
Gumperz’s work on contextualization makes a seminal contribution to the
development of a semiotics of action in which the functional integration of all
of the semiotic modalities which are involved is accounted for.
Notes
1. Thus, Wittgenstein in his Tractatus: “3.203 A name means an object. The object is its
meaning” (1969 [1921]:23).
2. See Kenneth Pike’s (1967) wave and Weld, as distinct from his particle, perspectives on
language, prosodies of all kinds, suprasegmentals, and so on.
3. It may well turn out, as further research is conducted on this question, that what we
conventionally refer to as ‘language’ and treat as a uniWed semiotic system will turn out to
be a diverse set of possibilities whereby linguistic forms are integrated with other semiotic
modalities. In the conventional view, speech and writing have been seen as diVerent
material realizations of the same lexicogrammatical forms in the oral-aural and optical-
visual channels. In my view, speech and writing are themselves folk-linguistic glosses on a
diverse set of multimodal meaning-making practices in which diVerent possible integra-
tions of lexicogrammatical, prosodic, paralinguistic, graphological, pictorial, kinesic, spa-
tial, rhythmic, and other resources characterize the discourse genres — the semiotic
action formations — of a given community. In such a view, genre will, increasingly, be
deWned as the typical modes of deployment of the multimodal meaning-making resources
available to the members of a given community.
4. Alan Rumsey (1990) points out how, in the Ungarinyin language of Australia, the
meanings of linguistic locutions are glossed, not on the basis of de-contextualized criteria
of reference, but, instead, on ‘pragmatic’ criteria concerning the ways in which the
locution, or some alternative, locally construed as equivalent to it, is used in the various
context-types of the culture.
5. Bolinger, for instance, recognized long ago that entire intonational conWgurations may
be related to speciWc context-types in culturally speciWc ways (1951:209). Further, and just
like lexicogrammatical forms, these are “ineVable” in the sense that they cannot be reliably
glossed in terms of some extra-semiotic absolute standard (Halliday 1988). Rather, the
Contextualization and social meaning-making practices 59
meaning of a particular intonation selection itself derives in the Wrst instance from the
place which any given selection is felt to occupy in an overall system of contrasting options
which deWnes its semiotic value.
6. In each case, Gumperz acknowledges the diverse sources which have inXuenced his
thinking on these issues. In relation to (1) above, he singles out Peirce’s notion of indexical
signs and GoVman’s frames; he relates (2) to GarWnkel’s inferential processes; and (3)
derives in part from Bakhtin’s theory of speech genres, as well as Silverstein’s (1992:60)
“metapragmatics”.
References
Bakhtin, Mikhail. 1981. “Discourse in the novel”. In The Dialogic Imagination: Four Essays
by M.M. Bakhtin, C. Emerson and M. Holquist (trans.), M. Holquist (ed.), 259–422.
Austin and London: University of Texas Press.
Bakhtin, Mikhail. 1986. “The problem of speech genres”. In Speech Genres and Other Late
Essays, V.W. McGee (trans.), C. Emerson and M. Holquist (eds.), 60–102. Austin:
University of Texas Press.
Bateson, Gregory. 1973. “The logical categories of learning and communication”. In Steps
to an Ecology of Mind, 250–279. London and New York: Paladin.
Bolinger, Dwight L. 1951. “Intonation: Levels versus conWgurations”. Word 7 (3):199–
210.
Di Luzio, Aldo. “Presenting John J. Gumperz”. This volume.
Gumperz, John J. 1992. “Contextualization revisited”. In The Contextualization of Lan-
guage, P. Auer and A. di Luzio (eds.), 39–53. Amsterdam and Philadelphia: John
Benjamins.
Halliday, M.A.K. 1978. Language as Social Semiotic. The Social Interpretation of Language
and Meaning. London: Arnold.
Halliday, M.A.K. 1988. “On the ineVability of grammatical categories”. In Linguistics in a
Systemic Perspective, J.D. Benson, M.J. Cummings, W.S. Greaves (eds.), 27–51.
Amsterdam and Philadelphia: John Benjamins.
Halliday, M.A.K. 1994. An Introduction to Functional Grammar. Second edition. London
and Melbourne: Arnold.
Langacker, Ronald W. 1987. Foundations of Cognitive Grammar, Vol.1. Theoretical Pre-
requisites. Stanford, CA: Stanford University Press.
Lemke, Jay. 1984. “Action, context, and meaning”. In Semiotics and Education. Victoria
University, Toronto: Monographs, Working Papers and Prepublications of the
Toronto Semiotic Circle 2:63–93.
Lemke, Jay. 1985. “Ideology, intertextuality, and the notion of register”. In Systemic
Perspectives on Discourse, Vol. 1. Selected Papers from the 9th International Systemic
Workshop, J.D. Benson and W.S. Greaves (eds.), 275–294. Norwood, NJ: Ablex.
Lemke, Jay. 1988a. “Genres, semantics, and classroom education”. Linguistics and Educa-
tion 1 (1):81–99.
60 Language and interaction
Lemke, Jay. 1988b. “Discourses in conXict: Heteroglossia and text semantics”. In Systemic
Functional Approaches to Discourse. Selected Papers from the 12th International
Systemic Workshop, J.D. Benson and W.S. Greaves (eds.), 29–50. Norwood, NJ:
Ablex.
Lemke, Jay. 1990. Talking Science. Language, Learning, and Values. Norwood, NJ: Ablex.
Lemke, Jay. 1993. “Discourse, dynamics, and social change”. In Language as Cultural
Dynamic. Cultural Dynamics VI (1–2):243–275.
Lemke, Jay. 1998. “Multiplying meaning: Visual and verbal semiotics in scientiWc texts”.
In Reading Science: Critical and Functional Perspectives on Discourses of Science, J.R.
Martin and R. Veel (eds.), 87–113. London and New York: Routledge.
Levinson, Stephen C. 1997. “Contextualizing ‘contextualization cues’”. This volume.
Martin, James R. 1992. English Text. System and Structure. Amsterdam and Philadelphia:
John Benjamins.
Nöth, Winfried. 1990. Handbook of Semiotics. Bloomington and Indianapolis: Indiana
University Press.
Pike, Kenneth L. 1967. Language in Relation to a UniWed Theory of the Structure of Human
Behavior. Second, revised edition. The Hague and Paris: Mouton.
Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This
volume.
Rumsey, Alan. 1990. “Wording, meaning, and linguistic ideology”. American Anthropolo-
gist 92 (2):346–361.
Silverstein, Michael. 1987. “The three faces of ‘function’: Preliminaries to a psychology of
language”. In Social and Functional Approaches to Language and Thought, 17–38.
Orlando and London: Academic Press.
Silverstein, Michael. 1992. “The indeterminacy of contextualization: When is enough
enough?”. In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 55–
76. Amsterdam and Philadelphia: John Benjamins.
Thibault, Paul J. 1986. “Thematic system analysis and the construction of knowledge and
belief in discourse”. In Text, Discourse, and Context: A Social Semiotic Approach.
Victoria University, Toronto: Monographs, Working Papers and Prepublications of
the Toronto Semiotic Circle 3:44–91.
Thibault, Paul J. 1989. “Semantic variation, social heteroglossia, intertextuality: Thematic
and axiological meaning in spoken discourse”. Critical Studies 1 (2):181–209.
Thibault, Paul J. 1990. “Questions of genre and intertextuality in some Australian televi-
sion advertisements”. In The Televised Text, R. Rossini Favretti (ed.), 89–131. Bolo-
gna: Pàtron.
Thibault, Paul J. 1991. Social Semiotics as Praxis. Text, Social Meaning Making and
Nabokov’s ‘Ada’ [Theory and History of Literature Series 74]. Minneapolis and
Oxford: University of Minnesota Press.
Thibault, Paul J. 1994. “Text and/or context?”. In The Semiotic Review of Books (Toronto)
5 (2):10–12.
Thibault, Paul J. 1997. Re-reading Saussure. The Dynamics of Signs in Social Life. London
and New York: Routledge.
Contextualization and social meaning-making practices 61
Thibault, Paul J. and Van Leeuwen, Theo. 1996. “Grammar, society, and the speech act:
Renewing the connections”. Journal of Pragmatics 25:561–585.
Van Leeuwen, Theo. 1992. “The schoolbook as a multimodal text”. Internationale Schul-
buchforschung 14:35–58.
Wittgenstein, Ludwig. 1969 [1921]. Tractatus Logico-Philosophicus / Logisch-Philoso-
phische Abhandlung. D.F. Pears and B.F. McGuinness (trans.). London: Routledge
and Kegan Paul.
62 Language and interaction
Gumperz and the minims of interaction 63
Chapter 5
Carlo L. Prevignano
1. Introduction
2. Some preliminaries
the activity models and the know-how which are used by participants them-
selves to categorize and monitor “what’s going on” and to act in life. Following
Gumperz, Svennevig (1999:11–12) prefers to call the latter “activity frames”.
As an analyst of the “fabric of our social life” (1982a:7), sub specie commu-
nicationis, Gumperz has contributed to the identiWcation of an interaction
principle which I intend to present here in three diVerent speciWcations, each
on a diVerent scale (see level i. to level iii. in Table 1). These range from a
general action-theoretical or praxeological interpretation to a linguistic one.
A caveat. Gumperz has contributed to the individuation of that principle
in the sense that there is a complicity in the crime (and his accomplices
include, at the very least, Bateson, GarWnkel, GoVman, Sacks, Erickson and
Levinson — an invisible college à la Winkin (1981/1995:21), now quite vis-
ible; see Kendon 1990 and Auer 1992:22–23).2 However, it is Gumperz who
has provided its most extensive analyses at the linguistic levels of speciWcation,
in particular studying what he deWnes as “contextualization cues”.
Here is the principle, which I propose to call the “semiotic principle of
interaction”: human agents do not simply act, but are also accustomed to
signaling or cueing the types of act/action/activity they are engaging in; in other
words, they make them interpretable at least via what could be called “Gum-
perzian cues”, rather than (only) with overtly lexicalized or verbalized explana-
tions and rationalizations. Among other reasons, they do so in order to
economically facilitate coordination, to simulate, and to lower or raise the
anxiety level of the other agents present. “Uncertainty reduction” (Giles et al.
1979:372; Berger and Bradac 1982) is accompanied by certainty induction,
and certainty reduction by uncertainty induction. Table 1 presents three
speciWcations of the semiotic principle of interaction.
I cannot resist the temptation of calling the Wrst level of speciWcation of the
principle, the “GarWnkel-and-Sacks level”, the second, the “Levinson level”
(see Levinson 1979), and the third, the “Gumperz level”.3 While presenting
“interpretive sociolinguistics”, Auer and di Luzio have anticipated a deWnition
of what I call the Gumperz level, since they have referred to “the participants’
continuous task and accomplishment to make interpretable and interpret
each other’s linguistic activities” also via cues (1984:viii). A distinct, but
68 Language and interaction
i. all human
ii. entirely and predominantly non- act/action/activity types are
verbal as well as predominantly signaled at least by cues
verbal
iii. even exclusively verbal
parallel, set of tasks (i.e., signaling and identifying or detecting signaled prag-
matic functions of diVerent linguistic activities or activity-types) has been
considered by Stati (1982, 1990) and others from the perspective of dialogue
analysis, and in relation to dialogue constituents other than contextualization
cues. In my view, the Gumperz level corresponds to both cueing itself (as the
process of proposing the use of x as a cue for y as a cued entity, i. e. of a cueing
function and its functors; see Kendon 1996) and cue processing (emission and
reception of both coded and still uncoded or unshared cues), with a distinc-
tion (which has already been used by Richard Petty and John Cacioppo in
their research on persuasion) between central and peripheral processing.
Bazzanella 1990, 1994; see also Dascalu Jinga and Vanelli 1996), if, impor-
tantly, Gumperzian signals also include clusters of co-occurring cues. As both
output and input, the latter require strategies, not simply processing rules. In
order to capture and represent the many-to-many semiotic circuitry which
Gumperz has characterized as “contextualization” (both in output and in
input), Peircian semiotics is not enough, though this question cannot be
explored here. Gumperzian cues can be coded and uncoded, shared and
unshared, but also plural, as multimodal clusters of co-occurrent triggers.
Further, there is another research problem, which is also pointed out by
Bateson’s metaphor of a shower of signals (1971/1981:125; for Bateson’s cues
and their functions, cf. Rawlins 1987), when these signals are considered as an
output from and an input for human agents. It is the problem of possibility
conditions for their human processing, and in general for the human process-
ing of semiotic multimodality (cf. Thibault, this volume).
If from a “structureless barrage-of-signals view” of human interaction (i.e.
an idea of interaction as an “unstructured exchange” of signals, again in the
words of Brown and Levinson (1979:293)), via a “cross-shower of signals”
view of it (à la Bateson), we come to a structural approach to interaction as a
multichannel event, it is important to avoid studying it only as a turn struc-
ture. An antidote against today’s triumphant reduction of human interaction
to a turn structure and its variants (from a turn-by-turn sequence to a bundle
of n tracks, for n participants, i.e. a multitrack space with tracks being multi-
channel and turns occurring on each of the multichannel tracks; I propose this
variant in Prevignano and Thibault, 2002:166), an antidote, I say, can be an n-
level theory and analysis of interaction, as well as a corresponding n-level
processing model.
In this direction, one of Gumperz’s contributions has been to analyze
occasions of human interaction on a move level, rather than on a turn level,
seeing “actors engaged in strategically formulating and positioning their
moves,” as well as interpreting and reinterpreting other participants’ moves.
Gumperz has distinguished “moves”, “countermoves”, “conWrmations”, and
“disconWrmations”, i.e. move sequences (see Prevignano and di Luzio, this
volume:12). What could be called “move processing” is analyzed by Gumperz
as conversationalists’ execution of “rhetorical strategies” and interpretive pro-
cedures, with positional (etc.) constraints on their applications. Among inter-
actional abilities, it is “rhetorical ability”, as the command or control of a set of
rhetorical strategies being developed in networks of interactions, that
70 Language and interaction
was discovered by Gumperz, working back to the utterance surface from some
participants’ and other members’ assessments, to be a cue accompanying the
utterance which was addressed by a black student to other black students.
Gumperz was able to consult four groups of “judges” (1982a:31–32). Accord-
ing to the members of the last group, it is the rhythm which gives the whole
Gumperz and the minims of interaction 71
According to (a), (b), and (c), both the output minim (“Signal via cues what
you are doing”) and the corresponding input minim (“Look for what the other
person is doing, i.e. for his or her activity-type cues”), seem to have been
applied by the student and his addressees, respectively. But can the results of a
reconstructive methodology be considered as a post-factum ‘explanation’ of
cases such as (1)? Postdictions (vs. predictions) presuppose the resolution of
postdictability and postdeterminability problems (Prevignano 1991, 1993).
There is an extraordinary game which Gumperz has been playing for more
than forty years: integration. Its risk is eclecticism, and Gumperz knows that
he takes “a somewhat eclectic approach” (1996a:359) to interaction via lan-
guages. From when he was Wfty (in 1972), he has subsequently presented
his research as speech event analysis (Gumperz 1972:17–18; cf. Gumperz
1984:281), conversational analysis of social meaning and sociolinguistic
analysis of conversations (Gumperz and Herasimchuk 1973), sociolinguistics
of interpersonal communication (Gumperz 1974), the interpretive approach
to conversation (Gumperz 1982a:6, 35–7; for a presentation of interpretive
sociolinguistics, as Auer called it, see Auer and di Luzio 1984), and interac-
tional sociolinguistic analysis (Gumperz 1982a:210) or interactional socio-
72 Language and interaction
6. The future
Notes
1. Revised and augmented version of a text previously published as Prevignano 1997. Paul
Thibault’s comments were valuable in the preparation of this version.
2. See also Murray (1994) for a group-sensitive Gumperzologist’s report.
3. See Gumperz 1974:18V, 1977:205V, 1980:119, on a distinction between speech acts and
speech activities; Gumperz 1982b:329V, and Gumperz and Cook-Gumperz 1982:11V, on
a distinction between activities and tasks; see also Gumperz 1989a:78–79, 1990a:439,
1992:43V.
References
Chapter 6
A commentary on a discussion
with John J. Gumperz1
Afzal Ballim
too far from the practical problems of using language for it to have much
bearing on what they do;
Of course, interaction does take place, and the work in computational linguis-
tics on phonology, morphology, part-of-speech tagging, and syntactical anal-
ysis plays an important role in modern NLP. It is when one goes beyond these
realms to semantics and pragmatics that the interaction breaks down. In part,
what I wish to say in this commentary is that the work of Gumperz is an
example of why the interaction should not fail.
In order to situate my detailed comments on the discussion, it is necessary
to give a brief description of my own work in NLP. I am concerned with all
forms of computer-mediated communication; in particular, with human-
machine interaction (and in this I include language-based human-machine
communication), as well as with human-human communication assisted by
computer (such as in Computer-Supported Collaborative Work [CSCW]). The
ultimate goal in this work is to create systems that communicate.
As such, I am less encumbered by having to provide neat theoretical
accounts of linguistic phenomena than are my colleagues in computational
linguistics (although such neat theories are still desirable). Instead, I can take
shortcuts around such philosophical problems as “symbol grounding” or even
“what is semantics?”. Indeed, in recent years my work has concentrated on
“robust interaction”, in which we admit that the computer interpretation and
model of the interaction is incomplete, but we try to work with what we know
the computer can do. This tendency towards robustness and practicality has
been notable in work on human-computer interaction since the mid-90s.
If one can describe the desired functionality of a system, and the system is
able to communicate with a human to achieve that functionality, then the goal
has been reached and it is not necessary to ask if the computer system has
understood. That said, the sheer complexity of the problem of achieving
communication between a computer system and a human is such that one
cannot help but formulate theories that may well be applicable to human-
human communication as well. For the same reasons, theories of human-
human communication must be considered, as they are still important.
Commentary on a discussion with John J. Gumperz 81
The crux of my work (see Ballim and Wilks 1991; Ballim and Pallotta,
forthcoming) is in interpreting utterances in the light of their context as
perceived by the speakers and hearers involved. That is, the system formulates
a model of the knowledge, beliefs, desires, intentions, and other attitudes of
the interlocutors that are relevant in the context of the dialogue, and further-
more it formulates models of how the interlocutors model each other at that
point. These models are then used as the basis for interpretation of the dis-
course. They are built using past and current experience with the individual
interlocutors, as well as through generalizations about the background of
various participants.
In general terms, I feel that this approach is very compatible with Gum-
perz’s ethnographically-based sociolinguistic analysis of dialogue, since no-
tions such as “networks of relationships”, “situated inference”, and “speech
events” are all fundamental to the processing involved. In turn, the eVec-
tiveness of this NLP method can be seen as providing indirect evidence of the
validity of Gumperz’s approach. This can be seen more clearly by considering
in more detail speciWc parts of the discussion.
Speech events
The notion of the context of a communication has long been central in NLP.
My own work with Yorick Wilks, as well as that of people such as Barbara
Grosz, and Barwise and Perry’s (1983) “situation semantics”, are all examples
of taking context as being of primary importance in discourse understanding.
However, Gumperz’s notion of “contextualization cues” is not something
accounted for or used by any of these theories. His aim to show the interaction
of indexical signs with symbolic signs may, I believe, make a great contribu-
tion to NLP, where the problems of deWning the structure of discourse and
demarcating shifts in context are far from resolved.
Gumperz presents a schema for overcoming some of the ambiguity inher-
ent in the inferential process associated with discourse. I believe that combin-
ing this schema with the methods commonly used in NLP, and those adopted
in my own work (of Wnding hypotheses that are maximally consistent with the
context and interlocutors’ background) may yield an interesting and novel
approach to problems of ambiguity.
The method that I mention in my work of Wnding maximally consistent
hypotheses that explain discourse events has undercurrents of Gumperz’s
“activity types”, in particular because they may be based on predictive reason-
ing on where the conversation is leading the participants. Such predictive
structures are of course a notable aspect of much NLP work (in particular,
Commentary on a discussion with John J. Gumperz 83
Schank’s work on scripts and the derivatives from that work can be seen as
aiming to embody form-context relationships).
Corpus pragmatics
Gumperz is questioned about the use of corpora. His problem with such
corpora is that these transcriptions lack many of the contextual cues that he
believes to be fundamental to understanding conversation. A large part of the
problem is the lack of standards for recording such cues. Given the growing
and already widespread use of corpora in computational linguistics and NLP,
a contribution in this domain which establishes standards for those cues
would be of great beneWt to many people.
Conclusion
Notes
References
Ballim, Afzal and Pallotta, Vincenzo. Forthcoming. Dialogue through the ViewFinder.
Ballim, Afzal and Wilks, Yorick. 1991. ArtiWcial Believers: The Ascription of Belief. Hills-
dale, NJ: Lawrence Erlbaum.
Barwise, Jon and Perry, John. 1983. Situations and Attitudes. Cambridge, Mass.: The MIT
Press.
Coakes, Elayne, Willis, Dianne and Lloyd-Jones, Raymond (eds.). 2000. The New Socio-
tech: GraYti on the Long Wall. London: Springer.
Goddard, CliV. 1998. Semantic Analysis: A Practical Introduction. Oxford: Oxford Uni-
versity Press.
Greif, Irene. 1998. Computer-Supported Cooperative Work: A Book of Readings. San
Mateo, CA: Morgan Kaufmann.
Horton, William S. and Keysar, Boaz. 1996. “When do speakers take into account com-
mon ground?” Cognition 59:91–117.
Lyons, John. 1977. Semantics. Cambridge: Cambridge University Press.
Sproull, Lee and Kiesler, Sara. 1988. “Reducing social context cues: Electronic mail in
organizational communication”. In Computer-Supported Cooperative Work: A Book
of Readings, I. Greif (ed.), 683–712. San Mateo, CA: Morgan Kaufmann.
John J. Gumperz: A review 85
Chapter 7
Susan L. Eerdmans
1. Introduction
In this chapter, I would like to outline some of John Gumperz’s most impor-
tant contributions to Interactional Sociolinguistics in the last few years. An
attempt will be made to deWne in a nutshell Gumperz’s particular analytic
approach, and to demonstrate, with examples taken from two of his case
studies, how this is applied to real data.
‘Interactional Sociolinguistics’ is a general term used to describe a number
of qualitative approaches to the analysis of interactive and dialogic processes
in interpersonal communication. Its roots lie in the Welds of sociology, anthro-
pology and linguistics, as reXected by interactional sociolinguists’ interest in
the closely-knit interplay existing between language, society and culture
(SchiVrin 1994, 1996), and in the “communicative import of diversity”
(Gumperz 1999b:453). The focus of their attention is on naturally-occurring
utterances in social context, the aim of their analysis being to discover and
understand “how interpretation and interaction are based upon the interrela-
tionship of social and linguistic meanings” (SchiVrin 1994:8).
2. Background
atic ways in which language shift both reXects and deWnes social and cultural
boundaries” (Gumperz 1992a:229). By focussing his sociolinguistic analysis
on the qualitative aspects of verbal and non-verbal processes, together with
those of shared cognition and understanding, Gumperz has demonstrated his
commitment to a general theory. This is not only a theory of language, but of
verbal communication, of communicative action, “which integrates what we
know about grammar, culture and intercultural conventions into a single,
overall framework of concepts and analytical procedures” (Gumperz 1982:4).
The analytical methodology adopted by Gumperz for his in-depth investiga-
tion of discursive processes in what GarWnkel calls “naturally organized activi-
ties” is based principally on the inductive procedure of analysis adhered to by
conversation analysts. 2 The aim of the conversation analytic, turn-by-turn
study of sequential processes is to discover the naturally-occurring and recur-
ring mechanisms and patterns of conversational management, seeking “to
describe the underlying social organization[...]through which orderly and
intelligible social interaction is possible” (Goodwin and Heritage 1990:283).3
In keeping with conversation analysts’ ethnomethodological background,
which leads them to avoid any premature generalization or idealization, this
research is carried out assuming very little about participants’ social and
cultural background a priori; aspects of the context such as the social identity
of participants and the setting are considered to be “a category of social life
and conduct that is subject to locally situated interpretive activity” (SchiVrin
1994:235).
Alongside elements of the conversation analytic method of investigating
sequential processes, Gumperz’s interactional sociolinguistics also employs
features of more general approaches to discourse analysis, including, for
example, Gricean analysis of inferential processes and implicature. One of the
key notions of Gumperz’s perspective on verbal communication is communi-
cative intent, the interpretation made by the listener(s) of what the speaker is
trying to communicate, roughly correspondent to the illocutionary force of
what is conveyed (Gumperz 1992a:230).4 The assessment of participants’
communicative intentions is an essential part of conversational inference,
another fundamental concept that Gumperz is concerned with, deWned as
“the situated and presupposition-bound interpretive processes by which in-
terlocutors assess what they perceive at any one point in a verbal encounter
and on which they base their responses” (Gumperz 1996c:375). In the context
of a “communicative ecology” (Gumperz 1990), conversational inference thus
John J. Gumperz: A review 87
how they are used in context, or, as Gumperz argues, in terms of how they
enter into a particular interactive or dialogic process; (2) what Silverstein calls
“pure indexes” whose signaling value is basically relational, indexical, in that
they possess no propositional content. These indexical signs refer to, or rather,
index a context; this signaling process does not depend on any kind of logical
entailment, but on “conventionalised associations between co-occurring clus-
ters of signs and contextual frames” (Gumperz 1995:118).
One set of pure indexical signs Gumperz is particularly concerned with is
that of contextualization cues, “verbal and non-verbal metalinguistic signs
that serve to retrieve the context-bound presuppositions in terms of which
component messages are interpreted” (Gumperz 1996c:379). These cues op-
erate principally at the level of prosody, paralinguistic signs, code choice and
choice of lexical expressions or formulaic sequences (Gumperz 1993), their
meaning being established through interaction and not by the assignment of
context-free semantic content. In this way, they convey limitations on inter-
pretation, for example by cueing “the implicatures or lines of practical reason-
ing that enter into the interpretation of a particular word or expression”
(Gumperz 1992b:50).
The background knowledge necessary for conversational inference is
made available by means of such contextualization cues, which “when pro-
cessed in co-occurrence with symbolic grammatical and lexical signs [serve]
to construct the contextual ground for situated interpretation, and thereby
[aVect] how constituent messages are understood” (Gumperz 1999b:461).
Gumperz’s approach to the analysis of discursive practices is very much
focussed on how this interaction between indexically-linked signals, on the
one hand, and symbolic signs, together with the sequential organization
of speech exchanges, cultural and other background information, on
the other, forms everyday social action, that is to say, goal-oriented commu-
nicative action.
3. Research method
The Wrst case study to be examined here, the job center interview, reXects very
clearly Gumperz’s ever-present interest in intercultural communication, and
demonstrates the extent to which interactional sociolinguistic analysis can
reveal the stated and unstated presuppositions that enter into interactions of
this kind.
As mentioned previously, the interview, aimed at obtaining information
about the candidate’s past work experience, was recorded in the Midlands as
part of a three-year project on intercultural communication. The interviewer
was a native speaker of the English spoken in that area, and the interviewee a
South Asian living locally.13 As one of a number of such situations, “compared
systematically with a range of other encounters both within and outside the
particular institutional context” (Gumperz and Berenz 1993:93), the analysis
of this interview reveals the interviewer’s preoccupation with Wnding out
exactly what the candidate does, the actual work he performs. With this
knowledge, she is able to judge the kind of technical skills involved and
John J. Gumperz: A review 91
The second case study concerns a rape trial cross-examination. The purpose
of cross-examining, at least in Anglo/American court proceedings, is to re-
veal, through questioning, inconsistencies in the witness’ testimony with re-
gard to topics covered in the direct examination. In other words, the aim is to
‘deconstruct’ the direct examination, and this in fact is what the following
extract deals with.
The two main roles in this particular cross-examination are that of the
defence counsel (C), who is intent on conveying (even though indirectly) the
fact that the victim was actually a consenting participant in the rape attack,
and that of the accuser (W), who claims that she was the unsuspecting victim
of unprovoked aggression. The verbal interaction between these two oppo-
nents can only be fully understood, however, when analyzed in terms of the
adversary situation seen as a whole. In fact, a cross-examination like this can
be seen as a typical Bahktinian situation: the dialogic exchanges between the
counselor and the victim are directed to moving the interaction forward in the
traditional conversation analytic way of turn-by-turn sequences, but, at the
same time, they have the function of persuading the judge and jury (in
Bahktin’s terms, the “supersubject”) to consider their opposite points of view.
According to Gumperz, if the transcript of this cross-examination testi-
mony is studied purely in terms of turn-by-turn sequential analysis, “just
concentrating on regularities of sequential organization across speech ex-
changes” (Gumperz 1992a:231), an incomplete picture of the interactive situ-
John J. Gumperz: A review 95
were *all talking,” and, even more clearly, “you went to a *bar? in *Boston”
with “it’s a *club.” In each of these cases, there is a falling prosodic contour
with the accent being placed towards the end of the utterance. By instinctively
copying this contour in such a way that she is isochronous with these and
other key elements of the counsel’s speech, the victim is able to incorporate
them unobtrusively into her own argument and metaphorically construct her
own version of the story. In so doing, she seems to manage “to neutralize the
counsel’s implicatures and to convey her own perspective on the facts without
appearing overly uncooperative” (Gumperz 1995:109).
3.2.1 Summary
To summarize, throughout this extract of a cross-examination testimony, the
impression of victimhood is conveyed partly by linguistic means, but to a
much greater extent indexically, by means of a process of indirect inferencing.
It is only when this use of indexicality is considered within the context, in
terms of the social factors that govern the conduct of this particular situation,
that it takes on its real signiWcance; in fact, only in these conditions is it
defensible as an essential part of the process of conversational inference.
Notes
1. This paper was originally inspired by a talk that John Gumperz gave at the 1995 IADA
Round Table on Dialogue Analysis held in Bologna, Italy. It has subsequently been revised
to present an up-to-date picture of Gumperz’s recent work in Interactional Sociolin-
guistics.
2. “Naturally organized activities” are typically face-to-face encounters possessing some
structure recognizable in everyday life (for example, casual chat, job interviews, court
interrogations etc.).
3. Gumperz refers to these mechanisms and patterns of conversational management as
“principles of conversing” (Gumperz 1992c:304).
4. Gumperz, Aulakh and Kaltman 1982, Gumperz, Kaltman and O’Connor 1984, cited in
Gumperz 1992a:232–233.
5. Cf. Tsui’s work on three-part exchanges (Tsui 1989, 1994).
6. Descriptions of this approach can be found in: Silverstein 1992; Mertz and Parmentier
1985; Lucy 1993.
7. For a slightly diVerent and more detailed version of Gumperz’s analysis of these data,
see Gumperz 1992c, 1995.
8. For the construct “typiWcation”, see Schutz 1971, cited in Gumperz 1982:22.
9. For a deWnition of cultural knowledge, see Clark 1992; for a discussion of the diVerent
types of knowledge constituting cultural knowledge, see Gumperz 1995:105.
10. For fuller details on Gumperz’s interpretive level of analysis, see Gumperz and Berenz
1993:92–94.
11. For the description of a case study based on this particular participant observation
experience, see Gumperz 1982.
12. Interactional texts are described by Silverstein as “transcripts that account for all the
communicatively signiWcant verbal and non-verbal signs perceived” (Silverstein 1992, in
Gumperz 2001). For a full description of the transcribing of conversational exchanges, see
Gumperz and Berenz 1993.
13. For a more detailed analysis of this job interview extract, see Gumperz 1995:110–114.
14. Accenting or stress placement is thought to be “crucial in signalling coreference and
alerting interlocutors to what is expected by way of an answer” (Gumperz 1992c:323).
15. Cf. Gumperz’s (2001) analysis of a job interview with an electrician; also in this case,
the interviewers’ indirect accenting strategies are not fully understood by the South Asian
candidate.
100 Language and interaction
16. For a detailed comparison of Indian and Western English prosody, see Gumperz
1982:119–129.
17. For the description of a telephone number test devised to ascertain the occurrence of
these diVerences in stress patterns and the extent to which they constitute a natural
characteristic of South Asian English, see Gumperz 1982:122–123.
18. In some cases, failure by non-native speakers to recognize and/or produce English
stress patterns can also be grammatically motivated, as, for example, in speakers of Hindi,
who convey the idea of emphasis in part by the discourse particle ‘hi’ (Gumperz, Aulakh
and Kaltman 1982; see also discussion in Roberts et al. 1992:92).
19. Cf. Tyler’s (1995) study of miscommunication between a Korean tutor and a student
who is a native speaker of American English. In this case, in order not to appear rude, it is
the Korean tutor who plays down his knowledge of a particular skill.
20. It is important to note, however, that these ‘criticisms’ are not directly attributable to
participants’ personal feelings or beliefs, but are instead “situated reactions” provoked by
“indexical ties between verbal form and context that function metapragmatically to create
interpretive eVects” (Gumperz 1996c:400–401).
21. In his talk on the concept of communicative competence in culturally complex societ-
ies, Gumperz proposed a deWnition of communicative competence as “the ability to create
and maintain communicative involvement” (Gumperz 1996a).
22. Cf. Gumperz 1999a for the construction of a cultural defence in a court case involving
a Northern California Indian.
23. Transcription conventions are taken from Drew 1992. Gumperz has added ‘*’ to
indicate syllable stress.
24. Gumperz is following the deWnition of isochrony advocated in studies on prosody as
an interactive phenomenon (Couper-Kuhlen and Auer 1991, Uhmann 1996).
25. For a comparison of the concepts of communicative and interactional competence,
see He and Young 1998.
“Interactive practices” are deWned as “structured moments of face-to-face interaction [...]
whereby individuals come together to create, articulate, and manage their collective
histories via the use of sociohistorically deWned and valued resources” (Hall 1995:207–
208).
26. For example, see Gumperz’s analysis of shared inferencing between two sisters in an
informal coVee shop conversation (Gumperz 2000 and 2002).
27. For example, see the studies Gumperz has carried out with others in bilingual, “coop-
erative learning” classrooms to investigate the peer group inferential processes students
rely on to establish collaborative interaction with each other (Gumperz and Field 1995;
Gumperz, Cook-Gumperz and Szymanski 1999).
John J. Gumperz: A review 101
References
Brown, Penelope and Levinson, Stephen C. 1978. “Universals in language usage: Polite-
ness phenomena”. In Questions and Politeness: Strategies in Social Interaction, E.N.
Goody (ed.), 56–311. Cambridge: Cambridge University Press.
Clark, Herbert H. 1992. Arenas of Language Use. Chicago: University of Chicago Press.
Couper-Kuhlen, Elizabeth and Auer, Peter. 1991. “On the contextualizing function of
speech rhythm in conversation”. In Levels of Linguistic Adaptation, J. Verschueren
(ed.), 1–18. Amsterdam and Philadelphia: John Benjamins.
Drew, Paul. 1992. “Contested evidence in courtroom cross-examination: The case of a
trial for rape”. In Talk at Work: Interaction in Institutional Settings, P. Drew and J.
Heritage (eds.), 470–520. New York: Cambridge University Press.
GoVman, Erving. 1976. “Replies and responses”. Language in Society 5:257–313.
Goodwin, Charles and Heritage, John. 1990. “Conversation analysis”. Annual Review of
Anthropology 19:283–307.
Gumperz, John J. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
Gumperz, John J. 1990. “Conversational cooperation in social perspective”. Parasession
on the Legacy of Grice. Berkeley Linguistic Society 16:429–441.
Gumperz, John J. 1992a. “Contextualization and understanding”. In Rethinking Context:
Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 229–252.
Cambridge: Cambridge University Press.
Gumperz, John J. 1992b. “Contextualization revisited”. In The Contextualization of Lan-
guage, P. Auer and A. di Luzio (eds.), 39–53. Amsterdam and Philadelphia: John
Benjamins.
Gumperz, John J. 1992c. “Interviewing in intercultural situations”. In Talk at Work:
Interaction in Institutional Settings, P. Drew and J. Heritage (eds.), 302–327. New
York: Cambridge University Press.
Gumperz, John J. 1993. “Culture and conversational inference”. In The Role of Theory in
Language Description, W.A. Foley (ed.), 193–214. Berlin and New York: Mouton de
Gruyter.
Gumperz, John J. 1994. “Foreword”. In Crosstalk and Culture in Sino-American Commu-
nication, L.W.L. Young, xiii-xix. Cambridge: Cambridge University Press.
Gumperz, John J. 1995. “Mutual inferencing in conversation”. In Mutualities in Dialogue,
I. Marková et al. (eds.), 101–123. Cambridge: Cambridge University Press.
Gumperz, John J. 1996a. “Communicative competence in culturally complex societies”.
Paper presented at invited colloquium on Creating Communicative Competence at
1996 Conference on Discourse Communities, Chicago: American Association for Ap-
plied Linguistics.
Gumperz, John J. 1996b. “On teaching language in its sociocultural context”. In Social
Interaction, Social Context, and Language. Essays in Honor of Susan Ervin-Tripp, D.I.
Slobin et al. (eds.), 469–480. Hillsdale, NJ: Lawrence Erlbaum.
Gumperz, John J. 1996c. “The linguistic and cultural relativity of conversational infer-
ence”. In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 374–
406. Cambridge: Cambridge University Press.
102 Language and interaction
Gumperz, John J. 1997. “On the interactional bases of speech community membership”.
In Towards a Social Science of Language. Papers in Honor of William Labov, Vol. II:
Social Interaction and Discourse Studies, G.R. Guy et al. (eds.), 183–203. Amsterdam
and Philadelphia: John Benjamins.
Gumperz, John J. 1999a. “Culture in the cultural defense”. Proceedings of the Sixth Annual
Symposium about Language and Society (1998), SALSA VI:115–132. Austin: Depart-
ment of Linguistics, University of Texas.
Gumperz, John J. 1999b. “On interactional sociolinguistic method”. In Talk, Work and
Institutional Order, S. Sarangi and C. Roberts (eds.), 453–471. Berlin and New York:
Mouton de Gruyter.
Gumperz, John J. 2000. “Inference”. Journal of Linguistic Anthropology 9 (1–2):131–133.
Gumperz, John J. 2001. “Interactional Sociolinguistics: A personal perspective”. In The
Handbook of Discourse Analysis, D. SchiVrin, D. Tannen and H.E. Hamilton (eds.),
215–228. Oxford: Basil Blackwell.
Gumperz, John J. 2002. “Sharing common ground”. In Soziale Welten und Kommuni-
kative Stile: Festschrift für Werner Kallmeyer zum 60. Geburtstag, I. Keim and W.
Schütte (eds.), 47–56. Tübingen: Gunter Narr.
Gumperz, John J. and Berenz, Norine. 1993. “Transcribing conversational exchanges”. In
Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and
M.D. Lampert (eds.), 91–121. Hillsdale, NJ: Lawrence Erlbaum.
Gumperz, John J. and Field, Margaret. 1995. “Children’s discourse and inferential prac-
tices in cooperative learning”. Discourse Processes 19:133–147.
Gumperz, John J. and Roberts, Celia. 1991. “Understanding in intercultural encounters”.
In The Pragmatics of Intercultural and International Communication, J. Blommaert
and J. Verschueren (eds.), 51–90. Amsterdam and Philadelphia: John Benjamins.
Gumperz, John J. and Toribio, Almeida J. 1999. “Codeswitching”. In The MIT Encyclope-
dia of the Cognitive Sciences, R.A.Wilson and F.C. Keil (eds.), 118–120. Cambridge,
MA: The MIT Press.
Gumperz, John J., Aulakh, Gurinder and Kaltman, Hannah. 1982. “Thematic structure
and progression in discourse”. In Language and Social Identity, J.J. Gumperz (ed.),
22–56. New York: Cambridge University Press.
Gumperz, John J., Cook-Gumperz, Jenny and Szymanski, Margaret. 1999. Collaborative
Practices in Bilingual Cooperative Learning Classrooms. CREDE Research Report 7.
Santa Cruz, CA and Washington DC: Center for Research on Education, Diversity
and Excellence.
Gumperz, John J., Kaltman, Hannah and O’Connor, Mary C. 1984. “Cohesion in spoken
and written discourse: Ethnic style and the transition to literacy”. In Coherence in
Spoken and Written Discourse [Advances in Discourse Processes XII], D. Tannen
(ed.), 3–19. Norwood, NJ: Ablex.
Hall, Joan K. 1995. “(Re)creating our worlds with words: A sociohistorical perspective of
face-to-face interaction”. Applied Linguistics 16:206–232.
He, Agnes W. and Young, Richard. 1998. “Language proWciency interviews: A discourse
approach”. In Talking and Testing: Discourse Approaches to the Assessment of Oral
John J. Gumperz: A review 103
ProWciency, R. Young and A.W. He (eds.), 1–24. Amsterdam and Philadelphia: John
Benjamins.
Jacoby, Sally and Ochs, Elinor. 1995. “Co-construction: An introduction”. Research on
Language and Social Interaction 28 (3):171–183.
Lucy, John A. (ed.). 1993. ReXexive Language: Reported Speech and Metapragmatics.
Cambridge: Cambridge University Press.
Mertz, Elizabeth and Parmentier, Richard J. 1985. Semiotic Mediation. Orlando: Aca-
demic Press.
Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This
volume.
Roberts, Celia and Sarangi, Srikant. 1999. “Introduction: Revisiting diVerent analytic
frameworks”. In Talk, Work and Institutional Order, S. Sarangi and C. Roberts (eds.),
389–400. Berlin and New York: Mouton de Gruyter.
Roberts, Celia, Davies, Evelyn and Jupp, Tom. 1992. Language and Discrimination: A
Study of Communication in Multi-Ethnic Workplaces. London: Longman.
SchegloV, Emanuel A., JeVerson, Gail and Sacks, Harvey. 1977. “The preference for self-
correction in the organization of repair in conversation”. Language 53:361–382.
SchiVrin, Deborah. 1994. Approaches to Discourse. Oxford: Basil Blackwell.
SchiVrin, Deborah. 1996. “Interactional sociolinguistics”. In Sociolinguistics and Language
Teaching, S.L. McKay and N.H. Hornberger (eds.), 307–328. Cambridge: Cambridge
University Press.
Schutz, Alfred. 1971. Collected Papers II: Studies in Social Theory. 3rd printing. The
Hague: NijhoV.
Selting, Margret. 1991. Prosodie im Gespräch. Habilitationsschrift, Universität Oldenburg.
Selting, Margret. 1995. Prosodie im Gespräch. Aspekte einer Interaktionalen Phonologie der
Konversation. Tübingen: Niemeyer.
Silverstein, Michael. 1992. “The indeterminacy of contextualization: When is enough
enough?” In The Contextualization of Language, P. Auer and A. Di Luzio (eds.), 55–
76. Amsterdam and Philadelphia: John Benjamins.
Tsui, Amy B.M. 1989. “Beyond the adjacency pair”. Language in Society 18 (4):545–564.
Tsui, Amy B.M. 1994. English Conversation. Oxford: Oxford University Press.
Tyler, Andrea. 1995. “The coconstruction of cross-cultural miscommunication. ConXicts
in perception, negotiation, and enactment of participant role and status”. Studies in
Second Language Acquisition 17(2):129–152.
Uhmann, Susanne. 1996. “On rhythm in everyday German conversation: Beat clashes in
assessment utterances”. In Prosody in Conversation: Interactional Studies, E. Couper-
Kuhlen and M. Selting (eds.), 303–365. Cambridge: Cambridge University Press.
104 Language and interaction
Response essay 105
Chapter 8
Response essay
John J. Gumperz
Let me begin with a little more background on the issues and Wndings that Wrst
motivated my research on interaction. As Levinson (this volume: 31) puts it in
discussing my approach to contextualization, “Readers will not understand
his work if they view it just as the study of conversation” or, for that matter,
“interaction”, I would add. When I replied to Prevignano’s initial question,
stating that my interests are both theoretical and applied, I meant to suggest
that I did not set out to study interaction in the abstract. What perhaps I did
not make clear was how my current theoretical thinking about language and
communication builds directly on earlier ethnographic Weld studies that led
me to conclude that I needed to learn more about interaction in order to deal
with the questions I had set out to resolve.
I have always had a strong interest in linguistic and cultural diversity, and
in particular in questions such as: How are sociocultural boundaries repre-
sented in language? How does this relationship aVect human understanding
and especially people’s ability to convey information, persuade others and
carry on their everyday aVairs? What is its import for our understanding of
how social institutions work?
Before responding to speciWc comments, I will brieXy describe, or, more
appropriately perhaps, reconstruct, how such interests led me to concentrate
on a theory of interaction. When I Wrst embarked on Weld research, the
nineteenth century assumption that human populations come naturally di-
vided into distinct, bounded groups was still widely accepted. It was assumed
that each of such units can be treated as an island unto itself, with its own
language or dialect and culture, which must be analyzed in their own terms.
Communication across group boundaries was said to be problematic, and
linguistic diversity was seen as potentially a major obstacle to societal consen-
sus. Yet, apart from abstract discussions about translation as the process of
106 Language and interaction
is said, when, and by whom in the course of such encounters, seemed to count
as an equally important factor.
The Norwegian study yielded additional conWrmation for the interaction
hypothesis (Blom and Gumperz 1972). The ideology of egalitarianism here
was so strong that residents refused to acknowledge the existence of sig-
niWcant class diVerences in the locality. And indeed it soon became evident
that income and occupational status distinctions were not clearly reXected in
talk. Nevertheless, it was possible to detect signiWcant diVerences, this time
not at the level of phonology or syntax, but in patterns of language use and
particularly in code-switching between the local dialect and the regionally
dominant variety of Standard Norwegian. The diVerences in language use
distinguished those who, by virtue of their occupation, maintained regular
and close relationships with speakers in the surrounding small towns and in
the more highly urbanized South of the country, from other residents whose
signiWcant contacts were conWned largely to local ones. Once more, density
and quality of interpersonal contacts turned out to be more signiWcant than
social categorizations.
Furthermore, as Wndings from research in previously little-known lin-
guistic areas became better understood, it also became evident that interac-
tion-based forces of linguistic convergence and diVerentiation do not
necessarily respect language boundaries. Throughout the known world, both
in large nation states, as well as in less urbanized regions, where small, sepa-
rate, localized language/culture units live among majority populations, regu-
lar and frequent translocal, interpersonal contacts set oV pressure for change.
Over time, these pressures tend to blur and ultimately eliminate established
lexico-grammatical distinctions (Gumperz 1996; Gomez-Imbert 1996).
A third ethnographic study of a multilingual village at the Indo-Aryan/
Dravidian language border largely supports this interpretation. In this Central
Indian site, the dominant group of landowners spoke a dialect of Kannada, a
Dravidian language, while Indo-Aryan Marathi was the oYcial administrative
language in the surrounding region and the medium of schooling. A sig-
niWcant minority population of artisan and untouchable castes, moreover,
spoke dialects of Urdu (Indo-Aryan) or Telugu (Dravidian). We were struck
by the frequent code-switching observed in everyday village talk. For example,
just as the small town residents in Norway had alternately used Standard
Norwegian and local dialect forms, people in this multilingual village readily
switched between the historically-unrelated Marathi or Urdu and Kannada,
Response essay 109
often in the course of a single brief encounter. And they did so for similar
communicative ends.
When we compared transcripts of everyday talk, moreover, we found that
texts in the local dialects of Urdu and Kannada could be described by a single
underlying set of surface syntactic rules. What local peoples referred to as two
separate languages were analytically distinct, for the most part, only at the
level of word morphology and lexicon, so that translation from one local
variety to the other was largely a matter of word-for-word substitution. In
other words, for the speakers involved in these situations, code-switching was
grammatically not more onerous than dialect-switching in the Norwegian
case. Apparently, centuries of co-residence in the same locality had brought
about large-scale structural convergence among local speech varieties. Yet, the
changes had stopped short of complete creolization. If comprehensibility is
not an issue, how is it that such structurally homologous varieties maintain
themselves in this relatively small population in the face of long-term close
contact? As the code-switching evidence shows, it is not because people
cannot understand each other. We are dealing with a socially-motivated lin-
guistic ideology of diVerentiation, a set of local norms similar to those found
in the North Indian situation, which keep speakers from taking over the
others’ modes of speaking.
The picture of variability that emerges from these Weld studies is quite
diVerent from the conventional one. Common practices that accept taken-
for-granted notions of community, explaining language usage in relation to a
presumably uniform, community-wide system of beliefs and values, that is,
treating each as a self-contained island, are seriously in need of revision.
Whereas truly isolated, homogenous communities have by now all but disap-
peared, as the pace of population migration continues to grow, multilingual or
multidialectal communicative economies like those I have sketched above are
rapidly becoming the rule rather than the exception. It would be more fruitful
therefore to turn to a new base metaphor for sociolinguistic research, a meta-
phor that incorporates diversity into the research design. One might for
example focus on a set of local formations marked by translocal economic,
political, religious and other similar networks of relationships. All this is not
to claim that communities do not have a basis in linguistic reality. I am
arguing that communities must be seen as ideological constructs. Whereas
communicative practices rely on common ground, that is, shared assump-
tions to achieve understanding, ideologies are by their very nature contested.
110 Language and interaction
Communities on one deWnition are treated as the oYcial units for much
governmental and educational planning, but how these communities are
deWned is always open to question, and cannot be taken for granted. Ulti-
mately, communicative practices that are part of a daily reality may have an
inXuence on governmental decisions. I am arguing that the relationships
between ideological constructs and practices are the most useful point of entry
for an investigation of how societal level forces can aVect and be aVected by
everyday talk.
In my earlier work, I used the term “linguistic repertoire” to refer to the
speaking practices of social collectivities, arguing, as di Luzio (this volume: 2–
3) points out, that repertoires are describable in terms of sets of functionally-
diversiWed speech varieties. I found the notion of repertoire useful because it
served to call attention both to diVerences among constituent varieties and to
commonalities due to convergence. To show, in other words, that we are
dealing with what, for want of a better term, we can call “communicative
economies”, that have their own local and translocal forms of social and
linguistic organization. Recent, more detailed work on everyday communica-
tion, however, has led me to doubt that it is useful to assume that discourse
level diversity can be explained in terms of alternation among sharply-
bounded speech varieties. I will return to this point later.
The ethnographic insights that initially seemed most in need of explana-
tion were the following. First, the surprising lack of correspondence between
people’s perceptions of boundaries and the linguistic nature of the boundary
markers. In none of the three Weld situations did we Wnd a direct one-to-one
relationship between what I originally termed “socially”, but would now call
“ideologically” grounded language/dialect distinctions, on the one hand, and
units of language, deWned by grammatical distance alone, on the other. Both
in multilingual and in multidialectal settings, what from a purely linguistic
perspective may count as minor distinctions can often, for largely ideological
reasons, attain great social import as badges of identity. If structure alone
cannot explain such communicative facts, how do we look at verbal interac-
tion to Wnd explanations? Secondly, people readily communicated across
linguistic boundaries. What locals refer to as language or dialect distinctions
do not by themselves prevent understanding, or as such limit interaction. In
the early sites, as well as in later European and U.S. Weld work, local speech
varieties formed part of a single communicative economy, where, depending
on circumstances, switching among distinct languages, dialects or even styles
Response essay 111
Practices (1996), which presents detailed arguments to support the need for
the above distinction.
My approach to semiotic phenomena is broadly in line with Michael
Silverstein’s Peircian semiotic functionalism. Silverstein argues, among other
things, that contextualization cues are a class of indexicals (1992); that is to
say, they function via vector-like cognitive processes radiating out in certain
directions from a known origin, but have no deWnite terminus (Silverstein
1993). Yet, in spite of its inherent indeterminacy, indexicality, as most of
today’s linguistic anthropologists will agree, is crucial for anyone concerned
with discursive representations of culture. Among other things, indexicality
suggests an explanation for what I referred to above as the lack of Wt between
grammatical distance and ideologically-based language distinctions. What
indexes do is act as Xags, cues or reminders to listeners to search their memory
for possible alternative ways of explaining or framing what they hear or
otherwise perceive or recall. That is to say, to arrive at explanations that make
coherent sense of what is going on in the situation at hand. There is thus no
generally agreed-upon, stable meaning relationship between the indexical
sign and any speciWc explanation. The interpretations are always highly con-
text-speciWc in that they depend on aspects of the communicative process that
resist strict formalization.
A key aspect of Peircian theory, not suYciently foregrounded in the
currently available literature, is that signs have meaning only by virtue of
being taken to stand for an object by some interpreter (Lucy 1993). This lays
the groundwork for an integrated semiotic framework, capable of dealing with
both the linguistic and socio-cultural aspects of everyday communication
within a single analytical perspective. Saussurean denotational meanings must
ultimately always be seen as what they are, abstracted from situated practice in
accordance with procedures that set aside the very signaling processes on
which our assessment of the situated interpretation of communicative intent
rests. Useful as they are for comparative grammatical analyses and historical
reconstruction, denotational meanings cannot by themselves explain every-
day human action, partly because they do not account for discourse-level
signaling processes. Once we look at verbal communication as relying on both
symbolic and indexical signs, we can see that human interpreters become
central to the interpretive process. That is to say, the interpreters’ communi-
cative background, their ability to retrieve background knowledge while en-
gaged in an interaction, the power relations aVecting access to relevant
114 Language and interaction
knowledge, and the ideological social and cultural constraints that aVect
understanding, all become integral constituents of the communicative pro-
cess. Although I have not explicitly said so before, I basically share this
perspective in my attempt to develop ways of dealing with interaction that
account for universal, as well as culturally and situationally-speciWc, aspects
of discourse.
Turning now to the speciWc comments, Levinson, Eerdmans and, to
some extent, Prevignano, all expand on my responses in the discussion.
Levinson explores the notion of contextualization from a more formal, that
is, more linguistically-oriented, pragmatic perspective. His discussion of
“context importation and context invocation” is useful and illuminating in
that it cites a number of examples from the linguistic literature to suggest
that the notion of contextualization cue is not as far removed as one might
think from more familiar linguistic pragmatic processes. One might say that
contextualization applies a pragmatic perspective to discourse analysis. But
recall that, especially in my current work, I focus on events or situations of
speaking where participants are treated as engaging in discursive practices,
rather than speaking particular languages. Participants in an event engage in
communicative practices, where verbal expressions, indexical signs like
code-switching, prosody or non-verbal signs, and other signaling modes
may all achieve similar communicative eVects. It is this notion of discursive
or communicative practice, which, in accord with Hanks (1996), I take to be
distinct from language as such, that is relevant here. It enables us to achieve
the broader, more inclusive level of semiotic analysis, not tied to the speciWcs
of particular linguistic signs, which we need to treat code-switching, non-
verbal signs, dress and other signaling modes as contributing to the interpre-
tation process. Along with Hanks, GoVman, conversational analysts and
others, I think of discursive practice as a separate level of analysis and would
argue that Levinson, illuminating as his points are, takes a more language-
oriented perspective; while, from this perspective, the distinction between
diVerent kinds of contextualization makes sense, it is less important at the
level of communicative practice.
Finally, as to what Levinson (this volume: 38) calls the “ineVability” of
understanding, I agree that my answers to Prevignano’s “persistent” questions
as to how to deal with the ambiguity that arises for the analyst as a result of the
complexities of the interpretive process seemed somewhat evasive. Let me try
to be more speciWc and rephrase Levinson’s second question (this volume: 38)
Response essay 115
as follows: What gives the analyst the right to say, “B perceives that A said this,
but he thinks A intended to convey that”? As pointed out before, I do not
claim that Interactional Sociolinguistic (hereafter, IS) analysis can provide
deWnite answers to such questions. Since situated understanding ultimately
always builds on indexical signs, we cannot avoid ambiguity. But then, as
Prevignano (this volume: 17) points out, even properly socialized natives can
understand diVerent things by the same utterances. In fact, we all know
political discourse, advertising, committee negotiations and similar genres
rely on ambiguities for rhetorical eVect (Cook-Gumperz and Gumperz 1994).
Interpretation, moreover, is not a unitary assessment. Typically, inferences
are interactively negotiated over several turns at speaking and such negotia-
tions rarely yield uncontestable understandings. As I see it, the IS analyst’s
task is to reconstruct the understanding process, relying on a varied array of
formally distinct sources of knowledge both at the level of content and form.
Let me turn once more to examples from my published writings to illus-
trate how I view the interpretive process: (i) in the cross-examination ex-
change described by Eerdmans (this volume:94–97), we understand what is
intended by virtue of the fact that we can check our local inferences against
widely-known background presuppositions about what goes on in a cross-
examination and what we can expect the individual participants to have
intended to achieve by means of their talk; (ii) take the case of the graduate
student from Discourse Strategies which I referred to in the discussion
(Prevignano and di Luzio, this volume:19–20), who responded, “Ahma git
me a gig,” when I asked him to see me in my oYce. I was at Wrst puzzled by
what seemed like an unexpectedly jarring code-switch to Black English, which
did not Wt into my initial understanding of what had transpired. It was only
when I compared my interpretation of the incident with comments of other
black, as well as white, students, that I arrived at a reasonable interpretation.
An interpretation, moreover, that Wts in quite well with what I later learned
about African American communicative practices in my work with racially-
mixed school populations; (iii) in a conversational exchange among American
Asian graduate students (Gumperz 1992b:48–9), A asks B what his brother’s
major in college is. When B answers “ah, bio*chem?”, A replies “ ...... *da::(ng),
.. the *bio, huh?”. And when B conWrms, he goes on with “we Asians man! I’m
telling you.” Although when I Wrst looked at the exchange, I did not quite
understand what the talk was about, it was clear from the way the two partici-
pants responded to each other and jointly went about developing a shared
116 Language and interaction
theme that they understood each other. A few turns later in the exchange, it
appeared that what the group was concerned with was the failure of Asian
professionals to take leadership roles in public aVairs or in other technical
Welds; (4) in the Wnal example (Gumperz 1996:388V), an older South Asian
student (D) came to see a lecturer (L) in an adult education center to ask about
enrolling in a new course in a nearby community college. When the lecturer
told him that, although she had participated in planning the course, she was
not involved with student admission, he directly contradicted her, arguing
that she did have a role in the selection process. He kept on with this for
several minutes, totally contradicting nearly all of her statements, and seem-
ingly oblivious of the fact that she was becoming more and more annoyed.
Taken at its face value, D’s behavior seemed at Wrst quite inexplicable both to
me and to the two native South Asian assistants who had helped with the
transcription. It was only when we stepped back from the details of the
exchange, trying to imagine what someone like D might have been seeking to
accomplish with his repeated denials, asking ourselves why he would want to
oVend someone he was asking help from, that we came up with a likely
solution. The assistants pointed out that D sounded as if he was pleading. And
as they said this, I was reminded of an encounter with an older man in a North
Indian village who asked me to help his son, a college student, obtain an
assistantship in the United States. When I tried to tell him that, as a junior
professor, I did not have the right connections, he replied “you can do every-
thing,” using an expression remarkably similar, i.e. intertextually-related, to
some of the statements D employed.
What is analyzed here are events made up of a set of utterances or speech
exchanges and the surrounding circumstances relevant to their interpretation.
The aim is to arrive at interpretations that a) account for what transpires in the
event as a whole and b) Wt into what we know independently about local
ideologies of language and interpersonal relations. The discussion highlights
the functioning of shared background knowledge or common ground in the
interpretive process. Ethnographic observation is clearly important as the
means by which the analyst acquires the knowledge of background presuppo-
sitions and of the indexical conventions that participants tend to rely on in
deriving what they assume is intended, i.e. Levinson’s ‘that’, from what they
perceive as ‘this’. But what I am referring to here is a type of ethnography of
communication that focusses on what people do in speciWc types of situations,
how they react to others and how they talk about it. Along with other
Response essay 117
[at any one point in an interaction], and on which they base their responses”
(Gumperz 1982:153). Given this deWnition, we cannot take understanding for
granted and must assume that the relation between adjacent turns is always
mediated by an evaluative assessment as a precondition for the production of
a succeeding turn. How constituent moves are assessed by participants is
crucial to the production of conversational exchanges. The object of IS analy-
sis is thus not a ready-made sequence of turns at speaking. From an interac-
tional sociolinguistic perspective, exchanges are best represented as sequences
such as the following: ‘move … evaluative assessment … countermove …
evaluative assessment … further move’, and so on. That is to say, any ex-
change depends on participants’ interpretive acts at any one of the many
transition points in the exchange. Apart from the usual semantic processes,
moreover, ideology, which in turn closely reXects societal power relationships,
also enters into the assessments. In all these matters, IS analysis clearly diVers
from conversation analysis.
Eerdmans’ two examples (this volume: 90V) illustrate these matters. In the
second case, where background assumptions are shared, each of the two
principals can rely on shared knowledge of indexical conventions to indirectly
convey conXicting accounts of what transpired. In the Wrst example, the job
center interview, the two participants diVer in background knowledge and are
unsuccessful in establishing a productive exchange, with the result that the
interviewer is unable to make a fair assessment of the applicant’s abilities.
Because of the prevailing ideology of minorization, and what are in fact
diVerences in background knowledge, the applicant’s abilities are disvalued
and his career prospects damaged. This one exchange clearly reveals how
macro-societal stereotypes can aVect our assessments of what transpires in
such encounters.
Ballim’s brief comments are helpful because they point out speciWc areas
of application where the interactional sociolinguistic approach may be useful.
The comments highlight areas of interest where IS and Natural Language
Processing converge. Among other things, the two share an interest in situ-
ated interpretation as it aVects participants’ perception of context. But as an
anthropologist, I am inclined to place more emphasis on shared traditions and
how they aVect speaking practices. For instance, building on Wndings from
work on intercultural communication, I have tried to isolate shared interpre-
tive principles or maxims demonstrably grounded in the typiWcations that
arise from participation in speciWc networks of relationships. Such maxims
Response essay 119
tion of shared linguistic and social knowledge is needed before we can resolve
this issue (see also Gumperz 1997).
Towards the beginning of his discussion of context and context-indepen-
dent meanings, Thibault refers to some of my statements that he interprets as
contrasting “pure indexicals with context-independent or propositional
meaning.” Perhaps my use of the term “propositional” and “context-free” is
confusing, but that is not what I intended to say. I agree with Thibault that all
naturally-occurring semiotic forms are indexical. To the extent that we can
talk about propositional content and context-independent meanings, we refer
to forms or interpretive assessments that have been extracted from some
context by analysts for comparative or other analytical purposes. I should
have made that clearer.
But other sections of Thibault’s comments reXect fundamental diVer-
ences in basic assumptions with respect to those on which IS rests. For
example, he argues that communicative practices have to be studied in rela-
tion to a community’s resource systems. These are described in terms of arrays
of semiotic modalities, including grammar, prosody, non-verbal signs etc., in
fact pretty much the same types of signs that IS treats as entering into the
inferential process. But whereas for IS the shared inferencing of human actors
is the locus of interpretive assessment and ultimately understanding (or
“meaning-making”, in Thibault’s terms), Thibault argues that meaning
construal is achieved by a conjunction of mutually contextualizing modalities.
In other words, meaning, on this view, resides in language, whereas IS follows
the linguistic anthropological tradition in arguing that it is humans who
construe meaning (Lucy 1993). Both approaches have overcome the limita-
tions of traditional approaches that see only denotation as properly linguistic.
But beyond that, I have problems with Thibault’s position. How does he
deWne a community in today’s conditions of ever-increasing diversity? How
does he account for the human factor in interpretation, particularly its ideo-
logical character and limited access to knowledge? Through IS analysis it is
possible, for example, to gain insights into how communicative barriers arise;
what it is that brings about their disappearance; how diversity can be studied
via communicative practices in ways that do not rely on a priori assumptions
about group or community membership; how interpretation reXects past
communicative experience. I believe that I have reason to claim that IS has, at
the least, made signiWcant steps towards the solutions of the problems I set
myself and I am not sure how Thibault would deal with such issues. But then
122 Language and interaction
the two approaches reXect diVerent interests and cannot be judged by their
inability to deal with matters they were not designed to deal with.
Elsewhere in this volume, I have already partially answered Prevignano’s
questions about what makes interaction possible and about the validity of
inference-based interpretive assessments. Let me brieXy recap: I see interac-
tion as a turn-by-turn process, where the production of each turn, including
the opening move, is preceded by an assessment that is simultaneously prag-
matic and meta-pragmatic. Interaction always presupposes conversational
involvement, so that the maintenance of involvement depends, among other
things, on the extent to which knowledge of linguistic, that is, symbolic and
indexical signs, and underlying background knowledge are shared.
Both interactants and analysts are thus always faced with ambiguities. But
the resulting interpretive diYculties seem less problematic if, instead of look-
ing at interpretation as a one-shot assessment, we think of it as an ongoing
interactive process where inferences are constantly revised and attuned as
circumstances change.
As to the question of minims and maxims, Prevignano (this volume: 63),
if I interpret him correctly, is comparing two scholarly positions, the Gricean
tradition, which postulates universal conversational principles or maxims at
the level of content, and a second one that he associates with Gumperz. The
latter argues that interpretive assessment also depends on context-speciWc
aspects of surface discursive form that do not ordinarily enter into other types
of interpretive analysis. I Wnd Prevignano’s discussion interesting because he
seems to be attempting to deal with issues of appropriate speech that many
pragmaticists and sociolinguists tend to account for by means of the much
criticized normative rules of appropriateness, of the form: “In situation A, do
or use Y”, where the term “situation” refers to extra-communicatively avail-
able sets of circumstances. The two traditions he alludes to both take a basi-
cally semiotic position on what makes for rhetorically eVective talk,
attempting to capture regularities of discursive practice without reifying them
in terms of a priori rules. The two diVer in that whereas the Wrst tradition
concentrates on content, the second argues that interpretation rests on par-
ticipants’ ability to relate form to content both to decode propositional mean-
ing and to derive contextually-grounded hypotheses about presuppositions
and possible entailments that aVect understanding. My insights into the com-
municative import of linguistic form derive from work in dialectology and
intercultural communication, where what from a structural perspective
Response essay 123
seemed like insigniWcant phonetic and prosodic processes, can have great
communicative import. I believe that to account for the facts of discourse-
level interpretive assessments, we need to treat discursive or communicative
practice as a separate level of analysis.
Prevignano brings up the issue of metacommunication in discussing the
validity of analysts’ models or descriptions. Let me point out, however, that
one of the main reasons linguistic anthropologists, building on Voloshinov’s
discussion in Marxism and the Philosophy of Language (1973), have begun to
concentrate on meta-communicative phenomena is to avoid their prede-
cessors’ reliance on analytically unjustiWed, dichotomous distinctions like
‘language and thought’, ‘language and culture’, and the like. Linguistic anthro-
pologists have turned to metacommunication as a strategy for avoiding these
dichotomies. In so doing, they open up new possibilities as well as empirical
methods for dealing with these issues. The reason for these new approaches is
the realization that the way people talk about talk provides at least an indirect
indication of what people might think about what they hear. In other words,
metacommunicative phenomena provide us with benchmarks for evaluating
analysts’ interpretations. The distinctions between explicitly metacommu-
nicative signs and indirect indexicals such as contextualization cues is, ana-
lytically, of great importance. The fact that the former are readily subject to
conscious manipulation, whereas the latter are automatically produced below
the level of consciousness, opens up interesting possibilities for interpretive
analysis, as several recent articles suggest (Gumperz 1999; Cook-Gumperz
and Gumperz 1996).
Finally, as to my current interests, at the moment I am working on a book
to be entitled Language in Social Theory, which reexamines Wndings from my
Weldwork over the years in the light of recent theoretical developments, ex-
panding, among other things, on the ideas expressed here. In addition, I am
about to Wnish a case study, Communicative Competence as Communicative
Practice, which reviews the notion of communicative competence in the light
of what we have learned about discourse since the seventies. I am also cooper-
ating with Marco Jacquemet on an edited volume of ethnographically-based
articles that will pick up where Gumperz and Hymes’ Ethnography of Commu-
nication (1972) left oV.
124 Language and interaction
Note
1. See also the article by Du Bois et al. in the same volume (1993), and the more recent, more
comprehensive, “Gesprächsanalytisches Transkriptionssystem (GAT)” (Selting et al.,
1998).
References
Chapter 9
Paul J. Thibault
which these forms realize. This means that both the semantic categories and
the lexicogrammatical forms which realize them are discrete. That is, the
categories and the forms which realize them are deWned by contrasts with
other discrete forms and categories in relation to the paradigmatic systems
and the structures whereby language is theorized as a system of contrasting
terms in which their respective values are deWned by the place of each term in
the overall system. For example, consider the clause complex, When tides are
very large, Wsh in the estuaries for golden snapper, trevally, salmon, cod and the
elusive barramundi. The second clause in this complex is an instance of the
grammatical category ‘imperative mood’. Imperative mood is usually seen as a
discrete type-category which corresponds with other mood categories such as
declarative, interrogative, exclamative moods in the mood system of English
grammar. The noun tides also illustrates the same point. This noun is plural
rather than singular, as indicated by the morphemic suYx -s. The noun is
either plural or singular; the grammar does not recognize continuously vary-
ing degrees of plurality or singularity between these two categorial dis-
tinctions. This is a discrete typological-categorial distinction which the
lexicogrammar of English realizes through, for example, the presence or ab-
sence of morphemes such as the one mentioned here. English grammar has no
resources for specifying topological-continuous variation between the seman-
tic distinctions [singular] and [plural].
Furthermore, the nouns tides, estuaries, snapper, trevally, salmon, cod
and barramundi specify discrete type-categories of the schematic category
Thing, which is prototypically realized by nouns in English. Thus, a given
Wsh which the fortunate angler catches in the estuaries around Darwin Har-
bour is likely to be categorized as an instance of the semantic type-category
represented by the nouns snapper, trevally, salmon, cod and barramundi.
These words refer to discrete categories of Wsh, just as the lexicogrammatical
forms qua exponents of the word class ‘noun’ are themselves discrete: the
noun snapper names the semantic category [snapper] rather than the cat-
egories [salmon] or [barramundi]. Nor does this noun name continuously
varying degrees of ‘snapperness’ or ‘salmonness’ or ‘barramundiness’ which
lie somewhere between these discrete distinctions. The typological mode of
semiosis is certainly a prominent characteristic of language. Nevertheless,
language is not exclusively organized according to typological-categorial
principles. As the observations made by Gumperz and Levinson on
contextualization cues show, language also exhibits topological-continuous
Body dynamics, social meaning-making, and scale heterogeneity 131
categorial phase space whereby semiosis cascades and ramiWes across space-
time scales not tied to the here-now scale of the organism’s material
interactivity with its immediate environment (Lemke 2000:194; Thibault
2000). Whereas the topological mode is more tightly cross-coupled to the
material processes of the body, the typological mode predisposes the linking
of a particular meaning-making event to space-time scales beyond the here-
now scale of the organism’s material interactivity with its immediate environ-
ment. This does not mean that the typological mode transcends the
topological mode. The point is that language-in-use, rather than language as
abstract system, operates on the basis of “mixed-mode” semiosis such that it
combines both topological and typological principles as central to its contex-
tual functioning.
The mapping of semiotic functions onto diverse organizational scales has
dynamical implications for the emergence of meaning, viz. the emergence of a
new level in the scalar hierarchy of semiotic functions can only occur, both
phylogenetically and ontogenetically, if and only if a new level in the hierarchy
of semiotic interpretance emerges. Each higher level in the hierarchy has
multiple realizability at lower levels. For example, it is the many-to-one map-
ping of topological-continuous variety at the N–1 level of the continuous
topological Xux of the phenomena of experience onto the N level of
lexicogrammar that allows for the Wltering of N–1 non-criterial Xuctuations at
the N level. Furthermore, discrete lexicogrammatical forms at this level can be
re-construed at the N+1 level as having an increased variety of meanings,
depending on the ways in which constraints on the N level of lexicogrammar
are taken up and responded to on the basis of their connections to a culture’s
intertextual networks on the more global, longer term N+1 level of the SI (see
Lemke 1999; Thibault, forthcoming).
Following Lemke’s Principle of Alternation, we see that the semiotic trans-
formation of continuous topological variation in the Xux of perceptual and
other phenomena of experience to discrete lexicogrammatical forms and of
these to continuous semantic meanings, in the progression from level N–1 to
level N+1, constitutes what Lemke deWnes as “a semiotic transformation of the
information content of lower levels as signs for higher levels, allowing many-
to-one classiWcations and one-to-many context-dependent reinterpretations”
(1999: 9). It is the dynamics of the semiotic transformations mapped onto the
scalar hierarchy that allows for the emergence of the symbolic possibilities of
Body dynamics, social meaning-making, and scale heterogeneity 135
language, and hence its capacity to form contextual links with many diverse
space-time scales in a given act of meaning-making.
vidual organisms or the locally perceived goals and intentions of social actors.
Rather, the behaviors of the individual organisms involved are entrained to
higher-scalar ecosocial patterns.
Furthermore, the organization of the higher-scalar level cannot be de-
scribed in terms of the organization of entities and processes existing on
lower scalar levels. For example, Cowley (1998) has shown how the interper-
sonal attunement of individuals to each other in conversational events can
depend on physical properties of utterances such as the rate at which partici-
pants vibrate their vocal chords as they tune in to and synchronize their
voice with variables such as Xuctuations in fundamental pitch of the other’s
voice (1998:549). Such sensori-motor events take place on a temporal scale
that is much faster than the rising and falling tones that are recognized as
meaningful contrasts in intonation. Intonational contrasts are perceivable
and are salient on a slower temporal scale than that of the centiseconds
characteristic of the micro-temporal scale of vocal chord vibration. Sono-
graph records show, as Cowley (1998:550–551) documents, that participants
in conversation tune into and act on each other at this physical acoustic level
of, for example, pitch Xuctuation. The diVerent temporal scales of the enti-
ties and processes on the lower scalar sensori-motor level as compared to
that of the higher-scalar level where contrasts in intonation are salient shows
how, in speaking, participants both act on and respond to each other on the
basis of voice dynamics existing on potentially many diVerent temporal
scales which implicate physical-material (bodily) processes not normally
considered to be intrinsic to linguistic form, but which are acted on by
participants as meaningful components of the act of conversing with others.
The process whereby participants converge on a particular fundamental
pitch frequency is an example of the embodied ways in which interactants
tune into each other in ways recognized as interpersonally salient.
In other words, the very fast, lower scalar vocal dynamics of individual
participants are entrained to higher-scalar rhythmic and harmonic patterns at
the level of the conversational event. In this way, the sensori-motor behavior
of the individual participants spontaneously coheres to higher scalar patterns
which have to do with, for example, relational dynamics existing on quite
diVerent temporal scales. Sensori-motor dynamics such as vibrating vocal
chords are biological initiating conditions that self-organize into patterned
behavior at the level of the particular conversational event. In turn, the spon-
taneous emergence of such patterns is entrained to still higher level patterns
138 Language and interaction
such as, for example, the historical-biographical unfolding of the social rela-
tions among the participants to any given conversational event. The relational
dynamics of this level provide boundary conditions which regulate and en-
train the dynamics of the lower levels to its patterns. In the examples analyzed
by Cowley (1998), the higher scalar level of the social group to which indi-
vidual dynamics are entrained is, in the Wrst instance, that of the family,
comprising the two parents and their daughter, which is a social unit having
its own dynamical patterns characteristic of that level and their history. In
postulating such patterns and their history on their particular temporal scale
— months and years as compared to centiseconds! — the analyst is, in eVect,
postulating an entity — a traditional one in this case — which has the possibil-
ity of becoming part of some theoretical discourse whereby diVerent scalar
levels come clearly into view. In this case, the focal level is that of the conversa-
tional event. The scalar hierarchy of relevant levels with reference to this focal
level can be described as a three-level system, each with its own processes and
entities unique to that level. Thus:
N+1: relational dynamics of family as social group and the history of its
evolving patterns
N: the conversational event comprising a number of individual partici-
pants whose behaviors cohere to the higher-scalar patterns of the overall
event
N–1: the sensori-motor dynamics of individual organisms as biological
initiating conditions
the ways in which the bodily potential of the organism is entrained to larger-
scale patterns of social interaction. In other words, a large number of degrees
of freedom of our bodily dynamics on level N–1 of the three-level hierarchy
described in Section 3 is entrained to a more restricted set of possibilities in
virtue of the conventional indexical associations on level N through our
attempts to reciprocally adapt to the behavior of others in our interactive
encounters with them on level N+1. In this way, we see how the conventional
indexical associations discussed by Gumperz result from our ongoing eVorts
to adapt our own bodily dynamics to the bodily dynamics of other individuals
with whom we engage in interactive, goal-directed or purposeful activity.
Such mutually modulated body dynamics constitute emergent collective pat-
terns which, in turn, exert their own eVects and constraints on the activities of
individuals.
Gumperz describes such indexical associations as “context-dependent”
because their contextual eVects arise in immediate situations in the here-now
of a given interaction. This fact is evidenced by the ways in which the given
(indexical) sign is dependent on speciWc sensori-motor cues for its interpreta-
tion. Indexicals of the sort discussed by Gumperz under the rubric of “con-
textualization cues” allow the individual to interact with other individuals on,
roughly speaking, the space-time scale of their shared perceptual purview. In
this case, the principle of semiotic mediation at work here means that the
here-now body dynamics — e.g. the particular vocal or manual-brachial
gesture — which take place on the scale of seconds or fractions of seconds in
the real-time of our perceptual-motor activity, indexes contextual associa-
tions which are commensurate with the scale on which the gesture qua indexi-
cal sign itself is articulated. This is the scale on which perceptual-motor
activity and its environmental eVects is both enacted and observed through
the perceptual systems of the body in interaction with its ecosocial environ-
ment. In other words, semiotic mediation by pure indexicals already entails
scales beyond the organismic one per se at the same time that it is constrained
by higher-scalar constraints in the form of the conventional associations that
have emerged between indexical signs and their respective contextual values.
Figure 1 models the complex, non-linear, inter-level constraints that operate
here across a number of diVerent scalar levels, ranging from the organismic
scale of the human body to the higher-scalar, socio-cultural conventions
involved.
Now, the fact that contextualization cues are conventional associations
Body dynamics, social meaning-making, and scale heterogeneity 141
already implies that they have a history and that this history has both indi-
vidual and social dimensions. That is, a given use of a conventional indexical
association implies scales of dynamical complexity in the given ecosocial
system which go beyond the momentary use of a particular contextualization
cue in some context. How did the user come to use the particular con-
textualization cue? In relation to which social practices in the ecosocial sys-
tem? How did the given cue come to be part of the particular user’s repertoire
of meaning-making resources? In relation to which social networks and their
characteristic bodily dispositions (cf. habitus)? Contextualization cues such as
rhythm and prosodies have their basis in the neurophysiological substrate and
perceptual-motor activity of the organism, yet their eVects extend beyond the
organismic scale of the individual and into the ecosocial environment of other
social agents for whom they are potentially meaningful. The questions posed
here point to the fact that contextualization cues qua indexical signs are
embedded in at least two interpenetrating scales which have a history.
First, there is the personal-biographical history of the individual’s trajec-
tory through speciWc social networks. Secondly, there is the history of the
ecosocial system within which individual trajectories are in turn embedded.
Contextualization cues can only have their meaning in relation to both em-
bodied individual memory and the accumulated history of the meaning-
making practices of a given culture or some part of this. Indexical signs
mediate bodily processes so as to amplify their eVects on the ecosocial envi-
ronment, including other individuals, beyond the scale of the micro-level
processes which constitute their bio-physical substrate on the organismic
142 Language and interaction
According to Gumperz:
Sociolinguistic explanation, if it is to be relevant to today’s concerns, cannot
implicitly accept traditional categories of language, culture and society. I believe
that interaction at the level of discursive or communicative practices […] must
be seen as separate from either linguistic or socio-cultural processes. It is consti-
tuted by the interplay of linguistic, social and cultural/ideological forces and
governed or constrained by partly universal and partly local-speciWc organiza-
tional principles.
(Prevignano and di Luzio, this volume: 9–10)
This formulation of discursive practices, on the one hand, and linguistic and
socio-cultural practices, on the other, suggests that socio-cultural groups and
their associated practices have a prior existence with respect to the ways in
which they are constituted, construed and evaluated through language and
other semiotic activities, material processes, and related objects and artifacts.
In my view, any such separation is an analytical, rather than a constitutive, one
(Thibault 1991:229–240). Social formations are neither logically nor onto-
logically prior to the social semiotic and material activities in and through
which meanings are made. Instead, they are constituted by these activities.
Take Gumperz’s example of the Afro-American speaker’s use of the utterance
Ahma git me a gig (Prevignano and di Luzio, this volume: 19–20).
As Gumperz’s analysis shows, the lexicogrammatical and phonological
features which co-pattern in this utterance instantiate semantic patterns
which are typical of register-speciWc semantic relations that occur in the
speech of the Afro-American speakers referred to. Such patterns are charac-
teristic of the discourse community to which these speakers belong. More-
over, it is not simply a question of the speaker talking about a particular topic
in a way which diVers from the ‘academic norm’ of standard American En-
glish referred to by Gumperz. Rather, the speaker activates a small fragment of
a much wider intertextual thematic formation which is recognizable and
interpretable by members of the discourse community in question. It is a
thematic formation which is also typically associated with a particular evalua-
tive orientation or social viewpoint insofar as it implicates an evaluative stance
both towards his interlocutors as well as towards the thematic content of his
utterance. For this reason, the utterance can be seen as the activation of what
Bakhtin (1981) called a particular “social voice”, i.e. a particular social view-
144 Language and interaction
Note
1. I wish to thank Carlo Prevignano for his valuable comments on an earlier version of this
chapter.
References
Bakhtin, Mikhail. 1981. “Discourse in the novel”. In The Dialogic Imagination: Four Essays
by M.M. Bakhtin, C. Emerson and M. Holquist (trans.) M. Holquist (ed.), 259–422.
Austin and London: University of Texas Press.
Bateson, Gregory. 1973 [1972]. “Problems in cetacean and other mammalian communica-
tion”. In Steps to an Ecology of Mind, 334–348. London and New York: Granada.
Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, Mass.: MIT Press.
Cowley, Stephen J. 1998. “Of timing, turn-taking, and conversations”. Journal of
Psycholinguistic Research 27 (5):541–571.
Firth, J.R. 1957 [1948]. “Sounds and prosodies”. In Papers in Linguistics 1934–1951, 121–
138. London and New York: Oxford University Press.
Gumperz, John J. “Response essay”. This volume.
Halliday, M.A.K. 1979. “Modes of meaning and modes of expression: Types of grammati-
cal structure, and their determination by diVerent semantic functions”. In Function
and Context in Linguistic Analysis: A Festschrift for William Haas, D.J. Allerton, E.
Carney and D. Holdcroft (eds.), 57–79. Cambridge and London: Cambridge Univer-
sity Press.
Halliday, M.A.K. 1992. “A systemic interpretation of Peking syllable Wnals”. In Studies in
Systemic Phonology, P. Tench (ed.), 98–121. London and New York: Pinter.
Body dynamics, social meaning-making, and scale heterogeneity 147
Halliday, M.A.K. 2000a. “Phonology past and present: A personal retrospect”. Folia
Linguistica XXXIV (1–2):101–111.
Halliday, M.A.K. 2000b. “Grammar and daily life: Concurrence and complementarity”. In
Functional Approaches to Language, Culture and Cognition, D.G. Lockwood,
P.H. Fries, and J.E. Copeland (eds.), 221–237. Amsterdam and Philadelphia: John
Benjamins.
Lemke, Jay L. 1999. “Opening up closure: Semiotics across scales”. Paper presented at the
conference, Closure: Emergent Organizations and their Dynamics, University of
Ghent, Belgium, May 1999.
Lemke, Jay L. 2000. “Material sign processes and emergent ecosocial organization”. In
Downward Causation: Minds, Bodies and Matter, P. Bøgh Andersen, C. Emmeche, N.
Ole Finnemann, P. Voetmann Christiansen (eds.), 181–213. Aarhus: Aarhus Univer-
sity Press.
Levinson, Stephen C. “Contextualizing ‘contextualization cues’”. This volume
Pike, Kenneth L. 1967. Language in Relation to a UniWed Theory of the Structure of Human
Behavior. Second, revised edition. The Hague and Paris: Mouton.
Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This
volume.
Thibault, Paul J. 1991. Social Semiotics as Praxis. Text, Social Meaning Making, and
Nabokov’s Ada. Minneapolis and Oxford: University of Minnesota Press.
Thibault, Paul J. 2000. “The dialogical integration of the brain in social semiosis: Edelman
and the case for downward causation”. Mind, Culture, and Activity 7 (4):291–311.
Thibault, Paul J. Forthcoming. The Signifying Body: An Ecosocial Semiotic Theory of
Situated Meaning-Making, Embodiment, and Consciousness.
Thibault, Paul J. and Van Leeuwen, Theo. 1996. “Grammar, society, and the speech act:
Renewing the connections”. Journal of Pragmatics 25:561–585.
148 Language and interaction
Continuing the discussion with John J. Gumperz 149
Chapter 10
this means we must see them as abstractions and cannot claim that any one
element or constituent thereof can by itself account for the phenomenological
facts of communicative practice.
Turning to the quote at the end of your question, in using that expression,
I meant to argue against scholars who resort to extra-communicatively
deWned, largely lay categories, like class, gender, ethnicity, professional iden-
tity, community or group membership and the like, to explain or naturalize
what happens in situated interactive exchanges, without demonstrating how
the relationship works. As the preceding paragraph implies, some forms of
social/cultural/ideological knowledge always organize our actions and de-
monstrably aVect CP interpretive assessments. But this is not equivalent to
claiming that participants’ membership in one or another human collectivity
as such is directly reXected in talk. What we are talking about here is how
interpretive presuppositions, ideologically associated with such a collectivity,
aVect interpretation. The nature and workings of such presuppositions must
be investigated through the analysis of context-bound verbal exchanges.
As to “regularities” of communication practice, I believe that these should
ultimately be derived from or related to in-depth analyses of situated encoun-
ters in a variety of settings. Because of the way interaction works, the usual
questionnaire studies are not satisfactory here. I made some informal com-
parisons between South Asian and English interviewing practices in a paper of
mine (Gumperz 1996), but these are observations that might inform ethno-
graphic studies. I don’t know of any systematic research along these lines.
C.L.P.: At the end of your Response essay, you indicate a “distinction [analy-
tically of great importance, I agree] between explicitly metacommunicative
[and consciously manipulable] signs and indirect indexicals [which are pro-
duced below the level of consciousness] such as contextualization cues”. Do
you see today any other diVerence between the former and Gumperzian cues?
J.J.G.: On second thought, it might be more fruitful here to speak of a distinc-
tion between overtly-lexicalized indexicals, on the one hand, and non-
lexicalized or indirect indexicals (including, among others, prosody, rhythm,
tempo, certain phonetic or grammatical variables, along with code- and style-
switching), on the other. That would avoid value-laden terms. Linguistic
anthropologists, following Silverstein (1992, 1993), sometimes use the term
“pure indexicals” to refer to the latter category. That is to say, they are pure in
the sense that they basically function relationally or inter-discursively. Unlike
152 Language and interaction
tures] existing on “larger spatial and temporal scales” has its origin in con-
ventions that arise in interaction and over time have become grammaticalized.
These may then be brought into interaction by participants as part of the
interpretation process. Like other grammatical knowledge, these conventions
are subject to preservation, reaYrmation or change through what happens in
later communicative practice.
I do of course agree that we need to make a distinction between what
I would call “interactional time” or, as you put it, “the space-time scale” of the
interactional exchange and “larger spatial and temporal scales”. But what I
have to say does not really contradict that.
Your remarks concerning my objection to the claim that “language and
other semiotic modalities construe experience” go to the heart of the diVer-
ences between your and the IS approach. Your systemic perspective leads
you to privilege formal semiotic structure as a source of insight into the
nature of linguistic processes. IS, on the other hand, argues that we have no
access to experiential reality, except as it is Wltered through our own or some
other human interpreter’s mind, so that we must look to communicative
practice. This is why I emphasize shared inferencing. The very fact that
interactants maintain conversational involvement over a string of move-
countermove-rejoinder sequences constitutes empirical proof that they
share the presupposed grammatical and indexical knowledge, since the in-
ferential process constitutes the cognitive space where grammatical and in-
dexical knowledge meet. But this does not mean that “language and other
semiotic modalities” existing “on space-time scales” that go beyond the
“shared inferencing of human actors” are not relevant for IS. What IS sug-
gests is that we should not assume there is a one-to-one relationship between
the linguist’s bodies of abstract rules and regularities and what goes into
shared inferencing. The relationship between the two needs to be investi-
gated in specially-designed studies. Until this is done, the two approaches
will remain largely incommensurable.
C.L.P.: Many years ago, you proposed the interaction-oriented notions of
“linguistic repertoire”, “linguistic community”, and later, “social network”.
How do you see these today?
J.J.G.: The notion of “speech community” or “linguistic community” (the two
terms were often used interchangeably, but see my Wnal remarks), has always
been rather loosely employed to refer to the social space in which language is
Continuing the discussion with John J. Gumperz 155
here, as in the study of other human interactional tasks and abilities, is an idea
of interactional intelligence (representable as a cultural set of strategies, tac-
tics, heuristics, default rules, meta-rules, and so on), as has been anticipated
and practiced by your IS research over a long period of time. The construct of
interactional intelligence is necessary, in my opinion, in order to deal with the
study of human, interactionally intelligent conversationalists.
After this long premise, I come to a short question: could you, as a de facto
analyst of interactional intelligence in and between diVerent cultures, give
some suggestions for future research on it?
J.J.G.: Let me restate parts of the question’s preamble in my own terms,
expanding on what I understand the questioner to have said and on argu-
ments I made elsewhere in this volume.
A. As to interaction analysis. Along with other Linguistic Anthropolo-
gists, I assume that we, that is, specialists as well as lay persons, have no direct
access to physical, social or psychic reality, except as it is processed by some
interpreter’s mind. It is this processing, as reXected in the inferences partici-
pants draw on, on the basis of what they believe they see, hear, recall, or
otherwise perceive in speciWc sets of circumstance, that lies at the center of IS
analysis. So that the focus is no longer on word-to-world relationships, but on
the ‘how’ of interpretation, i.e. the procedures we and others rely on to assess
experience.
B. For analytical purposes, it is useful to treat the inferential process as
made up of several stages. In the Wrst of these stages — the most basic one in
the sense that it underlies further, higher-level inferencing — perceptions are
(among other things) chunked, categorized into event types and phrases and
assigned phonetic, prosodic, rhythmic values. This yields what, using a
modiWed version of Silverstein’s (1993:36–8) term, I call an “interactional
text”. As I use the term, the notion of interactional text makes explicit the
assumptions of lay persons and analysts about what they perceive, and at the
same time provides the indexical ground for further interpretations. Unlike
Saussurian analysis where etic constituents of talk are seen as part of the
physical world, an IS approach treats what the preamble calls “interaction
surface and cues” as components of interactional texts and thus as interac-
tively constructed.
C. As to “cross-inferencing and misinteractions” (I prefer the term “inter-
action breakdown” for the latter). My research shows that these are frequently
Continuing the discussion with John J. Gumperz 159
References
Caldas-Coulthard, Carmen R. and Coulthard, Malcolm. 1996. Texts and Practices: Readings
in Critical Discourse Analysis. London and New York: Routledge.
Fairclough, Norman. 1989. Language and Power. [Language in social life series]. London:
Longman.
Ferguson, Charles and Gumperz, John J. (eds.). 1960. Linguistic Diversity in South Asia.
International Journal of American Linguistics 26 (3), special issue.
Gumperz, John J. 1968. “The speech community”. In International Encyclopedia of the
Social Sciences, J. Sills (ed.), 9:381–386. New York: Crowell Collier and Macmillan.
Gumperz, John J. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
Gumperz, John J. 1992. “Contextualization and understanding”. In Rethinking Context:
Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 229–252.
Cambridge: Cambridge University Press.
Gumperz, John J. 1996. “The linguistic and cultural relativity of conversational inference”.
In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 374–406.
Cambridge: Cambridge University Press.
Gumperz, John J. 2002. “Sharing common ground”. In Soziale Welten und Kommuni-
kative Style: Festschrift für Werner Kallmeyer zum 60. Geburtstag, I. Keim and W.
Schütte (eds.), 47–56. Tübingen: Gunter Narr.
Gumperz, John J. “Response essay”. This volume.
Continuing the discussion with John J. Gumperz 161
Gumperz, John J. and Levinson, Stephen C. (eds.), 1996. Rethinking Linguistic Relativity.
Cambridge: Cambridge University Press.
Hanks, William F. 1996. Language and Communicative Practices [Critical Essays in An-
thropology Series]. Boulder, CO: Westview Press.
Irvine, Judith T. 1993. “Insult and responsibility:Verbal abuse in a Wolof village”. In
Responsibility and Evidence in Oral Discourse, J. Hill and J.T. Irvine (eds.), 105–134.
Cambridge: Cambridge University Press.
Jacquemet, Marco. 1996. Credibility in Court: Communicative Practices in the Camorra
Trials. Cambridge: Cambridge University Press.
Leach, Edmund R. 1954. Political Systems of Highland Burma. London: G. Bell and Sons.
Prevignano, Carlo L. “On Gumperz and the minims of interaction”. This volume.
Prevignano, Carlo L. and di Luzio, Aldo. “A discussion with John J. Gumperz”. This
volume.
Silverstein, Michael. 1992. “The indeterminacy of contextualization: When is enough
enough?” In The Contextualization of Language, P. Auer and A. di Luzio, 55–76.
Amsterdam and Philadelphia: John Benjamins.
Silverstein, Michael. 1993. “Metapragmatic discourse and metapragmatic function”. In
ReXexive Language: Reported speech and Metapragmatics, J.A. Lucy (ed.), 33–58.
Cambridge: Cambridge University Press.
Silverstein, Michael. 1996. “Monoglot ‘standard’ in America: Standardization and meta-
phors of linguistic hegemony”. In The Matrix of Language: Contemporary Linguistic
Anthropology, D. Brenneis and R.K.S. Macaulay (eds.), 284–306. Boulder, CO: West-
view Press.
Thibault, Paul J, “Body dynamics, social meaning-making and scale heterogeneity: Re-
considering contextualization cues and language as mixed-mode semiosis”. This
volume.
Wodak, Ruth. 1997. Disorders of Discourse [Real Language Series]. London: Longman.
Young, Linda W.L. 1994. Cross-talk and Culture in Sino-American Communication. Cam-
bridge: Cambridge University Press.
162 Language and interaction
Bio-bibliographical note 163
Bio-bibliographical note
John J. Gumperz was born in 1922 in Germany, and emigrated to the United States in
1939. Currently Professor Emeritus of Anthropology at the University of California,
Berkeley, in 1954 he was awarded his Ph.D. in German Linguistics by the University
of Michigan, after Wrst completing a B.A. in Chemistry at the University of Ohio,
Cincinnati in 1947. His doctoral thesis, which was a study of the Swabian dialect of
third-generation farmers in Washtenaw County, Michigan, marked the beginning of
his interest in sociolinguistics. In this study, he formulated his early thinking on the
relations between social group membership and sociolinguistic processes. This was
the Wrst of a long series of research projects and Weld studies. On completion of his
doctorate, he undertook two years of Weldwork in India. He subsequently founded the
Hindi-Urdu program and was later Chairman of the Center for South and South East
Asia Studies (1968–71), both at Berkeley.
In 1956, Gumperz began collaborating with Charles A. Ferguson when both were
visiting faculty at Deccan College. The ensuing exchange of ideas on linguistic diver-
sity and language development led to their organizing a symposium at the 1958
annual meeting of the American Anthropological Association. The jointly published
volume, Linguistic Diversity in South Asia (1960), represents a landmark in the Weld.
Since 1966, Gumperz has been an active member of the Committee on Socio-
linguistics of the Social Science Research Council. This work led to the publication of
A Field Manual for Cross-Cultural Study of the Acquisition of Communicative Compe-
tence (1967), which has played an important role in the development of cross-cultural
research on sociolinguistic processes. He also collaborated with Dell Hymes, Susan
Ervin-Tripp, Edward T. Hall, Erving GoVman, Charles O. Frake, and William Labov
in the organization of symposia at the spring meeting of the Kroeber Anthropological
Society and the annual meeting of the American Anthropological Association. The
seminal volume, The Ethnography of Communication (1964), edited by Hymes and
Gumperz, was the product of these meetings.
Gumperz Wrst took up the position of Professor in Anthropology at Berkeley in
1965. In 1992, he was awarded a Dr. Hon. Causa by the University of Konstanz in
Germany. He was made Fellow of the American Academy of Arts and Sciences (1992),
the Max Planck Project Group in Cognitive Anthropology in Berlin (1991), Churchill
College, University of Cambridge (1987–88), as well as being a Senior Post-doctoral
Fellow at the National Science Foundation (1961–62). He was LSA Professor at the
Linguistic Institute of America, Georgetown University (1985). He has held Visiting
Professorships at the Universities of Bergen (1982) and Konstanz (1985).
Professor Gumperz’s major research interests lie in the areas of interactional
sociolinguistics, linguistic anthropology, social anthropology, ethnography of com-
munication, cross-cultural and interethnic communication, and bilingualism. His
major publications include Rethinking Linguistic Relativity, co-edited with Stephen
Levinson (Cambridge 1996); Engager la Conversation (Paris 1989); Sociolinguistique
164 Bio-bibliographical note
Subject Index
Author Index
Akinnaso, Niyi 17-18 41, 42, 44, 50, 53, 64, 67, 70, 71, 93, 110,
Auer, Peter 26, 67, 70, 71, 100 n. 24 112, 115, 127, 133, 135, 139, 142, 143,
Aulakh, Gurinder 99 n. 4, 100 n. 18 145, 153
Drew, Paul 27, 89, 100 n. 23
Bakhtin, Mikhail 43, 54, 55, 56, 59 n. 6, 94, Du Bois, John W. 124 n. 1
143, 144 Duranti, Alessandro 33, 63, 72
Ballim, Afzal ix, 66, 81, 118, 119, 120
Barwise, Jon 82 Eerdmans, Susan L. ix, 29 n. 1, n. 3, 114,
Bateson, Gregory 9, 27, 36, 46, 55, 67, 69, 115, 117, 118
129 Ensink, Titus 70
Bazzanella, Carla 69 Erickson, Frederick 67, 68
Berenz, Norine 72, 89, 90, 99 n. 10, 99 n.
12, 120 Fairclough, Norman 159
Berge, Kjell L. 65 Ferguson, Charles 156
Berger, Charles R. 67, 72 Field, Margaret 100 n. 27, 119
Blom, Jan-Petter 31, 66, 72, 108, 112, 156 Fillmore, Charles 13, 32
Bloomfield, Leonard 8, 106 Firth, John R. 131, 133
Blum-Kulka, Shoshana 66 Foppa, Klaus 27
Bolinger, Dwight L. 58 n. 5 Fraser, Colin 68
Bourdieu, Pierre 5, 8
Bradac, James J. 67 Gal, Susan 38 n. 1
Brown, Penelope 38 n. 1, 68, 69 Galilei, Galileo 66
Garfinkel, Harold 8, 10, 11, 19, 59 n. 6, 67,
Cacioppo, John 68 68, 72, 86, 119
Caldas-Coulthard, Carmen 159 Giles, Howard 67, 68
Cassell, Justine 72 Goddard, Cliff 83 n. 2
Castelfranchi, Cristiano 64, 68 Goffman, Erving 8, 10, 13, 14, 19, 27, 56, 59
Chomsky, Noam 136, 149 n. 6, 67, 72, 87, 114
Clark, Herbert H. 72, 99 n. 9, 120 Gomez-Imbert, Elsa 108
Coakes, Elayne 83 n. 2 Goodwin, Charles 72, 86
Cook-Gumperz, Jenny 70, 100 n. 27, 115, Goodwin, Marjorie H. 72
119, 123, 160 Greif, Irene 83 n. 2
Coulthard, Malcolm 159 Grice, H. Paul viii, 26, 32, 35, 63, 72, 86,
Couper-Kuhlen, Elizabeth 70, 73 n. 3, 100 117
n. 24 Grosjean, Michèle 64
Cowley, Stephen J. 137, 138 Grosz, Barbara 82
Grundy, Peter 72
Dascalu Jinga, Laurentia 69 Gumperz, John J. vii, viii, ix, x, xi, 1-6, 7-10,
Di Luzio, Aldo vii, viii, 14, 15, 26, 28, 31, 10-14, 15, 17-18, 19-21, 22-23, 24-25, 26-
170 Author index
27, 27-28, 29 n. 1, 29 n. 2, 29 n. 3, 31, 32, 67, 69, 70, 72, 105, 112, 114, 116, 127,
33, 35, 37, 38, 41, 42, 43, 44, 45, 46, 49, 130, 156
50, 53, 54, 55, 56, 57, 58, 59 n. 6, 63-73, Lucy, John A. 113, 121
79, 81, 82, 85-100, 107-111, 112, 115, Lyons, John 83 n. 2
116, 118, 121, 123, 127, 130, 131, 133,
135, 138, 139, 140, 143, 144, 145, 149, Martin, James R. 43
150, 151, 152, 153, 154, 155, 156 Mertz, Elizabeth 99 n. 6
Müller, Frank 70
Hall, Joan K. 100 n. 25 Murray, Stephen O. 73 n. 2
Halliday, M.A.K. 43, 47, 52, 58 n. 5, 131,
132, 138 Nöth, Winfried 49
Hanks, William F. 9, 28, 112, 114, 150
He, Agnes W. 100 n. 25 Ochs, Elinor 97
Herasimchuk, Eleanor 71 O’Connor, Mary C. 99 n. 4
Heritage, John 86 Orletti, Franca 68
Horton, William S. 83 n. 2
Hymes, Dell 3, 72, 111, 123, 149 Pallotta, Vincenzo 81
Parmentier, Richard J. 99 n. 6
Ionescu-Ruxandoiu, Liliana 15, 29 n. 2 Parret, Herman vii
Irvine, Judith 160 Peirce, Charles S. 10, 16, 49, 59 n. 6, 69, 87,
113
Jacquemet, Marco 23, 123, 152 Perry, John 82
Jacoby, Sally 97 Petty, Richard 68
Jakobson, Roman O. 87 Pike, Kenneth L. 58 n. 2, 132, 150
Jefferson, Gail 22, 95 Prevignano, Carlo L. vii, viii, x, 7, 10, 15,
16, 17, 18, 19, 21, 22, 23, 24, 25, 26, 27,
Kaal, Bertie xi 28, 41, 42, 44, 50, 53, 64, 66, 69, 71, 72, 73
Kallmeyer, Werner 150 n. 2, 93, 105, 112, 114, 115, 120, 122, 127,
Kaltman, Hannah 99 n. 4, 100 n. 18 133, 135, 139, 142, 143, 145, 146 n. 1,
Kendon, Adam 67, 68 149, 151, 152, 153, 154, 157-158, 159
Kerbrat-Orecchioni Catherine 70
Keysar, Boaz 83 n. 2 Raskin, Victor 72
Kiesler, Sara 83 n. 2 Rawlins, William K. 69
Koerner, E.K.F. 66 Roberts, Celia 97, 98, 100 n. 18
Rumsey, Alan 45, 58 n. 4
Labov, William 37
Lakoff, Robin 32 Sacks, Harvey 32, 67, 72, 95
Langacker, Ronald W. 48, 51 Sapir, Edward 8, 43
Lave, Jean 120 Sarangi, Srikant 72, 98
Leach, Edmund 155 Saussure, Ferdinand de 8, 10, 43, 158
Leech, Geoffrey viii, 63 Schank, Roger C. 83
Lemke, Jay 41, 43, 44, 49, 51, 52, 53, 54, 55, Schegloff, Emanuel A. 72, 95
129, 133, 134, 136, 138, 139 Schiffrin, Deborah 23, 68, 72, 85, 86, 98
Levinson, Stephen C. viii, 31, 52, 55, 64, 65, Schutz, Alfred 15, 99 n. 8
Author index 171