Sémantique et Pragmatique. Le tournant des années soixante-dix

 

La distinction et la rupture — une rupture épistémologique — intervenues il y a trente ou quarante ans entre Sémantique et Pragmatique sont fondatrices de notre discipline, l'Anthropologie linguistique, au début des années soixante-dix. C'est à partir de cette distinction que nous pouvons préciser la différence entre l'Ethnolinguistique (une sémantique) et l'Anthropologie linguistique (une pragmatique). C'est sur la base de cette distinction que nous pouvons placer au cœur de notre champ de recherches les problématiques de la Performance et de l'Indexicalité. Voici quelques jalons.

silverstein_theory_1972.pdf Michael Silverstein, “Linguistic Theory: Syntax, Semantics, Pragmatics,”, Annual Review of Anthropology, Vol. 1 (1972), pp. 349-382. Dossier ‘Silverstein’.

grice_conversation.pdf H[erbert] Paul Grice, “Logique et conversation,” Communications, 30, 1979, 57-72. Traduit de l'américain par Frédéric Berthet et Michel Bozon. Dossier ‘Grice’. La version originale en anglais de ce texte classique fut publiée en 1975. C'est le texte mentionné ci-dessous.

morgan_pragmatics.pdf Jerry L. Morgan, “Linguistics: The Relation of Pragmatics to Semantics and Syntax,” Annual Review of Anthropology, Vol. 6 (1977), pp. 57-67. Dossier ‘Pragmatique’.

 

1 / Compétence et Performance au début des années 1970

Dans les années soixante, la démarche épistémologique dominante vise à construire une théorie universelle de la grammaire fondée sur les universaux de la “compétence” linguistique. La performance est exclue du champ de la recherche. L'anthropologie linguistique va naître d'une réhabilitation de la performance. Les textes cités ci-dessous sont tirés des articles de Silverstein (1972) et de Morgan (1977) référencés plus haut.

(Silverstein, 362) What is the relationship of competence to performance? It seems to me that this is a fundamental question needing answer before competence has empirical content. Do the facts explained by a characterization of competence bear any regular relation to performance?

This is the approach taken by Weinreich, Labov and Herzog in their remarkable study (1968) of the synchronic bases of linguistic change, and again /363/ by Labov in his papers on methodology (Labov, W. 1971. Methodology) and on sociolinguistics (Labov, W. 1970. The study of language in its social context. Stud. Gen. 23:30-87)). Here are explored the consequences of accepting, as linguists from Paul to Chomsky have, such a distinction:

Linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogeneous speech community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic) in applying his knowledge of the language in actual performance… Observed use of language or hypothesized dispositions to respond, habits, and so on, may provide evidence as to the nature of this mental reality, but surely cannot constitute the actual subject matters of linguistics (Chomsky, 1965. Aspects of the Theory of Syntax, pp. 3-4).

As was mentioned above, the data for linguistics are constituted by speakers's intuitions about sentences, and so it devolves upon the use of these to establish the structure of linguistic competence. Methodologically, Labov has called this position “the Saussurian paradox: that the social aspect of language can be studied by the theorist asking himself questions, while the individual aspect can only be studied by a social survey” (Methodology, p. 437). The social aspect is the supposed shared norm, Saussure's langue, Chomsky's competence. In fact, it has become clear by simply noting the judgements of grammaticality and ungrammaticality (which are distinct, be it recalled, from acceptability and nonacceptability in performance) made by linguists in their presentations, that these vary so widely as to make it impossible to apply the notion in practice.

Il n'y a aucune raison de croire, en effet, que nos intuitions linguistiques et la grammaire que nous construisons à partir de ces intuitions reflètent fidèlement la structure de la langue. D'ailleurs, les données dont nous disposons, même si elles reflètent directement et fidèlement la compétence d'un locuteur indigène (the native speaker), n'en montrent pas moins une variation d'un locuteur à un autre. Comment décider alors entre ce qui est structural et ce qui ne l'est pas? La rupture épistémologique qui nous conduit à complexifier la compétence en la définissant à travers la performance, c'est la naissance du Variationnisme en linguistique.

(364) To return then to the conditions disputed by Weinreich et al, the first principle which they find abundant evidence against is that any speech community can be homogeneous. This is not in the trivial sense of memory limitations etc, but rather in the discovery that there are very strong statistical regularities in a number of significant points of variation within every speech community studied, and that these regularities are correlated very directly with such factors as age, socioeconomic class, and so forth. Moreover, native speakers control and have intuitions about functional varieties of language; we might ask if it is correct to ignore these in dealing with paraphrase relationships. This kind of “orderly heterogeneity” for them is a prime datum to be accounted for by a theory of competence, inasmuch as it is replicable, for the same linguistic feature, in comparable communities.

(368) This brings up the more general question of variation in language, which Chomsky rules out of the theory of competence, but which Weinreich et al claim is essential to any language. As was mentioned above, variability is patterned, it is observed in every linguistic community, and it is correlative for certain features. It would be conceivable to say, therefore, that this variability /369/ is the result of complex rules of performance which distort an otherwise homogeneous competence within a community. In fact, given the basis for grammaticality in Chomsky's statement quoted above, this follows necessarily.

Cette compétence étendue, qui “est le résultat de règles complexes de performance”, sera l'objet d'une partie de l'anthropologie linguistique développée plus tard par Silverstein sous le nom de métapragmatique.

 

2 / Du sens littéral de la «Proposition» au contexte d'énonciation de la «Phrase»

Morgan part de la distinction que faisait H. Paul Grice entre le sens littéral d’une phrase et ses implicitations ou implicatures quand elle est prononcée dans le cours d’une conversation, pour reformuler cette distinction de façon plus abstraite en opposant la sémantique (le sens) et la pragmatique (les implicatures).

(58) «Stalnaker (1972), motivated by an interest in presupposition and its relation to semantics, gives a broader characterization of pragmatics as “the study of linguistic acts and the contexts in which they are performed.” In his influential program, semantics is the study of “propositions,” not sentences. Sentences are not propositions, but are used to express them. Thus one cannot directly speak of the truth of falseness of a sentence like “they are here now” without constructing from context of utterance particular values for “they,” “here,” and “now” [des déictiques]; the result is a proposition whose truth can be evaluated. Pragmatics, then, is the study of “the ways in which the linguistic context determines the proposition expressed by a given sentence in that context,” and includes the study of speech acts, indexicals, knowledge, beliefs, expectations, and intentions of the speaker and hearer, and other aspects of context that bear on the determination of the proposition expressed by a sentence. On top of this, the term “pragmatics” has come to be used also for the study of meaning implied by the proposition the sentence is used to express. The philosopher Herbert Paul Grice (1975), in an extremely important and influential paper (circulated underground for several years prior to publication), gave an insightful account of aspects of indirectly conveyed meaning (“conversational implicatures”) that cannot be considered part of the literal meaning of the sentence, but are the result of inferences about the speaker’s intentions in saying what he says. Grice’s paper has led to a gradual broadening of interest in matters of context, communication, and intention, so that now the term “pragmatics” is applied even to studies of discourse structure, politeness, and social interaction in conversation. What unites all these apparently disparate areas under the same term is the crucial role in each of inference, in context, about the intentions of the speaker.»

(62) «Grice (1975) makes clear that much more is conveyed in the utterance of a sentence than merely the literal meaning of the sentence. For example… Grice’s example, if someone asks me how a friend is doing in his new job at a bank, and I reply:

“Oh quite well, I think; he likes his colleagues, and he hasn’t been to prison yet,”

then I will probably convey the opinion that the friend’s honesty is open to question, though it would be entirely implausible to attribute that meaning directly to any part of the sentence I uttered.»

(63) «The difficult and controversial issues in this literature revolve around the problem of distinguishing conversational implicature from literal meaning. For example, in saying:

“I took off my clothes and got in bed,”

one usually conveys clearly that disrobing preceded getting into bed. But should this be dealt with as part of the literal meaning of the sentence, or as conversational implicature? One could construct an initially plausible case for either approach. It is clear that semantic and pragmatic analyses compete as accounts of many phenomena, raising questions as to how much of the traditional territory of semantics is to be taken over by pragmatics.»

 

Dès ce moment, dans notre discipline, la pragmatique a pris le pas sur la sémantique. Ce n'est seulement qu'à la fin des années soixante-dix, néanmoins, que, glissant de la philosophie (Grice) vers la stylistique et la théorie littéraire (Jakobson, Bakhtine et le «tournant dialogique»), la littérature spécialisée de l'anthropologie linguistique mettra la problématique des intentions et le concept d'implicature en veilleuse pour privilégier la problématique de l'indexicalité et des concepts tels que la deixis et la métalepse.