discourse

discourse, discourse analysis The study of language, its structure, functions, and patterns in use. For Ferdinand de Saussure, language in use (or parole) could not serve as the object of study for linguistics, since as compared to langue (the underlying system of rules), it was individualized, contingent, and therefore intangible. Eventually, however, some of Saussure's successors in linguistics as well as in the wider structuralist tradition did turn their attention to parole, in the hope of discovering, behind it, additional structures to those of langue; structures which, in other words, would facilitate the completion of the analysis of meaning, and so allow semantics to take account of the connotative (secondary or implied) as well as the denotative (intended or explicitly signified) dimension of language.

In the event, the reversal of the privilege accorded by Saussure to the denotative over the connotative became one of the distinguishing characteristics of post-structuralism, and it is the sense given to the term discourse within this body of thought (rather than within linguistics) that has come to exercise a powerful influence in sociology. For this reason, then, discourse analysis in sociology has been more concerned to uncover the large patterning of thought that structures whole texts, rather than the finer patterning that structures sentences, and which concerns linguists.

As Roland Barthes pointed out in the conclusion to his Mythologies (1957), what one is confronted with in parole is a chain of ‘signifiers’ rather than one of ‘signs’. What is more, these signifiers often appear to mean more than is suggested by dictionary definitions. Barthes's suggestion was that, in order to discover what this might be, one has to be able to reconstruct the additional sets of underlying relations that determine the actual use of signifiers in particular contexts. Barthes himself termed these additional sets of relations ‘myths’—a term he and others later rejected because of its negative and economic-reductionist connotations.

It was Michel Foucault who eventually provided a conception of the additional structures that determine language use (and, indeed, although this is far less often acknowledged, of the sociological constraints upon them), which sits happily alongside the positive and non-reductionist conception of the ideological realm that commands wide support today. According to Foucault in his methodological text The Archaeology of Knowledge (1969), these additional structures are made possible by historically produced, loosely structured combinations of concerns, concepts, themes, and types of statement, which he terms ‘discursive formations’. Although such formations are far more loosely structured than the discourses they make possible, they are sufficiently determining to allow the differentiation of connotative structures from one another, for example, of sociology from racism and from the law.

What gives these formations their structuring quality are the particular conditions which made and still make them possible. These ‘rules of formation of a discursive formation’ include, so far as the objects they allow to be addressed are concerned, each of the following: the social or institutional contexts wherein they emerge, most often as the loci or sources of concern of some kind; the social identities of those who have or gain authority to pronounce on such problems and their causes; and the ‘grids of specification’, the intellectual templates so to speak, which are used to separate off the particular objects of concern from the many others with which each is intertwined in reality.

In order to indicate that the discourses produced in such ways add meaning to langue, Foucault describes their joint product not as a sentence, but as ‘statement’. He then defines this as a series of signs which, first, assumes the particular subject position given by the relevant discursive formation; second, projects a certain dynamic on to the set of signifiers that constitute it; and, finally, possesses a definite materialism by virtue of being recognizably different from other statements. A discourse is thus ‘a group of statements insofar as they are made possible by the same discursive formation’.

Despite the formidable nature of the intellectual underpinning made necessary by the counter-intuitive nature of nonrepresentationalist conceptions of social phenomena, and (ironically) its own somewhat opaque language (some idea of which can be gained from the terminology introduced in this entry), discourse analysis is not a horrendously difficult exercise, as Jonathan Potter and Margaret Wetherell make clear in the excellent discussion of its methodology contained in their Discourse and Social Psychology (1987). For an example of an empirical study see David Silverman 's Discourses of Counselling (1997
), a study of the conversations between HIV counsellors and their clients, which draws on the sociological traditions of interactionism, conversation analysis, and ethnomethodology. See also CONNOTATIVE VERSUS DENOTATIVE MEANING; SEMIOLOGY.