From Heritage and Greatbatch, 1991: 103

for a passing reference to Sacks et al’s study of turn-taking procedures in everyday conversation, there is no discussion of conversation analysis. This is a curious omission: Gilbert and Mulkay were trying to establish the need to study discourse as a topic, and to demonstrate the complexity of language use, to a largely sceptical sociological community which had hitherto regarded lan- guage as an unproblematic research resource. Acknowledgement of CA’s find- ings about the action-orientation of utterances, and the level of detail at which turns can be designed, would have been useful in establishing the relevance of their arguments. It would have at the very least supported their claim that lan- guage use can be studied sociologically as a topic in its own right. Discourse analysts, however, are much more explicit about the range of intellectual influ- ences. For example, Potter and Wetherell cite the importance of speech act theory, ethnomethodology and semiotics. The findings from conversation ana- lytic studies are regularly used to illustrate general claims about language, or to strengthen the force of criticisms of traditional social psychological approaches. Their chapter on accounts uses Atkinson and Drew’s 1979 research on courtroom interaction; and their critique of social psychological attempts to study how people categorise themselves into social groups is informed directly by material from Sacks’ early lectures. Edwards’ work on categorisations in everyday speech and the use of scripts in social life is informed, respectively, by Sacks’ work on membership categorisation devices and ethnomethodological arguments about the constitutive nature of rule following in social life Edwards, 1991, 1995a. Focus Although Gilbert and Mulkay’s discourse analysis was forged from a consideration of methodological issues which have a wider relevance in sociol- ogy, their research concerned issues in the sociology of scientific knowledge. This was also true of colleagues who were sympathetic to discourse analysis. This reduced the likelihood that their broader arguments would be influential in the discipline as a whole. Edwards, Potter and Wetherell, however, explored the implications of the variable and constructive properties of language in a variety of sub-topics within social psychology: for example, the study of attitudes, cate- gorisation, social representation theory and theories of the self. This ensured that the analytic method they offered as an alternative to traditional approaches in social psychology enjoyed a wider currency. It also allowed them to identify sev- eral important new lines of empirical inquiry, thus stimulating further research. Repertoires One major difference between the initial work in DA in sociol- ogy of scientific knowledge and the subsequent emergence in social psychology research concerns the use of the concept of repertoires. There is no clear-cut distinction, but it seems reasonable to propose that the investigation of reper- toires was a much more prominent feature of the earlier discourse analytic work. There was discussion of repertoires in Potter and Wetherell’s key DA 66 CONVERSATION ANALYSIS AND DISCOURSE ANALYSIS text; but even here many of the analytic themes which they addressed did not rely on the notion of repertoires. Their chapter on accounts, for example, draws much more heavily from conversation analytic work on courtroom interaction. And Edwards and Potter’s subsequent analysis of attributions in discourse concerning a political dispute focused more on the ways in which versions of events and descriptions of people had been formulated so as to manage imputations of blame and responsibility. In the subsequent develop- ment of discourse analysis, the use of repertoires has been variable. For exam- ple, in studies of the way we use psychological terms, or invoke cognitive or mental phenomena – a branch of discourse analysis which has become known as discursive psychology – there is little use of the concept of repertoires for example, Edwards, 1991, 1994. But in some more ideologically oriented dis- course analysis, analysts still seek to identify the repertoires through which versions of the world are constructed for example, Edley, 2001. Implications: from representations to action What unites these two strands of discourse analysis are their radical methodological implications. Discourse analysis proposes the language we use, and the way we use it, is not determined by, nor anchored in, some set of objective properties of the events to which we refer. Any state of affairs, any event, any person or group of people, yield potentially inexhaustible descriptive possibilities Waismann, 1965; Wittgenstein, 1953. Moreover, we do things with our words: they are not inert representations of social action, nor neutral disclosures which either exhibit the operation of underlying cognitive processes, or from which such processes may be inferred.They are actions, designed for the ‘here-and-now’ of their production. Social scientists are aware of the difficulties in using people’s accounts as data for research, but perhaps for the wrong reason. It has been assumed that such data are either anecdotal, and therefore of little analytic value, or that they are subject to distortion, revision, omission and so on, and thereby unreliable. A variety of methodological techniques have been developed in sociology and social psychology to overcome the perceived shortcomings of discursive data. However, these methodological responses still embody the assumption that lan- guage is a medium of representation, when in fact it is a medium of social action. Unless we recognise this key property of the language through which social and psychological life is conducted, and make it a central feature of our analytic enterprise, sociology and social psychology will rest on fragile foundations. Conversation analysis: continuities and implications Earlier we examined some findings from studies of interaction in particular kinds of work-related or institutional settings. How has this discussion advanced our understanding of the core methodological principles of CA and the significance of its findings? METHOD AND CRITIQUE 67