Sociology and ideas about language
Having things done: language and discourse
We examined what interactionist sociologies, and especially symbolic interactionism and ethnomethodology, had to say about the relationship between language and social action. Here, we shall concern ourselves with more contemporary developments, directing attention to various aspects of what has come to be called 'the linguistic turn' in social theory. The emphasis here is on the discursive nature of social life, the importance of the linguistic mediation of social relations and the role of language in human knowledge and the understanding of social action. Clearly, a full account of 'the linguistic turn' in social theory (which may also be regarded as a social/discursive turn in linguistics- see Coupland et 01. 2001) would be out of place here; rather the concern is with the views on the links between language and social action proffered by various perspectives.
The 'linguistic turn' in social theory is often associated with two other currents within social science: postmodernism and post-structuralism. Much has been written about these overlapping, but distinct, 'movements', but their significance for our argument rests on the criticisms they advance of the interactionist, phenomenological approaches. These centred on the privileging in such approaches of the human subject as both the origin of meaning and as the end point of analysis. These criticisms owed much to the work of writers such as Levi-Strauss, Barthes and Althusser in the 1950s, 60s and 70s who were, in different ways, responsible for establishing structuralism as an intellectual current in Western social theory. (See, for example, Anderson 1976 and Thompson 1978.)
Broadly speaking, structuralism took issue with those traditions of thought which emphasized the freedom of the individual and which regarded meanings as the creation of human consciousness. Against this view, structuralists pointed, on the one hand, to the powerful social and material forces (such as the distribution of wealth and political power) arrayed against the individual and severely curtailing their freedom, and on the other, to the equally powerful forces which shaped and structured social meanings. In the latter case, rather than seeing meanings as an emanation of individual consciousness- in which human beings creatively devise various ways of interpreting the world- structuralists saw them as the product of social discourses. Through the social organization of symbolic representations, social discourses confront people with historically limited and specific repertoires of meanings. Of course, the practices that people engage in influence social discourses in their turn, but practices and discourses are social, rather than individual, phenomena. In this way, structuralism 'decentres the human subject', and entails a radical break with the sociological perspectives considered earlier, with their emphasis on the importance of some notion of subjective understanding.
Post-structuralists make this break yet more decisive by insisting that the very idea of a subject is itself a social construction produced through social discourses. These discourses position people within particular sets of practices, already constructed through other social discourses, and it is these practices which provide the basis for the construction of human subjectivity. So the notion of the rational and autonomous human being that has been at the heart of much Western philosophy since the seventeenth century is extinguished; the autonomy has been much exaggerated and the rationality is the product of social discourses about rationality. These discourses structure the human psyche in complex and contradictory ways, often imperfectly understood (both by the individuals concerned as well as those seeking to analyze or interpret their behavior) and poorly managed. This fragmentary self-spells not only the end of any unitary or mechanical relation between human subjects and the class or power positions they occupy- this is the 'post' element of post-structuralism- but also brings to the fore the role oflanguage (and discourse more generally) in the constitution of the self and in the social relations of everyday life.
The pivotal role assigned to discourse by those writers influenced by post-structuralism requires a closer exploration of the term. We shall do this first by looking at the notion of discourse as it is developed and employed in the work of Foucault, possibly the most influential wielder of a discourse approach, and second by taking one or two examples of the use of a concept of discourse in contemporary social research. Our central argument will be that the notion of dis course employed in the Foucauldian tradition conflates practices and knowledge. That is to say, while it may be true that power effects are inscribed in practices, the two are not identical (Layder 1994).
Foucault's work has had a major influence within the social and human sciences, and this makes hazardous the enterprise of summary and interpretation. Furthermore, the status of the conception of discourse shifts in Foucault's work. In his earlier work (Foucault 1974) the focus is on 'discursive formations', on discourse as rules for constituting areas of knowledge; in the 'genealogical studies', such as Discipline and Punish (Foucault 1977), the emphasis is on the relation between knowledge and power; and towards the end of his life, Foucault became more concerned with the issue of 'ethics', with how the subject actively constitutes herself (Foucault 1978). Talk of a Foucauldian 'perspective' might be regarded in this light as premature. Bearing this in mind, the account that follows is narrowly concerned with the implications of Foucault's uses of a notion of discourse for an understanding of the relation between language and social action. There are, of course, many others who have considered his ideas within a much broader context. (See, for example, Barratt 1991, Carter 1996, Coupland et. al. 2001, Dean 1994, Dreyfus and Rabinow 1982, Fairclough 1993, Gordon 1980, Hoy 1986, Miller 1993, Owen 1994, Rabinow 1984, Simons 1995.)
Despite the various modifications in Foucault's use of the notion of discourse, we want to suggest that his definition in The Archaeology of Knowledge is compatible with the position we adopt insofar as it recognizes the irreducibility of discourse to either individual speakers or to linguistic structure in the Saussurean langue sense (Foucault 1974). This notion of discourse posits a stratified social ontology comprised of linguistic structure, discourse and individual speech activity, each with their distinctive properties and each irreducible to either of the others. Foucault expresses it thus:
Discourse lies at the level of pure atemporal linguistic 'structure' (langue) and the level of surface speaking (parole): it expresses the historical specificity of what is said and what remains unsaid.
Discourses are composed of signs, but what they do is more than use these signs to designate things. It is this move that renders them irreducible to the language and to speech. It is this 'move' that we must reveal and describe.
As writers such as Pennycook (1994a) and Sarangi (2001) have noted, this view of discourse has considerable potential relevance to applied linguistics. Nevertheless, there is a certain 'openness' about this definition that allows for less moderate interpretations. A discursive formation consists of 'rules of formation' for the particular set of statements that belong to it and in this way discourses both enable and constrain actors and agents through the rules that govern their operation. These rules, to follow Fairclough's (1993: 41-9) useful typology, concern the formation of objects, the formation of enunciative modalities or 'subject positions', the formation of concepts and the formation of strategies. The last two of these refer to the ways in which intra- and inter-discursive relations of complementarity, competition, analogy and so on operate to shape the meanings and possibilities available within a discursive formation. (Note that such relations are possible precisely because of the ontological status of discourse and its irreducibility, although there is little in Foucault's work to justify a strong claim that discursive relations maybe realized independently of social agency.) However, itis the first two sets of rules of formation that have been subject to readings which have pushed Foucauldian dis course in a decisively postmodern direction.
Common to the discursive formation of both objects and of subject positions is Foucault's view of discourse as constitutive, '... as contributing to the production, transformation, and reproduction of the objects ... of social life' (Fairclough 1993: 41). So discourse is not only constitutive, in significant or important ways, of our ideas about reality, it is also productive of the ideas we have about all sorts of social relations and thus about the social practices- of measurement, accounting, examining and imprisoning- that we found upon these ideas.
Besides this constitutive role in the formation of objects, dis courses, on this account, also constitute subjects and their subjectivities. They accomplish this via the subject positions that are associated with discursive activity: for example, 'immigrant' and 'British' are two subject positions constituted by discourses about race and immigration in postwar Britain. Subject positions are historically specific, and therefore liable to change, and are able to exert a conditioning influence on those subjects positioned in particular ways.
Both of these constitutive functions can be interpreted in what we shall call, following Rosenau (1992), 'weak' and 'strong' ways. In their 'weak' sense they have a great deal to offer. In particular, they can draw attention to what Taylor (1985: 234) has termed the expressive aspects of language and the ways in which they serve to'... express/constitute different relations in which we may stand to each other: intimate, formal, official, casual, joking, serious, and so on'. Our reservations about using discourse to analyze the social world arise from the 'strong' interpretations. With these the 'de-centring' of agency becomes the elimination of active social agency in any meaningful sense. The social world is seen as constructed by authorless discourses which themselves become agents; rather than tension between actors, agents and discourses, concretely negotiated in particular historical settings, there are merely discourses constructing objects and human subjects. Consequently, efforts to develop accounts using such a notion of discourse often get into difficulties when it comes to identifying agency in a specific, empirical sense.
One approach to researching social behavior using a notion of discourse is that of 'discursive psychology', and we have taken a study by Wetherell and Potter (1992), concerned with the 'language of racism', as an illustration of the problem. In this study of the attitudes of New Zealand 'whites' towards the Maori population, Wetherell and Potter make several claims about the nature and role of discourse. Most importantly, they argue that discourse is actively constitutive of both social and psychological processes. 'The psycho logical and social field- subjectivity, individuality, social groups and social categories- is constructed, defined and articulated through discourse' (ibid.: 59). Consequently, they place an 'over-riding stress on the study of discourse in action rather than language as an abstract system' (ibid.: 71).
The emphasis on 'discourse in action' conflates social practices with the ways in which such practices are thought about and enunciated through speech. The texts employed in their study- extracts from their interviews with 22 New Zealanders- are interpreted by the authors as instances of 'discourse in action'; others might interpret these texts differently. As we have argued elsewhere, viewing dis course as a practice raises other problems. Two in particular are pertinent here. Firstly, the actual content of discourses becomes secondary to its effects in defining and identifying them. Thus:
Racist discourse ... should be seen as discourse (of whatever content) which has the effect of establishing, sustaining and reinforcing oppressive power relations. ... Racist discourse is discourse which has the effect of categorizing, allocating and dis criminating between certain groups ... it is discourse which justifies, sustains and legitimates ... practices.
On this model, the identification of racist discourse becomes a task of Herculean proportions. How do we establish empirically what the effects of a discourse are? How do we distinguish these from the effects of other discourses? Such questions bring us to the second problem, namely the tension between reducing discourses to practices, which makes difficult analysis of their specific properties and the ways in which these may influence practices, and at the same time needing to render them distinct in order to account for people's decisions to construct certain forms of identity or to utilize certain types of narratives. The status of discourse is indeterminate: it is both a constitutive practice and yet it circulates as a resource:
... identity and fonus of subjectivity which become instantiated in discourse at any given moment should be seen as a sedimentation of past discursive practices. A sense of identity and subjectivity is constructed from the interpretative resources- the stories and narratives of identity- which are available, in circulation, in our culture. (ibid.: 78)
In some ways, this formulation comes close to the emergentist model we are advocating of regarding social and personal identities as products of the engagement of people's consciousness with a World 3 of cultural forms (including 'stories and narratives of identity'). However, the emergentist view entails a consistent analytical distinction between culture, in the form of World 3, and agency, that is what human beings do with culture, precisely so that their mutually influential effects of constraint and enablement may be assessed. It is to those social theorists who have suggested that such a distinction is key to sociological accounts of language.
