What is Critical Discourse Analysis?

We have seen one example above of what Critical Discourse Analysis can look like (Barker/Galasinski 2001). Let me begin again here with an account of what Critical Discourse Analysis is not. It is not the collocation of terms now common in cultural studies research to which I referred above, terms which have lost their histories in the translation into a cultural studies where one suspects that few, any longer, ever actually read the primary texts and research from which they were derived. These terms: representation, intertextuality, genre, narrative, discourse, performativity, corporeality, habitus, to name just a few are all used to support and argue a constructionist paradigm, that social realities and selves/bodies are constituted in discourse, and that social change requires both a deconstruction of the way these categories work and a rewriting of them to produce change. For those who are not linguists, or proponents of critical discourse analysis, this mode of operation with and on texts is seen as being primarily about language and representation, a form of, albeit poststructuralist, 'discourse analysis'.

The problem with the current uses of these tools of analysis is that in fact they have very little connection with the intricacies of language as such: they are macro not micro categories of analysis. They are large categories, identified with equally large chunks of often undeconstructed text. Once nominalised, they can be bandied about as names for things which apparently exist, but the work which would have to be done on the materiality of language to demonstrate that existence in recognisable and replicable linguistic terms is rarely done. Fairclough's 1995 a and b) work on what the linguistics of intertextuality might look like and my own (1997) work on discourses of race and gender and with Kress (Kress/Threadgold 1998) offer examples of what might be necessary here. The terms have also in crucial ways become, in the translation into cultural studies, predominantly tools for the analysis of the textual or the discursive (understood as language and sometimes the visual). I want to suggest that this completely elides the radical potential of the poststructuralist vocabulary as it was framed and contextualised in the texts in which it was first developed.

I have argued elsewhere (Threadgold 2000) that the now linear narrative of the overthrow of structuralism by poststructuralism needs to be rewritten in terms of the discontinuities and anticipations of the very complex histories that are actually involved. Much radical poststructuralist work was anticipated by structuralist linguists: and poststructuralism constructs itself always in relation to the inevitability and centrality of structuralism (Derrida 1972/1982; 1978/1981). There is also no longer a clear distinction between Critical Discourse Analysis (CDA) and poststructuralist theories and concepts, as the latter have increasingly been used to theorise and reposition linguistic work of the kind that locates itself within the CDA paradigm. However, here, as in cultural studies, there has often been a tendency for the incorporation of poststructuralist concepts and analytical terms to be assimilated to the linguistic or the discursive.

I want now to locate CDA historically and then return to the poststructuralist vocabulary in order to try to see what radical moves have been written out of its translation into both cultural studies and CDA itself. I then want to argue that it is not just a question of adding CDA to the current mix available in cultural studies but rather of rethinking both in relation to the work cultural studies and its affiliated disciplines and interdisciplines need to do in the twenty-first century.

CDA has several histories and emerges from a conjunction of different kinds of discursive activity. Critical Linguistics was an approach developed by a group based at the University of East Anglia in the 1970s (Fowler 1981; Fowler/Kress/Hodge/Trew 1979). The group was strongly influenced by the work of the linguist Michael Halliday and his systemic functional grammar although they also used stylistic approaches borrowed from Chomskeyan transformational linguistics and Fowler in particular used the work of Barthes and early French semiotics. They distinguished their approach from sociolinguistics and from the then mainstream paradigm of linguistics (which was Chomskeyan), refusing the oppositions form and content, and system and process or use, insisting on analysing real texts and their relations to real contexts. It is also important here to recognise the influence of Prague School linguistics and semiotics, British structuralist-functionalist anthropology, the work of the educational sociologist Bernstein (whose later (1990) work is specifically on pedagogic discourse) and the much earlier work of the linguists Sapir (1921) and Whorf (1956) in the work of Michael Halliday himself (Halliday 1985). This intertextual and discursive history needs to be thought of as still operating, albeit at a distance, in the work on discourse of this school of linguists.

Critical linguistics was concerned to read the meanings in texts as the realisation of social processes, seeing texts as functioning ideologically and politically in relation to their contexts. This was very much an approach in which discourse was text, but there was too little emphasis on the production and interpretation of texts, a too ready assumption of the transparent relationship between textual features and social meanings and a neglect of discourse as a domain of social struggle or of the ways in which changes in discourse might be related to wider processes of social and cultural change. There was also a typical Marxist top-down view of ideology and power and an emphasis on social structure rather than social action, social reproduction rather than social transformation (Fairclough 1992).

The work of the group that came to be known as social semioticians in Australia in the 1980s and 90s included some of the same people - Halliday, Hasan, Kress and Hodge - and worked with some of the same ideas, but in the Australian context there was a strong influence from poststructuralism and French structuralism and semiotics as well as the work of semioticians like Umberto Eco (1979), and a variety of feminisms, so that what Halliday had called a theory of language as social semiotic became something quite other by the early 1990s. There was a concern to rethink functional linguistics in the light of the work of Foucault on discourse, institutions and power, Bakhtin on heteroglossia and dialogism and the work of psychoanalytic feminisms on the unconscious and the questions of the body and subjectivity. There was a new focus on processes of textual interpretation and production, and a new understanding of the crucial importance of intertextuality and subjectivity in those processes as discursive processes involving struggle and change. What remained of the original theory was the importance of a functional theory of language as a way of grounding discourse analysis in a flexible linguistic analysis and a recognition of the importance of trying to integrate this with radical social and cultural theory. The work of Kress/Hodge (1988), of Kress/Threadgold (1988), of Thibault (1991), of Poynton (1985), the collection edited in 1988 by Birch/ O'Toole , Threadgold (1997) and the collection edited in 2000 by Lee/Poynton give some indication of the range, breadth and difference of this work from the earlier critical linguistics. The special issue of the journal Social Semiotics devoted to feminist poststructuralist work using functional grammar (Vol.3, No.1, 1993) is another example. Schirato/Yell (1996) was the pedagogic text-book version of this work, which was used as a text-book to teach students in education, cultural studies and semiotics in Australia. This text, much earlier than Barker/Galasinski (2001) combined poststructuralist theory and cultural theory with a functional linguistics.

Reflecting on it now, what is interesting about this work, despite the fact that some at least of its proponents were also very active in the National Australian Association of Cultural Studies at the time, was that it was seen as work in social semiotics, not cultural studies and was not anthologised in any of the collections of Australian Cultural Studies that define the field although when I left Australia the then President of the National Cultural Studies Association, Kay Schaffer, delivered a keynote address in 1999 in which she farewelled both myself and John Frow as having been major contributors to the field (Schaffer 2000). The Lee/Poynton (2000) book was published in the Australian Cultural Studies Series edited by myself and John Tulloch because at the time we were as editors committed to introducing new methods to cultural studies (see also Tulloch/Lupton 1999 where quantitative work was the issue). It is also notable that in the Grossberg et. al. (1992) account of Cultural Studies, while there are chapters which lament the absence of sociology and history, and which argue that the language theory available in cultural studies is inadequate to what it has to do (Steedman 1992), no-one remarks on the absence of linguistics or critical discourse analysis. Steedman's (1992) remarks on the detrimental effects of interdisciplinarity and intertextuality as methods of teaching history, used and popularised she suggests because they required less training and are thus cheaper, could it seems to me be extended to cover the decline in disciplinary rigour and the failure to teach critical theory and linguistics/CDA itself in recent cultural studies contexts.

Norman Fairclough's work in Britain is among the first to actually use the label CDA (Fairclough 1995a). Fairclough used this Australian work (Kress/Threadgold 1988; Thibault 1991) when he produced a theory of discourse and social change, drawing on Foucault and a number of neo-Marxist and other social theorists, that did bring together a version of functional linguistics with sophisticated social and cultural theory. Pecheux (1982) whose approach to discourse was strongly framed by Althusser's Marxist theory of ideology (1977) is one of the discourse analysts he uses. So are Bakhtin and Kristeva (Fairclough 1995a: 189). Fairclough regards his approach as 'critical' because it combines a Marxist theory of discourse with linguistic methods of text analysis. In Fairclough's work however, text analysis itself in its current forms came in for critique and considerable rewriting. He was as critical of some forms of text linguistics (1992; 1995a) as I and others have been of the research vocabulary of cultural studies above. In particular, rather as I have talked above of the nominalisation of categories like discourse, intertextuality and so on, he argues for the need to unpack categories like 'scripts' and 'frames', which as he says involve very little linguistic analysis. In most cases, he suggests, textual analysis is 'neither systematic nor detailed' (1992: 196) and intertextual analysis is rare. There is an implication here that discourse analysis should involve the two things, analysis of texture and intertextuality, and that no discourse (language in use in social process, language as action) can be understood except in relation to the larger discursive formations - orders of discourse - of which it is a part. That approach is the result, in Fairclough's work (e.g., 1995a )of bringing linguistics and language theory together with Foucault's theory of discourse and it signals as it did elsewhere when these approaches came together (Threadgold 1997) a new interest in understanding not just the workings of individual texts, but the ways in which they are traversed by traces of, and enter into networks of, other texts and discourses to form part of the hegemonic discursive structures which form social realities, subjectivities and bodies. That kind of work remains relatively rare in CDA and in cultural studies and yet, arguably, if Foucault's work on discourse had been really understood, work of this kind should have followed.

Fairclough's (1995 b, 112 ff., 183 ff.) work on discourse and intertextuality was the first in CDA in the UK to actually attempt a linguistic description of the poststructuralist categories of intertextuality and discourse. He chose the functionalist linguistics of Halliday (1985) for the analysis of 'texture', the structure and organisation of texts, but saw intertextual analysis rather than Halliday's original theory of the social semiotic (1978) as a crucial way of linking texts and contexts. He includes in the analysis of texture structures beyond the sentence (Halliday and Hasan's cohesion) and conversational analysis (Fairclough1995a: 188). Importantly for some of the arguments I have made earlier in this paper, he is concerned to argue that intertextual analysis is one of the ways in which social scientists might be persuaded to see the relevance of CDA to the kinds of work they want to do on larger social structures.

However, he does offer a caveat here. In my work with Kress (1988) intertextuality was identified with narratives, genres and discourses. These were seen as the categories of which intertextuality is constituted, categories which are realised linguistically in texts, and which in turn constitute the larger orders of discourse on which texts draw just as much as they draw on the resources of the linguistic system (see also Threadgold 1997, ch. 6). However, as Fairclough and others have pointed out, the recognition of these traces of orders of discourse in text is very much dependent on the analyst's experience, knowledge of the context of the text, and interpretative biases (Fairclough 1995: 212). This then requires that one 'engage in social and ethnographic research over significant periods of time in particular institutional settings' (Fairclough 1995: 212) in order to ground, and test, claims about intertextuality in and against more recognisable forms of social science research. I would add to this a comment that takes me back to my questions above about why certain kinds of textual work are characterised as empirical and other kinds as textual. If one is working with literary or philosophical texts, as Derrida for example does, the research that is required for justifying claims of intertextuality is wide reading in a textual field, 'dead texts' perhaps, but texts which, if we believe Bakhtin and Derrida, go on dialoguing in and with the present, go on being signed for in new contexts, go on constructing and forming worlds and people. My point here is that sometimes, in some fields of research, the materiality of the significant contexts with which one has to deal actually is constituted of other discursive (read here 'textual') events: although such events are always imbricated in materiality, especially for Derrida but for all poststructuralist theorists and for most proponents of CDA. Poststructuralist feminist work has, for example, argued that texts are always traces of bodies at work, that texts and bodies are always folded into one another (Threadgold 1997).

Teun van Dijk is among the best known and most influential of European critical discourse analysts and has recently produced two interesting text-book accounts of CDA (2001a; 2001b) which also identify other major European researchers in this field. Wodak's (2001) introduction to CDA provides another account of the development of the field and includes work from Germany not referred to in these other accounts. What is interesting about both these accounts is their difference from my own, and the questions both raise about CDA as any kind of established or clear-cut method for analysing texts or engaging in the typical politics that go with the field. Like van Dijk, Wodak points to the 'heterogeneity of methodological and theoretical approaches in this field' (2001: 2) and argues that what is common to different traditions is at best 'a shared perspective on doing linguistic, semiotic or discourse analysis' (2001: 2). Both trace the intellectual formation of the field back to Marx, the Frankfurt School and Habermas, Gramsci, Althusser, Foucault and Pecheux in different national contexts and van Dijk includes Stuart Hall and the Birmingham School of Cultural Studies as part of the narrative of the development of critical perspectives in the study of language (van Dijk 2001a: 301). Van Dijk is insistent on the anti-methodology stance we have seen earlier to be consistent with cultural studies and critical theory approaches. He argues that 'ready-made methods' are 'incompatible with a critical attitude' (2001b: 95). Nor, he says, is CDA a 'theory that can simply be applied to social problems' (2001b: 96). It is a 'perspective' on doing research which 'focuses on social problems' and 'on the role of discourse in the production and reproduction of power abuse or domination' (2001b: 96). His emphasis is on the necessity 'for a broad, diverse, multidisciplinary and problem-oriented CDA' (2001b: 97) which will select its methods and areas of analysis on the basis of a theoretical analysis of social issues (2001b: 98). His own example of 'how to do' CDA, his tool kit if you like, is distinctly different from the functional linguistic toolkits we have noted above. There are recognisable parallels in some of the areas of text and language explored but there is no common language here with Barker/Galasinski (2001), Fairclough or e.g., Kress/Threadgold.

These differences highlight some of the issues singled out by Wodak (2001: 12) as needing further research in the field: the problem of theorising the way texts mediate the relations between the social and the linguistic; which linguistic theory to apply (often 'a whole mixed bag of linguistic indicators and variables are used' with 'no theory of grammar' to support that use); the problem of defining 'context'; the accusation that the politics of CDA biases analyses; that the field has not yet achieved a real inter- or trans-disciplinarity. These questions are asked again by Wetherell (2001). They do not go away. Wetherell/Taylor/Yates (2001) have produced an Open University reader which attempts to construct the field as a field with methods in all of its diversity. It includes the following traditions: conversation analysis; discursive psychology; Foucauldian research; CDA and critical linguistics; interactional sociolinguistics and the ethnography of speaking; and Bakhtinian research. Wetherell too is in favour of 'the creative mixing of research questions and styles' (2001: 383) but points to the habits and the epistemological and methodological debates between different 'disciplines, domains and traditions' which often inhibits such creativity.

Among such resistances is one in particular which over the years I have not commented upon: the systematic failure of the key male figures in CDA with some exceptions (Teun van Dijk is one) to cite or use feminist work in the field. Women and men do work together. It is feminist theory which is resisted I think. A long time ago Meaghan Morris pointed to this same feature as a characteristic of early work in postmodernism and produced a whole bibliography of work by women in the field to challenge the typically masculine representation of it (1988). This absence in CDA may well impact on its reception in cultural studies where feminist theory has been strong.

These I think are very real issues which raise again the question of what value CDA might be to cultural studies. In a sense before we can ask that question we need to know: whether they are not already part of the same enterprise; what exactly it is that we are offering as CDA; whether CDA might still have things to learn from cultural studies and other places; whether the only thing we can really offer, 'a shared perspective' on doing the political work of CDA is not already shared with, and even derived from, encounters with poststructuralist and critical theory and cultural studies itself; and finally, what disciplinary and interdisciplinary resistances there may still be to change and innovation in research styles and methods, not only in cultural studies but in CDA itself.

Date: 2016-01-14; view: 852