Definining Discourse Analysis
The prevalence of discourse analysis in research has been traced to the ‘linguistic turn’ of the 20th century (Locke, 2004) by which is meant the development in philosophy which recognised that language was not so much to be understood as reflecting reality but as creating reality, that reality did not pre-exist to be discovered by language but that language brings reality to life in the sense of being understood and interacted with by humans.
This focus on language need not involve a social constructivist view of reality, or a denial of reality beyond words or discourse as some suggest (Potter, 2004, p.202; Bryman, 2008, p.370): indeed some key figures within discourse analysis research adopt a realist or critical realist position (Fairclough, 2003; Rapley, 2008). As Pring (2000, p.116) summarises: ‘The acceptance of a reality independent of the researcher does not contradict the possibility of many interpretations of that reality’.
Such a position holds that there is external reality, a world and a social sphere beyond words, but that such reality is addressed in human terms through language, and so that interrogating language is a worthwhile and indeed essential activity. While it may be that external reality is fixed and timeless, the same is not true of human understanding, nor is it true of the language by which understanding is formed and constituted: knowledge thus, immanent in language, is fluid, contingent, provisional, and perspectival (Coyle, 1995, p.243; Wetherell et al., 2001, p11).
Discourse analysis has been described as ‘the close study of language in use’ (Taylor, 2001, p.5), its importance reflected in the epistemological position outlined earlier. However, this definition has been added to helpfully by Fairclough (1992, p28): ‘Discourse is more than just language use: it is language use seen as a type of social practice’ and thus he argues that discourse analysis is not solely bound to the text but must also involve
‘analysing the relationship between texts, processes, and social conditions, both the immediate conditions of the situational context and the more remote conditions of institutions and social structures’(1989, p.26).
As a result, Fairclough outlines three integrated levels of discourse, involving analysis of text, of discursive practices, and of social practices (1992, p.73).
- at textual level, this involves critical linguistics;
- at discourse level is analysis of text production, distribution and interpretation, especially in terms of the way in which the readership is guided to a ‘preferred’ reading;
- at the level of social practice, analysis explores the extent to which the text upholds, or reproduces, hegemonic discursive or social practices, how it stands in relation to certain prevalent conditions.
This seems to be a helpful approach to bridge the polarities of Foucauldian sociopolitical analysis and more linguistic-oriented studies (van Dijk, 2001, p.263). Similarly, it is argued, in relation to the discourse analysis of policy texts, that it is necessary to scrutinize ‘the material conditions within which such texts are produced and to examine critically the institutional practices which they are used to defend’ (Olssen, Codd, & O’Neill, 2004, p.72). Fairclough thus identifies
‘a need to develop approaches to text analysis through a transdisciplinary dialogue with perspectives on language and discourse within social theory and research in order to develop our capacity to analyse texts as elements in social processes’ (2003, p6).
‘multidimensional, involving for instance economic, political and cultural issues, and always involving the relationship between discourse and other elements of social life. So the approach needs to be interdisciplinary. . .where the objective is to develop one’s theory and methodology through dialogue with other disciplines’ (Fairclough, 2005a, p.1).
The close alignment between Fairclough’s and Foucault’s understandings of discourse means that a rapport can be established readily between Foucault’s critical purpose and Fairclough’s analytical methods: the domain of inquiry is the same. As has been noted, however, Fairclough’s three-dimensional conception of discourse develops Foucauldian methods fruitfully by having a specific role for close linguistic analysis and this can be seen as strengthening the empirical basis for the emerging critique. Questions over the empirical foundations for some of Foucault’s work – particularly his genealogical inquiries – have become more prevalent in recent times (Paglia, 1992, 1998; Scull, 2007) and it seems prudent to guard against this by adhering more rigorously to Fairclough’s model.
Some discourse analysts have attempted to categorise Fairclough’s three dimensions – textual, discursive, social – in a way that renders them reducible to exercises in description, interpretation, and explanation respectively (Titscher et al., p.153). While there is some value in this understanding, it seems to be too clinical in its divisions and it would be more judicious to view these three activities as permeating each of Fairclough’s dimensions, which are themselves inter-linked, in any case. For example, to classify textual analysis as solely an exercise in ‘description’ seems to reduce it to some form of mechanical linguistic labelling when it could involve quite complex arguments about the significance and effects of word-choice and sentence structure. Such a discussion would also inevitably make reference to wider issues relating to discursive practice and social practice. Furthermore, using the term ‘explanation’ as descriptive of the role of Fairclough’s analysis of social practice may arouse positivist connotations which would make its usage here problematic, as will be discussed later.
Purpose of Critical Discourse Analysis
Critical discourse analysis (CDA) is distinguished by its objective of uncovering issues relating to power and domination and, as with all critical theory, is concerned with emancipation in the sense of opening up spaces in which people can identify dominant ideology and escape from such oppressive discourse. This approach is termed ‘dissident research’ in which the aim is to ‘understand, expose, and resist social inequality’ (van Dijk, 2001, p.352) . In Foucauldian terms, the role of critique is less pointed: the purpose is strictly revelatory, rather than emancipatory (although, as has been argued, these can be interpreted as being intertwined, albeit implicitly).
With specific reference to critical policy analysis, what is probed is how power is used to define the parameters of particular questions, set rules for particular practices, and shape agendas; it seeks to complicate policy, to include inquiry into underlying issues of power and ideology which are embedded in the very framing of policy problems and solutions (Woodhouse-Jiron, 2004, p.175). Ozga (2000, p. 95) summarises the role of analysis of policy texts as exploring the source, scope, and pattern of policy, this last being the way the nature of relationships and organizations are (re)framed in the text.