Ideology has a double face.
Ideology is inscribed in social practice.
‘Context’ is structured like a text.
Interpretation is a struggle.
Truth is always at risk.
Syntax is meaning.
2.8 Norman Fairclough on CDA
Norman Fairclough’s (1995) work in Britain is among the first to use the label CDA. Discourse is a category that is used by both social and linguistic theorists. Fairclough believes discourse refers to spoken or written language use and defines it as: ”the use of language seen as a form of social practice” (Fairclough, 1995a: 7). He mentions that discourse analysis is “the analysis of how texts work within sociocultural practice”. He defines CAD as the “analysis of the dialectical relationships between semiosis (including language) and other elements of social practices” (Wodak & Meyer, 2001: 123). In his later works and researches, he extends the definition of discourse to include other semiotic practices in other semiotic modalities such as photography and non-verbal language (see Lassen et al., 2006). Fairclough (1995: 131) believes that language use as social practice is a mode of action and it is a socially and historically situated mode of action, in a dialectic relationship with other facets of the society. Fairclough produced a theory of discourse and social change that tries to combine a theory of power based on Gramsci’s concept of intertexuality discussed by Bakhtin. He describes intertextuality as: ‘texts are constituted from other already produced texts and form potentially diverse text types’. Fairclough talks about Bakhtin’s theory of genre’ and states: ‘the productivity and creativity of discourse practice and its realization in texts which are heterogeneous in their forms and meanings, the heterogeneity emanating from their intertexuality’. He uses Gramscian ”theory of hegemony” in analysis of sociocultural practice when he says ”Gramsci highlights both how power relations constrain and control productivity in discourse and practice and how a particular relatively stabilized configuration of discourse practices constitutes one domain of hegemony”. Gramscian concept of hegemony suggests a focus in studies of language/ideology upon change in discoursal practice and structures regarding one aspect of change in the balance of social forces. Discourse conventions may represent nationalized ideologies which make them a most efficient mechanism for sustaining hegemonies. Moreover, control over the discursive practices of institutions is one aspect of cultural hegemony. Fairclough (1995) uses Halliday’s theoretical framework like many other CDA scholars to enact ”ideational”, ”interpersonal” and ”textual” function of language. Fairclough (1995: 132) regards his approach as ”critical” because it combines a Marxist theory of discourse with linguistic methods of text analysis and by this aims to methodically discover often obscure “relationships of causality and determination between: (a) discursive practices, events and texts and (b) wider social and cultural structures, relations and practices”. He states that there are seven ideological properties in a text: vocabulary and metaphors, grammar, presuppositions and implications, politeness conventions, turn-taking systems, generic structure and style. Fairclough’s (1995) analytical framework which explores the linkages between discourse, ideology and power is a three dimensional framework. He indicates that each discursive event has three dimensions. In this regard his framework of analysis is constituted of: analysis of language texts, analysis of discourse practice (process of text production, distribution, and consumption), and analysis of discursive events as instances of sociocultural practice. There is an implication here that discourse analysis should involve the two things, analysis of texture and intertexuality, and that no discourse can be understood except in relation to the larger discursive formations – orders of discourse- of which it is a part.
In particular, ”Language and Power” (Fairclough, 1989) explores the interwoven nature of language and social institutional practices and of wider political and social structures. In his book, Fairclough develops the concept of synthetic personalization to account for the linguistic effects providing an appearance of direct concern and contact with the individual listener in mass-crafted discourse phenomena, such as advertising, marketing, and political or media discourse (see Fairclough, 1989).
According to Van Dijk (1998 cited in Schiffrin & Hamilton, 2001: 271-280), Fairclough and Wodak summarize the main tenets of CDA as follows:

CDA addresses social problems.
Power relations are discursive.
Discourse Constitutes Society and Culture.
Discourse does ideological work.
Discourse is historical.
The link between text and society is mediated.
Discourse analysis is interpretative and explanatory.
Discourse is a form of social action.
Fairclough 1999 (Wodak & Meyer, 2001: 123) states that his particular concern is with ”the radical changes that are taking place in contemporary social life, with how semiosis figures within processes of change, and with shifts that take place in the relationship between semiosis and other social elements within networks of practices”.
An analytical framework for CDA is represented below by Fairclough. (Wodak & Meyer, 2001: 125):

Focus upon a social problem, which has a semiosis aspect.
Identity obstacles to it being tackled, through analysis of:
A. The network of practices it is located within.
B. The relationship of semiosis to other elements within the particular practice(s) concerned.
C. The discourse (the Semiosis itself):

Structural analysis: the order of discourse
Interactional analysis
Interdiscursive analysis
Linguistic and semiosis analysis
Consider whether the social order (network of practices) in a sense ‘needs’ the problem.
Identify possible ways past the obstacles.
Reflect critically on the analysis (1-4).
Fairclough’s approach to CDA is problem-based and for him CDA is a form of critical social science, which is visualized as social science which is general to revealing the problems that people are confronted with by particular forms of social life, and to contributing resources which people may be able to draw upon in tackling and overcoming these problems (Wodak & Meyer, 2001: 125).
Fairclough goes on to consider the social effects of CDA and mentions that like critical social science generally, CDA has emancipatory objectives, and is focused upon the problems confronting what we can loosely refer to as the “losers” within particular forms of social life, i.e. the poor, the socially excluded, those subject to oppressive gender or race relations, and so forth. (Wodak & Meyer, 2001: 125).

Tenn Adrianus Van Dijk on CDA
Van Dijk (1998, cited in Schiffrin & Hamilton, 2001: 271) defines Critical Discourse Analysis as ‘a type of discourse analytical research that primarily studies the way social power abuse, dominance and equality are enacted, reproduced and resisted by text and talk in the social and political context’. Van Dijk (1998, cited in Schiffrin & Hamilton, Handbook of Discourse Analysis, 2001: 282) mentions that ”CDA concentrates on social problems, and especially on the role of discourse in the production and reproduction of power abuse or domination”. According to him CDA can be conducted in, and combined with any approach and subdiscipline in the humanities and the social sciences. He declares CDA must account for the pa
rticulars of the relationships between discourse structures and social structures, according to its multidisciplinary theories. Van Dijk adds that CDA is a ‘dissident research’ and ‘critical discourse analysts take explicit position to understand, expose, and finally resist social inequality’. It does so from a perspective that is consistent with the interests of dominated groups. It takes the experiences and opinions of members of such groups seriously, and supports their struggle against inequality. that is, CDA research combines what perhaps somewhat pompously used to be called ‘solidarity with the oppressed’ with an attitude of opposition and dissent against those who abuse text and talk in order to establish, confirm or legitimate their abuse of power (Wodak& Meyer, 2001: 96). Unlike many other studies, CDA does not deny but explicitly defines and defends its own sociopolitical position. Specialized also in the critical analysis of scholarly discourse, CDA of course recognizes the tactical nature of such accusations as part of the complex mechanisms of domination, namely as an attempt to marginalize and problematize dissent and precisely because of its combined scholarly and social responsibilities.
According to multidisciplinary orientation chosen by Van Dijk, the overall label used for his way of doing CDA is that of ‘socio-cognitive’ discourse analysis (Wodak & Meyer, 2001: 20). However he believes that this label does not mean that he thinks that CDA should be limited to social and cognitive analysis of discourse, or to some combinations of these dimensions and insists that for him this only means that he is ”personally most interested in the fascinating socio-cognitive interface of discourse analysis”. According to Van Dijk 1997(cited in Wodak & Meyer, 2001: 97) “because of the fundamentally verbal nature of discourse, CDA also needs a solid ‘linguistic’ basis, where ‘linguistic’ is understood in a broad structural– functional sense”. In other words, CDA as a specific form and practice of discourse analysis obviously always needs to account for at least some of the detailed structures, strategies and functions of text and talk, including grammatical, pragmatic, interactional, stylistic, rhetorical, semiotic, narrative or similar forms of verbal and paraverbal organization of communicative events.
Van Dijk tries to give some guidelines and formulate a method in carrying CDA researches and put his effort to limit his tries to the domain defined by ”the theoretical discourse–cognitive–society triangle”. (Wodak & Meyer, 2001: 98) Knowing the interpretations which may be caused because of labels of this sort and hence responsible for reductionist misconception’ is here meant in the broad sense of a ‘communicative event’, including conversational interaction, written text, as well as associated gestures, typographical layout, images and any other ‘semiotic’ or multimedia dimension of signification. Similarly, ‘cognition’ here involves both personal and social cognition, beliefs and goals as well as evaluations and emotions, and any other ‘mental’ or ‘memory’ structures, representations or processes involved in discourse and interaction. And finally, ‘society’ is meant to include both the local, microstructures of situated face-to-face interactions, as well as the more global, societal and political structures variously defined in terms of groups, group-relations (such as dominance and inequality), movements, institutions, organizations, social processes, political systems and more abstract properties of societies and cultures. he goes on to state that adequate discourse analysis at the same time needs detailed cognitive and social analysis, and vice versa, and that it is only the combination of these accounts that may reach descriptive and especially critical sufficiency in the study of social problems( Wodak & Meyer, 2001: 100). Van Dijk explains that CDA does not provide a clichéd approach to social analysis, but emphasizes that for each study a thorough theoretical analysis of a social issue must be made, so as to be able to select which discourse and social structures to analyze and to relate (Wodak & Meyer, 2001: 100). He believes that in any practical sense there is no such thing as a complete discourse analysis: a full analysis of a short passage might take months and fill hundreds of pages. He mentions the reason for that as decades of specializations in the field have ‘discovered’ many hundreds of relevant units, levels, dimensions, moves, strategies, types of acts, devices and other structures of discourse. There may be visual, phonological, syntactic, semantic, stylistic, rhetorical, pragmatic, and interactional levels and structures. Hence in CDA a researcher must make choices, and select those structures for better analysis that are relevant for the study of a social issue (Wodak & Meyer, 2001: 98). Thus, if we want to study the ways some speakers or writers exercise power in or by their discourse, it only makes sense to study those properties that can vary as a function of social power. Thus the features which are under the control of speaker are topics for CDA and not those obligatory forms such as the form of words and many structures of sentences are grammatically obligatory and contextually invariant and hence usually not subject to speaker control, and hence irrelevant for a study of social power.
Van Dijk (1998, cited in Schiffrin & Hamilton, 2001: 275) suggests critical research on discourse demands to satisfy a number of prerequisites in order for its aim to be realized effectively:

“As is often the case for more marginal research traditions, CDA research has to be ‘better’ than other research in order to be accepted.
It focuses primarily on social problems and political issues, rather than on current paradigms and fashions.