An Overview of Focal Approaches of Critical Discourse Analysis

This article aims to present detailed accounts of central approaches to Critical Discourse Analysis. It focuses on the work of three prominent scholars such as Fairclough’s critical approach, Wodak’s discourse-historical approach and Van Dijk’s socio-cognitive approach. This study concludes that a combination of these three approaches can be useful to critical analysis of texts.


Introduction
This paper attempts to review the central theoretical and methodological approaches to critical discourse analysis (CDA). First, a background on CDA is presented and then three main approaches within this field are discussed. Critical discourse analysis is a method that analyzes language as discourse, which means that "language is conceived as one element of the social process dialectically interconnected with others" (Fairclough & Graham, 2002, p. 188). CDA analyzes real instances of social interaction which take a complete or partial linguistic form as it aims to make visible "the ideological loading of particular ways of using language and the relations of power" which underlie them (Fairclough & Wodak, 1997, p. 258). Critical discourse analysis examines social practices based on their discourse moments (Chouliaraki & Fairclough, 1999). It emphasizes "the substantively linguistic and discursive nature of social relations of power" and the way they are used and discussed in discourse (Fairclough & Wodak, 1997, p. 272). CDA is used to analyze texts in order to discover what "structures, strategies or other properties of text, talk, verbal interaction or communicative events play a role" in production or reproduction of unequal power relations (Van Dijk, 1993a, p. 250). Wetherell, Taylor and Yates (2001) also describe CDA as: the study of talk and texts. It is a set of methods and theories for investigating language in use and language in social contexts. Discourse research offers routs into the study of meanings, a way of investigating the backand-forth dialogues which constitute social action, along with the patterns of signification and representation which constitute culture (Wetherell, Taylor & Yates, 2001, p. i).
discourses are knowledge systems of the human sciences (medicine, economics, linguistics, etc.) that inform the social and governmental 'technologies' which constitute power in modern society. They are partly realized in ways of using language, but partly in other ways (for example, ways of designing prisons or schools) (Fairclough & Wodak, 1997, p. 261).
CDA approach is critical in the way that it involves "having distance from the data, embedding the data in the social, making a political stance explicit, and having a focus on self-reflection as scholars doing research" (Martin & Wodak, 2003, p. 6). In the same way, Fairclough (1995b) points out that his approach to discourse analysis is critical which intends: to make visible through analysis, and to criticize, connections between properties of texts and social processes and relations (ideologies, power relations) which are generally not obvious to people who produce and interpret those texts, and whose effectiveness depends upon this opacity (Fairclough, 1995b, p. 97).
In sum, according to Fairclough and Wodak (1997), the principles of CDA are outlined as follows: The first principle highlights that CDA addresses social problems. CDA not only concentrates on language and the use of language, but also focuses on the "linguistic character of social and cultural processes and structures" (Fairclough & Wodak, 1997, P. 271).
The second principle is that power relations are discursive. CDA draws attention to "the substantively linguistic and discursive nature of social relations of power in contemporary societies" (ibid., p. 272). This shows that power relations are performed and constructed in and through discourse.
The third principle is that discourse constitutes society and culture. It emphasizes a dialectical relationship between them i.e., discourse shapes society and culture, as well as being shaped by them (ibid.).
The next principle highlights that discourse does ideological work. It shows that discourse is not neutral; there are some ideologies behind it. Ideologies refer to as "particular ways of representing and constructing society which reproduce unequal relations of power, relations of domination and exploitation" (Fairclough & Wodak, 1997, p. 275).
Another principle underlines that discourse is historical. It means that "discourse is not produced without context and cannot be understood without taking the context into consideration" (ibid., p. 276).
The next principle emphasizes that the link between text and society is mediated. CDA is "about making connections between social and cultural structures and processes on the one hand, and properties of text on the other" (Fairclough & Wodak, 1997, p. 277). Fairclough (1993) highlighted that the link between text and society is mediated through Foucault's notion of "orders of discourse". For example, "the order of discourse that organizes, say a university will be characterized by a host of interrelated textual practices such as the discourses of essays, meetings, lectures, seminars, administrative texts and so on" (Simpson & Mayr, 2010, p. 53).
Another principle points up that discourse analysis is interpretative and explanatory. CDA aims to move beyond textual analysis to the interpretation and explanation stages of analysis. CDA shows that "discourse can be interpreted in very different ways, due to the audience and the amount of context information which is included" (Fairclough & Wodak, 1997, p. 278). CDA may deconstruct a contradiction within a text to demonstrate "the different implications of different readings for social action" (ibid., p. 279). Fairclough and Wodak (1997) argued that critical analysis of a text requires "systematic approach to inherent meanings", "scientific procedures", and "self-reflection of the researchers", which make the critical readings of a text different from uncritical reading by an uncritical audience. This interpretative and explanatory nature of discourse analysis is "dynamic and open, open to new contexts and new information" (ibid., p. 279).
The last principle refers to discourse as a form of social action. CDA intends to find out opaqueness and relations of power. It refers to "a socially committed scientific paradigm, and some scholars are also active in various political groups" (ibid., pp. 279-280).
The focus of CDA is in the relationships between discourse and social power, and its main goal is to "describe and explain how power abuse is enacted, reproduced or legitimised by the text and talk of dominant groups or institutions" (Van Dijk, 1996, p. 84). Similarly, CDA aims to find out "unequal relations of power" and "to reveal the role of discourse in reproducing or challenging socio-political dominance" (Garret & Bell, 1998, p. 6).

Fairclough's Critical Approach
Fairclough's (2001a) theoretical objectives have stemmed from linguistics and those studies in sociolinguistics, which focus on "language in its social context" and on relationship between language and power (p. 1). However, in spite of their contribution to Fairclough's critical approach, these approaches have major weaknesses from a critical point of view: Linguistics describes "language as a potential, a system, an abstract competence, rather than attempting to describe actual language practice" (Fairclough, 2001a, p. 5). Following the notion of langue (language) and parole (speaking) by Ferdinand de Saussure, the main concern of linguistics is on langue rather than parole. Saussure assumed that language of a community for all practical purposes is invariant for that particular community. In consequence, he stated that the study of langue should be synchronic, as a static system at a given point in time, rather than historical, which regards the study of langue dynamically as it changes through time (Fairclough, 2001a). Therefore, Fairclough (ibid.) criticized Saussure's viewpoint as it fails to consider that language is shaped socially.
Sociolinguistics, in contrast emphasizes, "Language use is shaped socially and not individually" (Fairclough, 1993, p. 63). It concentrates on socially variable language use and shows that there are systematic relationships between variations in linguistic form (phonological, morphological, syntactic) and social variable (social relationship between participants, differences in social setting, differences of topic, etc.) (Fairclough, 2001a). According to Fairclough (ibid.), while sociolinguistics is strong to describe what the facts of variations are, it fails to explain that these facts are the production of power relations and struggles.
In Fairclough's (2001a) approach discourse is viewed as "a form of social practice" which shows that it is a mode of action (recognized by Austin, 1962;Levinson, 1983). This means that spoken or written utterances constitute the performance of speech acts such as promising, asking, asserting, warning, and so forth. Moreover, Fairclough's (2001a) view on discourse considers language as a part of society which highlights that there is a dialectical relationship between language and society whereby language is a part of society. That is to say, on the one hand "linguistic phenomena are social" in the sense that interactions are both determined socially and have social effects, and on the other hand "social phenomena are linguistic" in the sense that language activity which occurs in social contexts is part of social processes and practices rather than merely an expression and reflection of them (ibid.).
Subsequently, Fairclough (2001a) regards language as a social process which means that discourse is the process of text production and interpretation. Finally, he considers language as a socially conditioned process. Namely, discourse involves social conditions of production and interpretation, which relate to different levels of social organizations: "the level of the social situation, or the immediate social environment in which the discourse occurs; the level of the social institution which constitutes a wider matrix for the discourse; and the level of the society as a whole" (Fairclough, 2001a, pp. 20-21). Fairclough (2001b) adds that intertextual analysis can be considered as a complementary part to linguistic analysis. The intertextual analysis "focuses on the borderline between text and discourse practice" (Fairclough, 1995a, p. 61), and acts as a bridge between language and social contexts or between texts and discourse contexts in his three-dimensional analytical framework (Fairclough, 2003). The term "intertextuality" was coined by Kristeva in the late 1960s in her accounts on Bakhtin's work. Kristeva's (1986) concept of intertextuality means "the insertion of history (society) into a text and of this text into history" (p. 39). Influenced by Bakhtin's (1986) theory, Fairclough (1993) states that texts are inherently intertextual, which means that they are constituted by elements of other texts. For Fairclough (1993) discourse representation which is a major part of the news is a form of intertextuality in which parts of other texts are integrated into a text through devices such as quotation marks and reporting clauses.
Besides, Fairclough concentrates on the correlation between discourse, power and ideology, which he combines the view of discursive practice inspired by Bakhtin (1986) via Kristeva's (1986) concept of intertextuality with Gramsci's (1971) theory of hegemony (Fairclough, 1993). According to Fairclough (1993, p. 92) the concept of hegemony refers to "a way of theorizing change in relation to the evolution of power relations which allows a particular focus upon discursive change, but at the same time a way of seeing it as contributing to and being shaped by wider processes of change." Fairclough (ibid.) also views hegemony as domination across various domains of a society, including economic, political, ideological, etc., and as "the power over a society as a whole of one of the fundamental economically-defined classes in alliance with other social forces." According to Chouliaraki and Fairclough (1999) hegemony is defined as: relations of domination based upon consent rather than coercion, involving the naturalisation of practices and their social relations as well as relations between practices, as matters of common sense -hence the concept of hegemony emphasises the importance of ideology in achieving and maintaining relations of domination (Chouliaraki & Fairclough, 1999, p. 24).
The concept of dominance and hegemony is employed in analyzing orders of discourse (Fairclough, 2001b). He states that "social practices networked in a particular way constitute a social order" (Fairclough, 2003, p. 206). Drawing on Foucault notion of orders of discourse Fairclough states: An order of discourse is a network of social practices in its language aspect. The elements of orders of discourse are not things like nouns and sentences (elements of linguistic structures), but discourses, genres and styles (…). These elements select certain possibilities defined by languages and exclude others -they control linguistic variability for particular areas of social life (Fairclough, 2003, p. 24).
Orders of discourse may change overtime and these changes are determined by changing the power relations at social interaction. As Fairclough (2001a) notes: How discourses are structured in a given order of discourse, and how structurings change over time, are determined by changing relationships of power at the level of the social institution or of the society. Power at these levels includes the capacity to control orders of discourse; one aspect of such control is ideologicalensuring that orders of discourse are ideologically harmonized internally or (at the societal level) with each other (Fairclough, 2001a, p. 25). Fairclough (2001b) looks at the relationships between the orders of discourse which he calls 'interdiscursivity', and adds that, "interdiscursivity of a text is a part of its intertextuality, a question of which genres, discourses and styles it draws upon, and how it works them into particular articulations" (p. 124).

Wodak's Discourse-historical Approach
In accordance to Fairclough's critical approach, the discourse-historical approach perceives discourse as a form of social practice (Fairclough & Wodak, 1997;Wodak, 2001). Wodak (2001) infers a dialectical relationship between discursive practices and fields of action (situations, institutional and social structures), in which they are situated. It means that: on the one hand, the situational, institutional and social settings shape and affect discourses, and on the other, discourses influence discursive as well as non-discursive social and political processes and actions. In other words, discourses as linguistic social practices can be seen as constituting non-discursive and discursive social practices and, at the same time, as being constituted by them (Wodak, 2001, p. 66). Wodak (2001) defines discourse "as a complex bundle of simultaneous and sequential interrelated linguistic acts, which manifest themselves within and across the social fields of action as thematically interrelated semiotic, oral or written tokens, very often as 'texts'" (p. 66). Reisigl and Wodak (2009) consider discourse to be: · a cluster of context-dependent semiotic practices that are situated within specific fields of social action; · socially constituted and socially constitutive; · related to a macro-topic; · linked to the argumentation about validity claims such as truth and normative validity involving several social actors who have different points of view (Reisigl & Wodak, 2009, p. 89).
Wodak (2001) like Fairclough views texts as the products of discourse and defines texts "as materially durable products of linguistic actions" (p. 66). Discourse historical approach considers intertextual and interdiscursive relationships between texts, genres and discourses, as well as sociological variables, and situational frames. In this approach, intertextuality means that texts are connected to other texts, while interdiscursivity means that discourses are connected to each other. The approach focuses on all these relationships to explore how discourses, genres, and texts change in relationship to socio-political change (Wodak, 2001).
For the analysis of the interrelationship between discursive and other social practices and structures, Wodak (2001) employs a principle of triangulation which combines different interdisciplinary approaches. For example, for investigating discursive construction of collective groups like races, nations and ethnicities, the interdisciplinary approach has combined historical, socio-political, as well as linguistic perspectives. Moreover, the principle of triangulation implies different methods of collecting data, and analysis of different corpora and genres. Wodak's (2001, p. 67) triangulatory approach is based on context which takes into account four dimensions: (1) "the immediate language or text internal co-text"; (2) "the intertextual and interdiscursive relationship between utterances, texts, genres and discourses"; (3) the social/sociological variables and institutional frames of a particular context of situation; and, (4) the broader socio-political and historical context which the discursive practices are embedded within and related to.
According to Reisigl and Wodak (2001), in analyzing texts related to races, ethnicities, nations, or national identities, the presentation of discursive strategies orients towards five questions: · How are persons named and referred to linguistically? (referential strategies). · What traits, characteristics, qualities and features are attributed to them? (predicational strategies). · By means of what arguments and argumentation schemes do specific persons or social groups try to justify and legitimate the exclusion, discrimination, suppression, and exploitation of others? (argumentation strategies, including fallacies). · From what perspective or points of view are these namings, attributions and arguments expressed?
(perspectivation, and framing strategies). · Are the respective discriminating utterances articulated overtly, are they even intensified or are they mitigated?
(mitigation and intensification strategies) (Reisigl & Wodak, 2001, p. xiii). Wodak et al. (2009) further illustrate the discourse-historical framework via their study which examines the discursive construction of national identity in Austria. Wodak et al. (ibid.) analyze interviews, focus-group discussions, and media products, including newspaper, posters, politicians' speeches, and so on. They distinguish between three levels of discourse and textual analysis, including contents, strategies, as well as means and forms of realization. In the first level of discourse analysis, Wodak et al. (2009)  The second level of Wodak et al.'s (2009) analytical framework is concerned with strategies which can be adopted to achieve a certain goal such as political and psychological objectives. Drawing on Bordieu's definition of strategy, Wodak et al. (2009) state that "strategic action is oriented towards a goal but not necessarily planned to the last detail or strictly instrumentalist; strategies can also be applied automatically" (p. 32). Wodak et al. (2009, p. 33) identify four kinds of macro-strategies: First, constructive strategies which aim at constructing and establishing "a certain national identity by promoting unification, identification and solidarity, as well as differentiation." Second, perpetuating strategies which "attempt to maintain and to reproduce a threatened national identity i.e., to preserve, support and protect it." This entails strategies of justification, which are "employed primarily in relation to problematical actions or events in the past which are important in the narrative creation of national history." Next, transformational strategies which intend "to transform a relatively well-established national identity and its components into another identity the contours of which the speaker has already conceptualised." Finally, destructive strategies which "aim at dismantling or disparaging parts of an existing national identity construct." Subsequently, at the third level of discourse analysis, Wodak et al. (2009) focus on linguistic means involved in the discursive construction of national identity, especially on lexical items and syntactic devices which "serve to construct unification, unity, sameness, difference, uniqueness, origin, continuity, gradual or abrupt change, autonomy, heteronomy and so on" (p. 35) which the most important of them are as follows: 1. Personal reference (anthroponymic generic terms, personal pronouns, quantifiers); 2. Spatial reference (toponyms/ geonyms, adverbs of place, spatial reference through person, by means of prepositional phrases such as 'with us', 'with them'); 3. Temporal reference (temporal prepositions, adverbs of time, temporal conjunctions, temporal references by means of nouns, semi-prefixes with temporal meaning) (Wodak et al., 2009, p. 35). Wodak et al. (2009) highlight that deixis (personal, spatial and temporal references) is the most important linguistic device in national identity discourses. Deixis, according to Fowler (1991), "consists of the devices which link a text with the time and place of communication and with the participants; which 'orient' speaker and addressee in relation to the content of the discourse" (p. 63). Furthermore, Wodak et al. (2009, p. 35) examine linguistic and rhetorical devices, including euphemisms, allusions, rhetorical questions, passive versus active voice, agency personification, and so on. Wodak et al.'s (2009) study assumes that national identities are produced and reproduced through discourse. It shows that construction of national identities is not entirely discursive; indeed, institutional and material social structures determine national identities. In addition, the study assumes that "there is no such thing as one national identity in an essentialist sense, but rather that different identities are discursively constructed according to context" (ibid., p. 186). Namely, they are constructed according to the degree of public exposure of a discourse, the situation and mode of the exposure, the topic, and the addresses, among other things. Moreover, their study highlights the importance of certain intertextual connections, including: Literal repetitions of passages from commemorative addresses and text extracts from books and articles by historians, political scientists and essayists, and hackneyed formulations transferred from the areas of politics and the media into semi-public and quasi-private areas (recontextualisation) (Wodak et al., 2009, p. 189).
Further, Wodak et al.'s (2009) study demonstrates that discourses about nations and national identities are perceived in combination with social macro-strategies. Finally, Wodak et al. (2009) compare the discourse of politicians and media elites with the semi-public and semi-private discourse and conclude that: The discursive national identification 'products' offered by these political and media élites to their targeted audiences was influenced partly by the demand of these target groups for images to reinforce their national confidence. At the same time these élites endeavoured to satisfy such demands for national identity, at times by creating, emphasising, or-as illustrated by the myth of permanent neutrality-by playing down particular features of this identity (Wodak et al., 2009, p. 202).

Van Dijk's Socio-cognitive Approach
Van Dijk's approach like Fairclough's approach attempts to connect the micro-structure of language to the macrostructure of society (see Kintsch &Van Dijk, 1978). However, instead of discursive practice, Van Dijk (1993a) focuses on social cognition as the mediating part between text and society. Van Dijk (ibid.) defines the social cognitions as "socially shared representations of societal arrangements, groups and relations, as well as mental operations such as interpretation, thinking and arguing, inferencing and learning" (p. 257). Kintsch and Van Dijk (1978) have distinguished between text's micro-structure and macro-structure. Macro level refers to power, dominance and inequality between social groups, whereas micro level refers to language use, discourse, verbal interaction and communication (Van Dijk, 2001b). According to Van Dijk (ibid.) societal structures is related to discourse structures through actors and their minds.
Van Dijk (1991) has applied his approach of discourse analysis to media texts. He believes one of the areas in which discourse plays an important role in the (re)production of inequality is that of race and ethnic relations (see Wodak & Reisigl 2001;Richardson, 2004). Van Dijk's (1984, 1993b, 2000b, 2000c) studies on discourse and racism have developed a general theory of the relations between discourse and racism. The major point of his work is that "racism is a complex system of social and political inequality that is also reproduced by discourse" (Van Dijk 2001b, p. 362; see Wodak & Reisigl, 2001).
Van Dijk's (1991) study, Racism and the Press, focuses on news discourse and ethnic minorities to show the media's role in reproducing racism and unequal power relations in society. Van Dijk (1991, p. 20) argues that ethnic minorities are marginalized in news report, and the press is "a representative of the white power structure" (see also Campbell, 1995;Allan, 1999). Van Dijk (1991) points out that minorities and immigrants are presented as a problem and threat, and are portrayed "in association with crime, violence, conflict, unacceptable cultural differences, or other forms of deviance" (p. 21) (see Campbell, 1995;Reisigl & Wodak, 2001). Van Dijk (1991) observes that the minority groups are quoted less often compared to the majority groups, even if the topics directly concern the minorities (see also Teo, 2000).
Van Dijk (2000a) has approached critical discourse analysis on the basis of understanding ideological structures and social relations of power involved in discourse. There is an argument that news texts are controlled by dominant power (Herman & Chomsky, 2002;Van Dijk, 1991). According to Van Dijk (2000a), ideologies may determine all structures of text or talk and they may be expressed explicitly or implicitly in the structure of discourse. Van Dijk (2001b) defines social power as control and asserts that groups have power if they can control the acts and minds of other groups. Van Dijk (ibid., p. 355) distinguishes two main types of power: (1) the "coercive power", which is based on force i.e., power of the military, power of violent men, etc.; and, (2) the "persuasive power", which is "based on knowledge, information, or authority" such as "the power of parents, professors, or journalists" (see Gramsci, 1971).
Van Dijk (2000a) introduces four principles for the analysis of ideology, which he calls "ideological square": "1-Emphasize positive things about Us; 2-Emphasize negative things about Them; 3-De-emphasize negative things about Us; 4-De-emphasize positive things about Them" (P. 44). Van Dijk's (ibid.) ideological square contributes to polarization of in-groups versus out-groups to represent Us favorably but Them unfavorably. Van Dijk (2000a, 2004 has proposed categories of ideological analysis conducive to the representation of Us versus Them which some of them are as follows: · Actor description: the ways we describe actors are based on our ideologies, for example, the description of ingroup as positive and out-group as negative. · Authority: mentioning authorities to support one's argument. · Categorization: classifying people to different groups and attributing them positive or negative characteristics. · Lexicalization: the expression of lexical items creating an overall ideological strategy for negative otherpresentation. · Polarization: categorizing people in in-group and out-group and assigning good attributes to Us and bad attributes to Them. · Vagueness: using vague expressions which do not have definite referents. · Victimization: emphasizing the "bad" nature of out-group by telling horrible stories about them. Finally, Van Dijk's (2001a) socio-cognitive approach focuses on the tripartite discourse-cognition-society model of ideology. According to Van Dijk (ibid.) here "discourse" means "communicative event", including conversation, written text, and any "semiotic" or multimedia dimension of signification. Personal or social "cognition" involves "mental" or "memory" structures, representations and processes in discourse and interaction such as beliefs, evaluations, and emotions. "Society" includes both microstructures of interactions, as well as societal and political structures such as group relations, institutions, and political system. Van Dijk (2001a) views "the combined cognitive and social dimensions of the triangle as defining the relevant (local and global) context of discourse" (p. 98).
As it is obvious from most of Van Dijk's studies, his critical analysis of texts tends to make explicit the ideological dimension of Us versus Them and to demonstrate the discursive structures and strategies used in exercising the dominant power.

Conclusion
According to the literature, CDA is not a single theory but rather a "research program with many facets and numerous different theoretical and methodological approaches" (Wodak, 1999, p. 186) and most prominent scholars who made a great contribution to CDA are Fairclough, Wodak, and Van Dijk (Chilton, 2005;Simpson & Mayr, 2010). Fairclough's socio-critical approach has been influenced by Foucault's (1972) theory of ideology and has drawn on Foucault notion of orders of discourse "covering a range of institutional discourse practices" (Simpson & Mayr, 2010, p. 53). Wodak's discourse-historical approach has been influenced by socio-linguistics and ethnography (Chilton, 2005). Van Dijk's socio-cognitive approach has drawn on various methods such as "textual, pragmatic, ethnomethodological, and cognitive approaches" to analysis of different discourses (e.g., political, racist) (Chilton & Schäffner, 1997, p. 211).
The study showed that Fairclough's socio-critical approach views discourse (language) as a form of social practice and Wodak's discourse-historical approach is fundamentally compatible with Fairclough's approach in a way that both consider discourse as a form of social practice. Besides, Wodak's approach is related to the socio-cognitive theory of Van Dijk which views discourse as a form of knowledge and memory. All these approaches highlight correlation between language, ideology and power relations. Hence, for critical analysis of texts, researchers can draw on a combination of theoretical and methodological approaches advocated by these prominent scholars.