Ideology, Identity, Interaction: Contradictions and Challenges for Critical Discourse Analysis Copyright © 2012 Critical Approaches to Discourse Analysis across Disciplines http://cadaad.net/journal Vol. 5 (2): 1 – 18 ISSN: 1752-3079 BOB HODGE Centre for Cultural Research University of Western Sydney, Australia [email protected] Abstract This article reflects on the condition of CDA, by analyzing key terms in the 2010 CADAAD conference: ideology, identity, interaction. It uses ideological-complex theory to emphasize contradiction as key to ideological effects in a highly complex world, source of both dynamism and vulnerability in theory, analysis and action. It argues for a single diverse and inclusive analytic project, including social, cognitive and linguistic lines, studying all media, including verbal, operating across all scales of space and time. Only an inclusive, contradictory CDA can have the impact it deserves. Keywords: CDA, social semiotics, ideology, identity, complexity, contradiction 1. Introduction This article takes the opportunity offered by the CADAAD conference, in the University of Lodz, Poland, September, 2010, to reflect on Critical Discourse Analysis (CDA). Where did it come from? What challenges does it face today? What resources could this tradition draw on to manage its own self- transformation? I put this review in a particular framework. Urry (2003) proclaimed a ‘complexity turn’ across the social sciences, seeing the major problems of today as characterised by new levels of complexity. CDA was part of the ‘turn to language’ that transformed the social sciences in the 1960s and 1970s. I argue that the ‘turn to language’ leads inexorably to this second turn. CDA today must accept the challenge of complexity, or it will condemn itself to irrelevance. 2.1 What is CDA? Origins are always indefinite, but in this paper I trace a trajectory from 1974, from what can be seen as the first manifesto of what became CDA (Hodge and Kress 1974), to an arbitrary present, 2010. The manifesto, an unheralded article in a little-known journal, declared the need for a new form of lingustics, to study language in a way that would ‘explore the relations between language 2 | Pag e CA D AA D and thought, language and society’ (1974: 5). As indication of CDA’s success today, there are two journals in the field, plus at least four others in which CDA is important. A recent Google visit had 270,000 hits for ‘Critical Discourse Analysis’, 110,000 for ‘Critical Discourse Studies’, with 146,000 for ‘Social Semiotics’, which I will argue is part of a common field. Google did not exist in 1974, but if it had we can be sure there would be no hits for any of these terms. That may look good, as it is, even if Google is an imprecise measure. However, another Google search for two terms in the CADAAD conference title, ‘Ideology’ and ‘Identity’, found 24,200,000 and 425,000,000 hits respectively. The numbers are rough guides, but the contrast is stark enough to draw a rough conclusion. These objects of analysis are far more salient than CDA is, as a means to study them. From this I draw a contradictory judgement on CDA I explore later in the article. CDA is both highly successful AND not nearly successful enough. Rather than begin by defining CDA I will deploy what I understand as a CDA perspective on CDA itself. This is an ‘ostensive’ definition: showing it in action. CDA aims to guide analysis, so ostensive definitions are especially suitable. I use two attributes to frame the demonstration: close attention to social functions and meanings, and scrutiny of features of linguistic form which other traditions treat as meaningless. ‘CDA’, for instance, can seem a neutral set of letters to refer to Critical Discourse Analysis, itself seen as semantically and socially equivalent to ‘critical analysis of discourse’, or ‘analysing discourse critically’. However, this transformational chain is motivated at every stage, and that motivation is a key to its social meaning. From this perspective, the acronym does a lot of work. It homogenises the practices that give the group its identity. The claimed homogeneity can then become a cover for heterogenous practices. It allows individuals to belong to the group without having to say what the group does or stands for. As I probe this piece of language, my analysis has social effects, challenging the cohesion of the group at the same time as it explores the basis of their and my belief system. There is a tension here between ‘identity’, understood as what gives groups a sense of being all ‘the same’, and ‘ideology’, beliefs that underpin that sense of unity. Both of these are in tension with ‘practice’, the set of engagements by many people facing different problems in a complex and contradictory world. All three component words, ‘Critical-Discourse-Analysis’, can be subjected to similar scrutiny. For instance, ‘critical’: who ‘criticises’ whom, and why? What social relationship is frozen and removed from scrutiny through this term? Given the left-wing affiliations that commonly identify this group, the referents might be understood to be ‘Left-wing’ criticizing the ‘Right’, but this binary is too simplistic. Could not the ‘Right’ equally use these forms of analysis against ‘the Left’? What mechanisms in the theory might stop that happening? ‘Critical’ could be referenced to the Marxist tradition, an ideological position in which ‘critical’ refers to analysis that sides with the oppressed, taking apart the ideological weapons of the oppressors carried through forms of language, Ho dg e Pag e | 3 in order to even the terms of the battle. But what if the oppressed (to use this binary for the moment) need a different form of help? To enhance their own communications, rather than demolish their enemies'? Relentless criticism might be irrelevant, or counter-productive. ‘Analysis’ as the term defining the approach has similar problems. It comes from Greek ana-lysis, breaking up something, loosening (lysis) bonds. Complementing ‘analysis’ is ‘synthesis’, putting things together. Combined with ‘critical’, analysis implies a destructive approach. But sustainable struggles need more scope and flexibility, to build and unite as well as destroy and take apart meanings, ideas, movements, people, in a world characterised by connections and alliances, local and global, not just struggle and difference. In a step towards inclusiveness, the new journal in this area calls itself Critical Discourse Studies, while CADAAD includes Critical Approaches. ‘Discourse’ proved a productive alternative to ‘language’, the earlier defining term (as in ‘Critical Linguistics’, Fowler et al 1979, ‘Critical Language Awareness’, Fairclough 1992). ‘Discourse’ had some decisive advantages over ‘language’ when it was first proposed (e.g. Van Dijk 1985, Fairclough 1989) because of different meanings it covered, contradictions it allowed. Compared with ‘language’ it included studies of processes and structures, language and thought, social processes and meanings in circulation. ‘Language’ brought another limitation into the study of social meaning, its default restriction to verbal language. Between ‘Critical Linguistics’ in 1974 and now, an information revolution transformed all media, introducing new conditions of communication for all forms studied by ‘Critical Analysis’. This drove the development of ‘Social Semiotics’ (Hodge and Kress 1988, Van Leeuwen 2005), which aimed to broaden the base of CDA, not be an alternative. ‘Discourse’ could have include meanings carried by all signifying systems, in all media: social semiosis. The struggles it was concerned with often took place in media systems, even more reason to take an inclusive view. However, the dominance of verbal language exerted powerful influence over ‘discourse’. CDA evolved with a boundary between verbal and other forms of language, creating an artificial separation from ‘Social Semiotics’. I argue that whatever CDA refers to, its object includes social semiotic phenomena, and whatever Social Semiotics refers to, it includes CDA. The difference between the terms is mostly political, to establish primacy in a single academic territory. Political differences, of course, matter for CDA and Social Semiotics. However, they should not be confused with a supposed difference in intellectual projects, nor get in the way of building alliances that both groups need. But the main force distorting both fields has been the role of the dominant Linguistics. This emphasised Language as a unitary object of study, separated from both society and thought, confined to the study of verbal language. This was the target of the 1974 manifesto quoted above. CDA challenged the exclusion of the social. It has a rich array of ways of studying social processes, especially power, as they act in ‘discourse’. However, like the mainstream Linguistics it supposedly rejects, CDA has problems with studying social meanings as allied to thought. Van Dijk (1985) insisted early on in including the study of cognitive structures in CDA, but he won few followers then. CDA 4 | Pag e CA D AA D cannot do the job it sets itself unless it can explore complex meanings that emerge in social interaction, and the complex processes which produce them, no less social for being located in minds. The key factor here has been the influence of Chomsky’s dominant brand of Linguistics, and a contradiction it transmitted. Chomsky declared Linguistics a branch of cognitive psychology, but in practice marginalised the study of language and thought from the discipline he dominated. This story is too complex for more than a brief summary here. It deserves a full study by a CDA approach, one with a strong cognitive component, which asks what key agents and typical actors thought and meant. In Linguistics the re-instatement of mind and meaning belatedly re-emerged in a minority branch, Cognitive Linguistics, which excited many participants in the 2010 CADAAD conference. As one example of the value of this direction, Chris Hart (2008) studied processes surrounding metaphors and shifting frames of meaning as applied to migrants in the British media in the election. Metaphors and shifts of meaning alike bring into the frame of analysis the kinds of instabilities that mark social interactions in times of stability as of crisis. Cognitive processes and meanings always underpin the seemingly inexorable operations of the power of the dominant. I hope it will be useful to others in CDA, broadly defined, for me to insist that this emphasis was central in earlier definitions of the field (Hodge and Kress 1974, Kress and Hodge 1979). Social meaning was a central object of study from the outset, represented in various texts and processes. Thought and cognitive processes were vital for study and analysis. This line of research does not need to be brought into CDA. It was always there. 2.2 Ideology, Identity, Interaction This triad of terms forming the theme of the CADAAD conference were well- chosen to provoke reflections on CDA. ‘Interaction’, the third of the three (171,000,000 Google hits) is the hardest to pin down, yet it has the greatest effect on the other two, and on CDA itself. Etymologically ‘discourse’ comes from Latin dis-currere, to run forwards and back, as in a chariot race, or in the passage of ideas or speech between two or more participants. Some kind of interaction is basic to ‘discourse’. The decisive break marked by ‘discourse’ versus ‘language’ was its incorporation of interaction, the dynamics of change and social process, compared to the huge, abstract static entity invoked by ‘language’. Interaction is subtly powerful and transformative. It introduces a dynamic perspective, challenging idealised, static structures with a different ontology, where process and function are what needs to be explained, and a new ‘post- structuralist’ epistemology loosens up monolithic structures, and irreducible complexity and chaos become ever-present possibilities. Theories that seek to understand social meanings in such a world need concepts and models that can cope with interactions on this scale. Interaction is a key driver for Urry’s ‘turn to complexity’ (2003). ‘Ideology’ has been a key term for CDA from the outset, a strong link with the Marxist tradition out of which it grew. Starting with Marx himself, that tradition defined this key term variously, (Williams 1974), but for most it was Ho dg e Pag e | 5 a coherent but false picture of reality, partial and distorted to serve or reflect the interests and assumptions of a particular group. Typical is Kress and Hodge’s (1979: 6) early formulation: ‘Ideology is a systematic body of ideas, organized from a particular point of view’. Fairclough followed this tradition (e.g. 1989: 2). A clear break with this tradition appeared in the concept of the Ideological complex: Ideological complexes [are] a functionally related set of contradictory versions of the world, coercively imposed by one social group on another on behalf of its own distinctive interests or subversively offered by another social group in attempts at resistance in its own interests. (Hodge and Kress 1988:3) What is important here is that unity or consistency in ideology is no longer expected. Contradiction is not occasional and accidental, but ubiquitous. It is not dysfunctional, but key to how ideology normally functions and achieves its effects. Contradictions come from the process of struggle, as meanings from the other are incorporated into discourse, in complex structures which risk incoherence to manipulate better. ‘Identity’ is a slippery term, posing many problems for its victims and for CDA. The concept of contradiction in a theory of ideological complexes is a key to understand how this potent term works. On one hand it refers to a unique, individual entity: e.g. Bob Hodge and no other. On the other hand it refers to total loss of that individual identity in a collective: e.g. Bob Hodge as Australian, identical to all other Australians. This sharp contradiction is highly functional. The same word, ‘identity’, applies to my unique individuality and to my ascription as standard-average- Australian-male. This implies that I am no less a unique individual for being Australian, no less Australian for being unique. I am also no less unique for being male, and no less male for being unique, and so on. It is a brilliant ideological move, almost too rapid to be seen, effective because not in spite of the fact that it is a bewildering contradiction. Etymology is a sadly neglected branch of Linguistics and CDA alike. Here as elsewhere it is a helpful way into understanding the complexity of the word today. ‘Identity’ comes from Latin idem-et-idem, ‘that one and that one’, or ‘the same and the same’. This history brings out the deictic basis of the word. Identity does not represent a quality, it points to elements in the world, in a primal act of classification. ‘Essentialist’ understandings of identity have been criticised, e.g. by Bhabha 1994, Nederveen Pieterse 2004, as problems in practice as well as theory, because they remove identity from social processes and fix it in rigid, non- negotiable eternal forms. Bhabha (1994) proposed the controversial term ‘hybridity’ to describe a new kind of identity he associated with late colonialism and a globalised world. ‘Hybridity’ puts multiplicity on the agenda, but still as a quality of individuals. A deictic concept of identity removes it from a connection with any supposed inner essence. Deictic identity comes from outside, from a pointing finger and a classifying gaze. It is part of an apparatus of control. Even strategic uses of it for resistance are shaped by that primary social fact. It has only such unity as is 6 | Pag e CA D AA D maintained by the identifying agency, which commonly seeks to remove all differences that complicate their control. Within any category, ‘sameness’ sometimes corresponds to recognised similarities, but it often masks relationships of complementarity or antagonism. A deictic concept of identity relocates contradictions, from what is represented to the social practices which achieve them: pointing a finger that can be mobile, and point in many different directions. Multiple identities are the rule not the exception, and have always been. Framing the situation in terms of the ideological complex, we can see how this contradiction makes ‘identity’ an especially convenient form for ideological use. 3. The Problem of Scale Critical Linguistics had no explicit model of the different scales at which processes of social meaning take place. Fairclough (1989: 25) introduced a simple, three-level scheme which has proved useful and influential, but 20 years later this framework needs to be strengthened and extended. In a complex, multiscalar world, processes go across levels of space and time (Gilmore 2002). Meanings are at play at every level. How are they to be captured and analysed? Without some ideas of what form these higher-level meanings take, and explicit analytic methods to track and interpret them, there is a danger that meanings at levels above the text will come only from the prejudices of the analyst. In the CAADAD conference I was struck by how many presenters used corpus linguistics to analyse data. This would not have happened 20 years ago. The way it is done now has problems of fit with older methods, but I believe these problems must be faced and overcome. To do so, CDA must develop a comprehensive, multiscalar model. The idea of ‘fractals’ from theories of chaos and complexity could play a role in such a model. According to Mandelbrot (1993) fractals are naturally-occurring self-similar non-Euclidean patterns across different scales or at the same scale. Mandelbrot claims that fractals at different levels have equivalent degrees of complexity. Large-scale patterns of discourse are not inherently more complex than smaller-scale patterns. So patterns at any one level, small or large, are guides to patterns at other levels. Multiscalar structures with many layers are richer than simplistic three-level models of sociology, micro, meso and macro, even in recent more complex forms (e.g. Foster and Potts 2009). I illustrate how fractal models can frame software analysis of corpus for multi- scalar CDA research. The program, Leximancer 3 (Smith and Humphreys 2006) is based on word-frequency algorithms, from which it builds up ‘concepts’, formed of high-frequency words which travel together, which algorithms identify as ‘themes’. Themes can be varied in size, from few (a higher scale of integration) to many. Thismulti-scalar analysis of levels of meaning can be reframed in fractal terms. The size and position of circles signify the size and relationship of the themes at this scale. The corpus I analyse in the figure below was the text of the 2010 CADAAD program, 88 speakers and their topics. The rationale was the fractally- Ho dg e Pag e | 7 informed hypothesis that each title would be the presenter's own micro- version of their fuller text, and that these 88 speakers, the CADAAD community attending the conference, would have a fractal (self-similar but not identical) relationship to the larger CDA community. The text I am analysing is a picture of a composite text, produced by a composite virtual speaker, as an authorised version of the larger composite text produced by them all in the conference. My analysis is a form of CDA, even though it is directed at a visual object, a computer-produced composite map, not at text produced directly by speakers. The map aggregates such speech acts, and the speakers’ collective identity is subsumed into the meta- identity of this meta-speaker, who is CDA itself. Figure 1. Leximancer map of CADAAD 2010 Figure 1 I will make a few observations on this text, to illustrate how this form of CDA may work. Firstly, I understand this text in interactional terms, as a response of this community to the Conference title text, Ideology, Identity, Interaction. The text of figure 1 shows that ‘ideology’ is a key term for the conference community, as it was for the committee, but ‘identity’ and ‘interaction’ are not. 8 | Pag e CA D AA D Instead, the semantic world of the community is distributed mainly between ‘ideology’, ‘media’ and ‘discourse’. Significantly, ‘media’ is more prominent than ‘discourse’, even in a conference on Critical Discourse Analysis. It is significant that these three main themes do not connect. This syntagmatic fact of the visual text suggests that the three themes organise distinct thematic universes, for distinct communities. This picture shows that those who referenced ideology in their titles did not reference either discourse or media. The same lack of connection characterises the other two terms. A specific pathway links the themes of discourse and media, and press (not discourse itself) with ideology, but the main message of the picture is disconnection. Disconnection is not full contradiction, but it identifies fissures in what would be otherwise understood as a single, cohesive field, giving a single cohesive identity. From this picture it seems that only those who frame their work around ‘discourse’ include political analysis strongly. Surprisingly, the concept of ideology is not strongly inflected politically. Nor is media analysis. Interpreted in terms of the concept of the ideological complex, these disconnects and contradictions do not unambiguously identify weaknesses in the field, or problems of identity. In this as in other cases of the kind of CDA I am advocating, fissures are diagnostic, showing tensions in the field, not fault lines about to open up. The key question with contradictions is: what function do they serve, in what dynamic condition of the field in question? 3. Analysing Identity One strength of CDA from its earliest days till now has been the practice of generating theory out of analytic practice. This has allowed the theory to grow by accumulation, becoming something richer than individual analysts could have hoped for, more contradictory than most would want to admit. I will illustrate how productive this strategy is by examining ‘identity’ and ‘ideology’ in a single, challenging textual instance. I came across the text in figure 2 by chance, as I wandered into a small cafe near Brighton, England. I could hardly believe my eyes as I read this front- page story. At a first glance, this text may seem too obvious to need analysis. It is racist in content and intent, clearly designed to arouse or reinforce racist sentiments in readers who share Clarkson's anti-Muslim sentiments, also guaranteed to anger people like myself with opposite views. CDA is not needed to show that British media and society carry large streaks of racism which generate offensive articles like this. Yet it can drill deeper, to bring out the complex, sometimes surprising processes which surround this kind of event, illuminating the non-linear causality which needs to be understood for effective interventions. CDA’s value as a heuristic device is I believe under- estimated and under-used. It is at its best in close readings of individual ideologically laden texts like this one, situated in an implicit or explicit fractal framework. Ho dg e Pag e | 9 Figure 2: Front page, Daily Star Wednesday July 28, 2010 In the form of CDA I am using, my own interaction is part of my reading. The content of the page constructs an opposition between male and female, British and Muslim, seemingly designed to connect with a presumed British male viewer/ reader through Clarkson's direct gaze and complicit smile, while the Muslim woman gazes into the distance. This would have been called the ‘preferred’ or ‘dominant’ reading by Marxist/Semiotician/Cultural Studies theorist Stuart Hall (1980) in a major contribution to CDA that is not usually seen as such. Hall developed a form of analysis which prefigured ideological complex theory, in which ‘dominant’ meanings co-exist with ‘negotiated’ and ‘oppositional’ meanings in an on-going struggle. In these terms I am male and English-speaking, but not British or Muslim, and hence an oppositional reader. In terms of ideological complex theory, however, my aberrant position can re-configure the key categories and identities which are at play in the ideological work of the image. In these terms, my role as an oppositional reader is not an accident. I am meant to be there, to be as irritated as I am. My expected irritation is part of the pleasure of this text for the dominant racist readers, re-inforcing their sense of dominance, constructing my own sense of being marginalised. Yet this trick, designed to co-opt my resistance, may not succeed. I may oppose on my own terms, aided by CDA. 10 | Pag e CA D AA D A fractal framework allows an intensive CDA/Social Semiotic reading of this text which for positivist forms of science may seem unscientific, but is experimental good-practice for non-linear science. This open, engaged kind of reading is called ‘Reading as analysis’ by Carbó (2002). CDA from the outset added new resources to reading/analysis by including meanings carried by features of language, often aspects of grammar, syntax or form, which act as auxiliary signs, that carry social meanings no less potent for being so often unconscious or invisible. This has been CDA’s major contribution to analysis of social meaning from the beginning, yet it has often been misunderstood as if it only involved a privileged and limiting relationship with Linguistics. Instead, it is a way of seeing the social meaning and effects of what is treated by most forms of Linguistics as semantically empty, socially disconnected aspects of form. I begin with the fact that this is clearly a multi-modal text in Kress and Van Leeuwen’s sense (2001). The text consists of words and image, and the words are visual images. The size of the type-face is the typographic equivalent of a shout, a complex, ambiguous statement declaring its importance but not its meaning, or more precisely, declaring that vehemence is more important than content. The black background is continuous with the woman’s burka, which makes it seem dangerous. Only the woman's eyes can be seen, seemingly just behind the words, part of the syntax yet also not part of the verbal text. Kress and Van Leeuwen attend to signs signified by the placement of elements of texts in an image, adapting Halliday’s categories of ‘given’ and ‘new’ (1985). In this case, the veiled woman is in the left (‘new’) slot in three lines of text: ‘Burka’ + eyes: ‘Babe’s’ + veil: ‘Undies’ + rest of Burka extending down to the feet, outside the frame of the image. These theorists see the main meanings communicated by this kind of sign as 'modality', the status or credibility of the message, (see Halliday 1985 and Hodge and Kress 1988). In this interpretation, the woman’s image is both content, an image containing important messages, and ‘modality’, affecting how it is interpreted. This whole text has low levels of logic. The sentence ‘Burka babe’s undies ad fury’ has at least three possible interpretations. In one the possessive refers to her undies. In the second it refers to the ad. In the third it refers to ‘fury’, fury felt by either her or another. As the story unfolds it is revealed that the third is what is being referred to. A video-clip for a cosmetics firm showed the woman putting clothes on till she was fully clothed, wearing a burka. The ‘story’ concerns hypothetical Muslim objections to the ad. However, this is not the ‘correct’ interpretation. The ambiguities are functional, and probably deliberate. They include suggestions which may lead salacious anti-Muslim males to buy the paper to see this ‘babe’s’ fury. In practice the single ‘correct’ interpretation is not correct. Contradiction and ambiguity are crucial to its effect, as it functions within a racist ideological complex. We can see contradictions present even in the smallest fractal level, in the first word, in the next fractal level up, the phrase ‘burka babe’s’, continuing into the phrase ‘burka babe’s undies’. ‘Burka’ uses a non-English word to refer to a form of clothing seen by Muslims and non-Muslims alike as a signifier of Muslim identity. Since it uses a Muslim term to refer to a Muslim practice it