The Field of Discourse Analysis PDF

3035-01.qxd 9/6/02 6:01 PM Page 1 1 The Field of Discourse Analysis For at least ten years now, ‘discourse’ has been a fashionable term. In scientific texts and debates, it is used indiscriminately, often without...

3035-01.qxd 9/6/02 6:01 PM Page 1 1 The Field of Discourse Analysis For at least ten years now, ‘discourse’ has been a fashionable term. In scientific texts and debates, it is used indiscriminately, often without being defined. The concept has become vague, either meaning almost nothing, or being used with more precise, but rather different, meanings in different contexts. But, in many cases, underlying the word ‘discourse’ is the general idea that language is structured according to different patterns that people’s utterances follow when they take part in different domains of social life, familiar examples being ‘medical discourse’ and ‘political discourse’. ‘Discourse analysis’ is the analysis of these patterns. But this common sense definition is not of much help in clarifying what discourses are, how they function, or how to analyse them. Here, more developed theories and methods of discourse analysis have to be sought out. And, in the search, one quickly finds out that discourse analysis is not just one approach, but a series of interdisciplinary approaches that can be used to explore many different social domains in many different types of studies. And there is no clear consensus as to what discourses are or how to analyse them. Different perspectives offer their own suggestions and, to some extent, compete to appropriate the terms ‘discourse’ and ‘discourse analysis’ for their own definitions. Let us begin, however, by proposing the preliminary definition of a discourse as a particular way of talking about and understanding the world (or an aspect of the world). In this chapter, three different approaches to social constructionist discourse analysis will be introduced – Ernesto Laclau and Chantal Mouffe’s discourse theory, critical discourse analysis, and discursive psychology. In the three following chapters, we will present the approaches individually. All three approaches share the starting point that our ways of talking do not neutrally reflect our world, identities and social rela- tions but, rather, play an active role in creating and changing them. We 3035-01.qxd 9/6/02 6:01 PM Page 2 2 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D have selected these approaches from the range of different perspectives within discourse analysis on the grounds that we think that they repre- sent particularly fruitful theories and methods for research in communi- cation, culture and society. They can be applied in analysis of many different social domains, including organisations and institutions, and in exploration of the role of language use in broad societal and cultural developments such as globalisation and the spread of mass mediated communication. Let us give a few examples of possible applications of discourse analy- sis. For instance, it can be used as a framework for analysis of national identity. How can we understand national identities and what conse- quences does the division of the world into nation states have? Many different forms of text and talk could be selected for analysis. The focus could be, for instance, the discursive construction of national identity in textbooks about British history. Alternatively, one could choose to explore the significance of national identity for interaction between people in an organisational context such as a workplace. Another research topic could be the ways in which expert knowledge is conveyed in the mass media and the implications for questions of power and democracy. How are claims to expert knowledge constructed and con- tested in the mass media and how are competing knowledge claims ‘consumed’ by media audiences? The struggle between different know- ledge claims could be understood and empirically explored as a struggle between different discourses which represent different ways of under- standing aspects of the world and construct different identities for speakers (such as ‘expert’ or ‘layperson’). The three approaches on which we have chosen to focus as frame- works for discourse analysis share certain key premises about how enti- ties such as ‘language’ and ‘the subject’ are to be understood. They also have in common the aim of carrying out critical research, that is, to investigate and analyse power relations in society and to formulate nor- mative perspectives from which a critique of such relations can be made with an eye on the possibilities for social change. At the same time, though, each perspective has a range of distinctive philosophical and theoretical premises, including particular understandings of discourse, social practice and critique, which lead to particular aims, methods and empirical focal points. The purpose of this introductory chapter is to outline the field to which social constructionist approaches to discourse analysis belong.1 We are interested both in those aspects which are common to all approaches – and, in particular, to our three approaches – and in those aspects in relation to which the approaches diverge. 3035-01.qxd 9/6/02 6:01 PM Page 3 T H E F I E L D O F D I S C O U R S E A N A LY S I S 3 The approaches are similar to one another in their social constructionist starting point, in their view of language, stemming from structuralist and poststructuralist linguistics, and in their understanding of the individual based on a version of structuralist Marxism. In this chapter, we will present these common roots and sources of theoretical inspiration, and during our account will touch on a series of concepts – for example, ‘power’ and ‘ideology’ – that often accompany the concept of discourse. Notwithstanding the shared premises, important differences exist between the approaches. First, there is disagreement as to the ‘scope’ of discourses: do they constitute the social completely, or are they them- selves partly constituted by other aspects of the social? Secondly, the approaches also vary with respect to their focus of analysis. Some analyse people’s discourse in everyday social interaction, others prefer a more abstract mapping of the discourses that circulate in society. We will elaborate on these points of divergence towards the end of the chapter. The division of the field into three approaches among which there are both similarities and differences should, to some extent, be understood as a construction of our own. We have picked out the three approaches and have chosen to allot one chapter to each and to compare and con- trast them to one another in Chapter 5, in order to provide a clear intro- duction to the field of discourse analysis. This representation should not be taken to be a neutral description or transparent reflection of the field. With respect to our choice of approaches, we cover only three approaches within the field of social constructionist discourse analysis, excluding, for example, the Foucauldian approach.2 And in relation to our identification of points of convergence and divergence among the three approaches, we acknowledge that comparison between the approaches is not a straightforward exercise. The three approaches emanate from different disciplines and have their own distinctive char- acteristics. At the same time, many discourse analysts work across disciplinary borders, and there are many theoretical points and method- ological tools that cannot be assigned exclusively to one particular approach. A C O M P L E T E PA C K A G E Although discourse analysis can be applied to all areas of research, it cannot be used with all kinds of theoretical framework. Crucially, it is not to be used as a method of analysis detached from its theoretical and 3035-01.qxd 9/6/02 6:01 PM Page 4 4 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D methodological foundations. Each approach to discourse analysis that we present is not just a method for data analysis, but a theoretical and methodological whole – a complete package. The package contains, first, philosophical (ontological and epistemological) premises regarding the role of language in the social construction of the world, second, theoreti- cal models, third, methodological guidelines for how to approach a research domain, and fourth, specific techniques for analysis. In dis- course analysis, theory and method are intertwined and researchers must accept the basic philosophical premises in order to use discourse analy- sis as their method of empirical study. It is important to stress that, while the content of the package should form an integrated whole, it is possible to create one’s own package by combining elements from different discourse analytical perspectives and, if appropriate, non-discourse analytical perspectives. Such multiperspectival work is not only permissible but positively valued in most forms of dis- course analysis. The view is that different perspectives provide different forms of knowledge about a phenomenon so that, together, they produce a broader understanding. Multiperspectival work is distinguished from an eclecticism based on a mishmash of disparate approaches without serious assessment of their relations with each other. Multiperspectivalism requires that one weighs the approaches up against each other, identifying what kind of (local) knowledge each approach can supply and modifying the approaches in the light of these considerations.3 In order to construct a coherent framework, it is crucial to be aware of the philosophical, theoretical and methodological differences and simi- larities among the approaches. Obviously, this requires an overview of the field. The aim of our presentation of the three perspectives in the following three chapters is to contribute to the acquisition of this overview by introducing the key features of three important discourse analytical approaches as well as the central themes in academic debates concerning these features. In addition, we will provide extensive references and suggestions for further reading. Ke y P re m i s e s The three approaches on which we have chosen to concentrate are all based on social constructionism.4 Social constructionism is an umbrella term for a range of new theories about culture and society.5 Discourse analysis is just one among several social constructionist approaches but it is one of the most widely used approaches within social constructionism.6 Furthermore, many use approaches that have the same characteristics as 3035-01.qxd 9/6/02 6:01 PM Page 5 T H E F I E L D O F D I S C O U R S E A N A LY S I S 5 those of discourse analysis without defining them as such. We will first provide a brief outline of the general philosophical assumptions that underpin most discourse analytical approaches, drawing on the accounts of social constructionism given by Vivien Burr (1995) and Kenneth Gergen (1985). Then we will focus specifically on the assumptions about language and identity that all discourse analytical approaches embrace. Burr (1995: 2) warns about the difficulty of giving one description that seeks to cover all social constructionist approaches, since they are so mani- fold and diverse. This notwithstanding, in Burr (1995: 2–5) she lists four premises shared by all social constructionist approaches, building on Gergen (1985). These premises are also embraced by our three approaches. They are as follows:7 A critical approach to taken-for-granted knowledge Our knowledge of the world should not be treated as objective truth. Reality is only accessible to us through categories, so our knowledge and representations of the world are not reflections of the reality ‘out there’, but rather are products of our ways of categorising the world, or, in dis- cursive analytical terms, products of discourse (Burr 1995: 3; Gergen 1985: 266–7). This premise will be explained further on (p. 9–12.) Historical and cultural specificity (Burr 1995: 3) We are fundamentally historical and cultural beings and our views of, and knowledge about, the world are the ‘products of historically situated interchanges among people’ (Gergen 1985: 267). Consequently, the ways in which we understand and represent the world are histori- cally and culturally specific and contingent: our worldviews and our identities could have been different, and they can change over time. This view that all knowledge is contingent is an anti-foundationalist position that stands in opposition to the foundationalist-view that knowledge can be grounded on a solid, metatheoretical base that tran- scends contingent human actions. Discourse is a form of social action that plays a part in producing the social world – including knowledge, identities and social relations – and thereby in maintaining specific social patterns. This view is anti-essentialist: that the social world is constructed socially and discursively implies that its character is not pre-given or determined by external conditions, and that people do not possess a set of fixed and authentic characteristics or essences. Link between knowledge and social processes Our ways of understanding the world are created and maintained by social processes (Burr 1995: 4; Gergen 1985: 268). Knowledge is cre- ated through social interaction in which we construct common truths and compete about what is true and false. 3035-01.qxd 9/6/02 6:01 PM Page 6 6 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D Link between knowledge and social action Within a particular worldview, some forms of action become natural, others unthinkable. Different social understandings of the world lead to different social actions, and therefore the social con- struction of knowledge and truth has social consequences (Burr 1995: 5, Gergen 1985: 268–269). Some critics of social constructionism have argued that if all knowledge and all social identities are taken to be contingent, then it follows that everything is in flux and there are no constraints and regularities in social life. There are certainly social constructionist theorists, such as Kenneth Gergen and Jean Baudrillard, who might be interpreted in this way. But, by and large, we believe that this is a caricature of social construction- ism. Most social constructionists, including adherents of our three approaches, view the social field as much more rule-bound and regula- tive. Even though knowledge and identities are always contingent in principle, they are always relatively inflexible in specific situations. Specific situations place restrictions on the identities which an individual can assume and on the statements which can be accepted as meaningful. We will resume this discussion in the next chapter in relation to Laclau and Mouffe’s discourse theory. The Three Approaches The key premises of social constructionism have roots in French post- structuralist theory and its rejection of totalising and universalising theories such as Marxism and psychoanalysis. But both social construc- tionism and poststructuralism are disputed labels and there is no con- sensus about the relationship between the two. We understand social constructionism as a broader category of which poststructuralism is a subcategory. All our discourse analytical approaches draw on struc- turalist and poststructuralist language theory, but the approaches vary as to the extent to which the poststructuralist label applies. Ernesto Laclau and Chantal Mouffe’s discourse theory, which we present in Chapter 2, is the ‘purest’ poststructuralist theory in our selec- tion. The theory has its starting point in the poststructuralist idea that discourse constructs the social world in meaning, and that, owing to the fundamental instability of language, meaning can never be permanently fixed. No discourse is a closed entity: it is, rather, constantly being trans- formed through contact with other discourses. So a keyword of the theory is discursive struggle. Different discourses – each of them representing 3035-01.qxd 9/6/02 6:01 PM Page 7 T H E F I E L D O F D I S C O U R S E A N A LY S I S 7 particular ways of talking about and understanding the social world – are engaged in a constant struggle with one other to achieve hegemony, that is, to fix the meanings of language in their own way. Hegemony, then, can provisionally be understood as the dominance of one particular per- spective. We will elaborate on this in Chapter 2. Critical discourse analysis, which we discuss in Chapter 3 with special focus on Norman Fairclough’s approach, also places weight on the active role of discourse in constructing the social world. But, in contrast to Laclau and Mouffe, Fairclough insists that discourse is just one among many aspects of any social practice. This distinction between discourse and non-discourse represents a remnant of more traditional Marxism in Fairclough’s theory, rendering critical discourse analysis less poststruc- turalist than Laclau and Mouffe’s discourse theory. A central area of interest in Fairclough’s critical discourse analysis is the investigation of change. Concrete language use always draws on earlier dis- cursive structures as language users build on already established meanings. Fairclough focuses on this through the concept of intertextuality – that is, how an individual text draws on elements and discourses of other texts. It is by combining elements from different discourses that concrete language use can change the individual discourses and thereby, also, the social and cultural world. Through analysis of intertextuality, one can investigate both the reproduction of discourses whereby no new elements are introduced and discursive change through new combinations of discourse. Discursive psychology, the subject of Chapter 4, shares critical discourse analysis’ empirical focus on specific instances of language use in social interaction. But the aim of discursive psychologists is not so much to analyse the changes in society’s ‘large-scale discourses’, which concrete language use can bring about, as to investigate how people use the avail- able discourses flexibly in creating and negotiating representations of the world and identities in talk-in-interaction and to analyse the social conse- quences of this. Despite the choice of label for this approach – ‘discursive psychology’ – its main focus is not internal psychological conditions. Discursive psychology is an approach to social psychology that has devel- oped a type of discourse analysis in order to explore the ways in which people’s selves, thoughts and emotions are formed and transformed through social interaction and to cast light on the role of these processes in social and cultural reproduction and change. Many discursive psycho- logists draw explicitly on poststructuralist theory, but with different results than, for example, Laclau and Mouffe. In discursive psychology, the stress is on individuals both as products of discourse and as producers of dis- course in specific contexts of interaction whereas Laclau and Mouffe’s dis- course theory tends to view individuals solely as subjects of discourse. 3035-01.qxd 9/6/02 6:01 PM Page 8 8 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D In Chapters 3 and 4 on respectively critical discourse analysis and discursive psychology, we set out the theoretical foundations and methodological guidelines for discourse analysis and present some con- crete examples of discourse analysis within each tradition. Laclau and Mouffe’s discourse theory, however, is short on specific methodological guidelines and illustrative examples. To compensate for this, we have extrapolated from their theory a range of analytical tools which we pre- sent in Chapter 2 together with an example of analysis based on some of these tools. The purpose of the guidelines and examples in the three chapters is to provide insight into how to apply the different approaches to discourse analysis in empirical work. In each of the chapters, we delin- eate the distinctive features of each perspective, whilst indicating the aspects which they share with one or both of the other perspectives. Throughout, we stress the links between theory and method. In Chapter 5, we home in on the theoretical and methodological differences and similarities among the approaches. We compare the approaches, weigh up their strengths and weaknesses, and point at ways in which they can supplement one other. Finally, we address some questions that are rele- vant to all the approaches. How do we delimit a discourse? How can we get started doing discourse analysis? How can we do multiperspectival research combining different discourse analytical approaches and differ- ent non-discourse analytical approaches? As in the other chapters, we present illustrative examples of ways of tackling these questions in empirical research. The final chapter of the book presents a discussion of the nature of critical research within the paradigm of social construc- tionism. Here, we discuss and evaluate a range of attempts to deal with the problems of doing critical research along social constructionist lines, focusing on their different stances in relation to the question of rela- tivism and the status of truth and knowledge.8 F RO M L A N G U AG E S Y S T E M TO D I S C O U R S E In addition to general social constructionist premises, all discourse analytical approaches converge with respect to their views of language and the subject. In order to provide a common base for the discussions in the coming chapters, we will now introduce the views that the approaches share followed by the main points of divergence. Discourse analytical approaches take as their starting point the claim of structuralist and poststructuralist linguistic philosophy, that our access to reality is always through language. With language, we create 3035-01.qxd 9/6/02 6:01 PM Page 9 T H E F I E L D O F D I S C O U R S E A N A LY S I S 9 representations of reality that are never mere reflections of a pre-existing reality but contribute to constructing reality. That does not mean that reality itself does not exist. Meanings and representations are real. Physical objects also exist, but they only gain meaning through discourse. Let us take as an example a flood associated with a river overflowing its banks. The rise in the water level that leads to the flood is an event that takes place independently of people’s thoughts and talk. Everybody drowns if they are in the wrong place, irrespective of what they think or say. The rise in the water level is a material fact. But as soon as people try to ascribe meaning to it, it is no longer outside discourse. Most would place it in the category of ‘natural phenomena’, but they would not necessarily describe it in the same way. Some would draw on a meteoro- logical discourse, attributing the rise in the water level to an unusually heavy downpour. Others might account for it in terms of the El Niño phenomenon, or see it as one of the many global consequences of the ‘greenhouse effect’. Still others would see it as the result of ‘political mis- management’, such as the national government’s failure to commission and fund the building of dykes. Finally, some might see it as a manifes- tation of God’s will, attributing it to God’s anger over a people’s sinful way of life or seeing it as a sign of the arrival of Armageddon. The rise in the water level, as an event taking place at a particular point in time, can, then, be ascribed meaning in terms of many different perspectives or discourses (which can also be combined in different ways). Importantly, the different discourses each point to different courses of action as possible and appropriate such as the construction of dykes, the organi- sation of political opposition to global environmental policies or the national government, or preparation for the imminent Armageddon. Thus the ascription of meaning in discourses works to constitute and change the world. Language, then, is not merely a channel through which information about underlying mental states and behaviour or facts about the world are communicated. On the contrary, language is a ‘machine’ that gener- ates, and as a result constitutes, the social world. This also extends to the constitution of social identities and social relations. It means that changes in discourse are a means by which the social world is changed. Struggles at the discursive level take part in changing, as well as in repro- ducing, the social reality. The understanding of language as a system, which is not determined by the reality to which it refers, stems from the structuralist linguistics that followed in the wake of Ferdinand de Saussure’s pioneering ideas around the beginning of this century. Saussure argued that signs consist 3035-01.qxd 9/6/02 6:01 PM Page 10 10 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D of two sides, form (signifiant) and content (signifié), and that the relation between the two is arbitrary (Saussure 1960). The meaning we attach to words is not inherent in them but a result of social conventions whereby we connect certain meanings with certain sounds. The sound or the writ- ten image of the word ‘dog’, for example, has no natural connection to the image of a dog that appears in our head when we hear the word. That we understand what others mean when they say ‘dog’ is due to the social convention that has taught us that the word ‘dog’ refers to the four-legged animal that barks. Saussure’s point is that the meaning of individual signs is determined by their relation to other signs: a sign gains its specific value from being different from other signs. The word ‘dog’ is different from the words ‘cat’ and ‘mouse’ and ‘dig’ and ‘dot’. The word ‘dog’ is thus part of a network or structure of other words from which it differs; and it is precisely from everything that it is not that the word ‘dog’ gets its meaning. Saussure saw this structure as a social institution and therefore as changeable over time. This implies that the relationship between lan- guage and reality is also arbitrary, a point developed in later structural- ist and poststructuralist theory. The world does not itself dictate the words with which it should be described, and, for example, the sign ‘dog’ is not a natural consequence of a physical phenomenon. The form of the sign is different in different languages (for example, ‘chien’ and ‘Hund’), and the content of the sign also changes on being applied in new situa- tions (when, for example, saying to a person, ‘you’re such a dog’). Saussure advocated that the structure of signs be made the subject matter of linguistics. Saussure distinguished between two levels of language, langue and parole. Langue is the structure of language, the network of signs that give meaning to one another, and it is fixed and unchangeable. Parole, on the other hand, is situated language use, the signs actually used by people in specific situations. Parole must always draw on langue, for it is the structure of language that makes specific statements possible. But in the Saussurian tradition parole is often seen as random and so vitiated by people’s mistakes and idiosyncrasies as to disqualify it as an object of scientific research. Therefore, it is the fixed, underlying structure, langue, which has become the main object of linguistics. Poststructuralism takes its starting point in structuralist theory but modi- fies it in important respects. Poststructuralism takes from structuralism the idea that signs derive their meanings not through their relations to real- ity but through internal relations within the network of signs; it rejects structuralism’s view of language as a stable, unchangeable and totalising structure and it dissolves the sharp distinction between langue and parole. 3035-01.qxd 9/6/02 6:01 PM Page 11 T H E F I E L D O F D I S C O U R S E A N A LY S I S 11 First we turn to the poststructuralist critique of the stable, unchangeable structure of language. As we have mentioned, in Saussure’s theory, signs acquire their meaning by their difference from other signs. In the Saussurian tradition, the structure of language can be thought of as a fishing-net in which each sign has its place as one of the knots in the net. When the net is stretched out, the knot is fixed in position by its distance from the other knots in the net, just as the sign is defined by its distance from the other signs. Much of structuralist theory rests on the assump- tion that signs are locked in particular relationships with one another: every sign has a particular location in the net and its meaning is fixed. Later structuralists and poststructuralists have criticised this conception of language; they do not believe that signs have such fixed positions as the metaphor of the fishing-net suggests. In poststructuralist theory, signs still acquire their meaning by being different from other signs, but those signs from which they differ can change according to the context in which they are used (see Laclau 1993a: 433). For instance, the word ‘work’ can, in certain situations, be the opposite of ‘leisure’ whereas, in other contexts, its opposite is ‘passivity’ (as in ‘work in the garden’). It does not follow that words are open to all meanings – that would make language and communication impossible – but it does have the conse- quence that words cannot be fixed with one or more definitive mean- ing(s). The metaphor of the fishing-net is no longer apt since it cannot be ultimately determined where in the net the signs should be placed in rela- tion to one another. Remaining with the metaphor of ‘net’, we prefer to use the internet as a model, whereby all links are connected with one another, but links can be removed and new ones constantly emerge and alter the structure. Structures do exist but always in a temporary and not necessarily consistent state. This understanding provides poststructuralism with a means of solving one of structuralism’s traditional problems, that of change. With structuralism’s focus on an underlying and fixed structure, it is impossible to understand change, for where would change come from? In poststructuralism, the structure becomes changeable and the meanings of signs can shift in relation to one another. But what makes the meanings of signs change? This brings us to post- structuralism’s second main critique of traditional structuralism, bearing on the latter’s sharp distinction between langue and parole. As mentio- ned, parole cannot be an object of structuralist study because situated language use is considered too arbitrary to be able to say anything about the structure, langue. In contrast to this, poststructuralists believe that it is in concrete language use that the structure is created, reproduced and changed. In specific speech acts (and writing), people draw on the 3035-01.qxd 9/6/02 6:01 PM Page 12 12 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D structure – otherwise speech would not be meaningful – but they may also challenge the structure by introducing alternative ideas for how to fix the meaning of the signs. Not all discourse analytical approaches subscribe explicitly to post- structuralism, but all can agree to the following main points: Language is not a reflection of a pre-existing reality. Language is structured in patterns or discourses – there is not just one general system of meaning as in Saussurian structuralism but a series of systems or discourses, whereby meanings change from discourse to discourse. These discursive patterns are maintained and transformed in discur- sive practices. The maintenance and transformation of the patterns should therefore be explored through analysis of the specific contexts in which language is in action. F o u c a u l t ’s A r c h a e o l o g y a n d G e n e a l o g y Michel Foucault has played a central role in the development of discourse analysis through both theoretical work and empirical research. In almost all discourse analytical approaches, Foucault has become a figure to quote, relate to, comment on, modify and criticise. We will also touch on Foucault, sketching out his areas of contribution to discourse analysis – not only in order to live up to the implicit rules of the game, but also because all our approaches have roots in Foucault’s ideas, while rejecting some parts of his theory. Traditionally, Foucault’s work is divided between an early ‘archaeo- logical’ phase and a later ‘genealogical’ phase, although the two overlap, with Foucault continuing to use tools from his archaeology in his later works. His discourse theory forms part of his archaeology. What he is interested in studying ‘archaeologically’ are the rules that determine which statements are accepted as meaningful and true in a particular historical epoch. Foucault defines a discourse as follows: We shall call discourse a group of statements in so far as they belong to the same discursive formation […Discourse] is made up of a limited number of statements for which a group of conditions of existence can be defined. Discourse in this sense is not an ideal, timeless form […] it is, from beginning to end, historical – a fragment of history […] posing its own limits, its divisions, its transformations, the specific modes of its temporality. (Foucault 1972: 117) 3035-01.qxd 9/6/02 6:01 PM Page 13 T H E F I E L D O F D I S C O U R S E A N A LY S I S 13 Foucault adheres to the general social constructionist premise that knowledge is not just a reflection of reality. Truth is a discursive con- struction and different regimes of knowledge determine what is true and false. Foucault’s aim is to investigate the structure of different regimes of knowledge – that is, the rules for what can and cannot be said and the rules for what is considered to be true and false. The starting point is that although we have, in principle, an infinite number of ways to formulate statements, the statements that are produced within a specific domain are rather similar and repetitive. There are innumerable statements that are never uttered, and would never be accepted as meaningful. The histori- cal rules of the particular discourse delimit what it is possible to say.9 The majority of contemporary discourse analytical approaches follow Foucault’s conception of discourses as relatively rule-bound sets of state- ments which impose limits on what gives meaning. And they build on his ideas about truth being something which is, at least to a large extent, created discursively. However, they all diverge from Foucault’s tendency to identify only one knowledge regime in each historical period; instead, they operate with a more conflictual picture in which different discourses exist side by side or struggle for the right to define truth. In his genealogical work, Foucault developed a theory of power/ knowledge. Instead of treating agents and structures as primary cate- gories, Foucault focuses on power. In common with discourse, power does not belong to particular agents such as individuals or the state or groups with particular interests; rather, power is spread across different social practices. Power should not be understood as exclusively oppres- sive but as productive; power constitutes discourse, knowledge, bodies and subjectivities: What makes power hold good, what makes it accepted, is simply the fact that it does not only weigh on us as a force that says no, but that it traverses and pro- duces things, it induces pleasure, forms knowledge, produces discourse. It needs to be considered as a productive network which runs through the whole social body, much more than as a negative instance whose function is repression. (Foucault 1980: 119) Thus power provides the conditions of possibility for the social. It is in power that our social world is produced and objects are separated from one another and thus attain their individual characteristics and relationships to one another. For instance, ‘crime’ has gradually been created as an area with its own institutions (e.g. prisons), particular subjects (e.g. ‘criminals’) and particular practices (e.g. ‘resocialisation’). And power is always bound up with knowledge – power and knowledge 3035-01.qxd 9/6/02 6:01 PM Page 14 14 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D presuppose one another. For example, it is hard to imagine the modern prison system without criminology (Foucault 1977). Power is responsible both for creating our social world and for the particular ways in which the world is formed and can be talked about, ruling out alternative ways of being and talking. Power is thus both a productive and a constraining force. Foucault’s conception of power is adhered to by Laclau and Mouffe’s discourse theory and discursive psychology, while critical discourse analysis is more ambivalent towards it. We discuss the position of critical discourse analysis in Chapter 3. With respect to knowledge, Foucault’s coupling of power and knowl- edge has the consequence that power is closely connected to discourse. Discourses contribute centrally to producing the subjects we are, and the objects we can know something about (including ourselves as subjects). For all the approaches, adherence to this view leads to the following research question: how is the social world, including its subjects and objects, constituted in discourses? Foucault’s concept of power/knowledge also has consequences for his conception of truth. Foucault claims that it is not possible to gain access to universal truth since it is impossible to talk from a position outside dis- course; there is no escape from representation. ‘Truth effects’ are created within discourses. In Foucault’s archaeological phase, ‘truth’ is understood as a system of procedures for the production, regulation and diffusion of statements. In his genealogical phase, he makes a link between truth and power, arguing that ‘truth’ is embedded in, and produced by, systems of power. Because truth is unattainable, it is fruitless to ask whether some- thing is true or false. Instead, the focus should be on how effects of truth are created in discourses. What is to be analysed are the discursive processes through which discourses are constructed in ways that give the impression that they represent true or false pictures of reality. THE SUBJECT It is also Foucault who provided the starting point for discourse analy- sis’ understanding of the subject. His view is, as already noted, that subjects are created in discourses. He argues that ‘discourse is not the majestically unfolding manifestation of a thinking, knowing, speaking subject’ (Foucault 1972: 55). Or as Steinar Kvale expresses the position, ‘The self no longer uses language to express itself; rather language speaks through the person. The individual self becomes a medium for the culture and its language’ (Kvale 1992: 36). 3035-01.qxd 9/6/02 6:01 PM Page 15 T H E F I E L D O F D I S C O U R S E A N A LY S I S 15 This is very different from the standard Western understanding of the subject as an autonomous and sovereign entity. According to Foucault, the subject is decentred. Here, Foucault was influenced by his teacher, Louis Althusser. Althusser’s structural Marxist approach links the subject closely to ideology: the individual becomes an ideological subject through a process of interpellation whereby discourses appeal to the individual as a subject. First, we will outline Althusser’s understanding of ideology and, following that, his understanding of interpellation. Althusser defines ideology as a system of representations that masks our true rela- tions to one another in society by constructing imaginary relations between people and between them and the social formation (Althusser 1971). Thus ideology is a distorted recognition of the real social relations. According to Althusser, all aspects of the social are controlled by ideol- ogy, which functions through ‘the repressive state apparatus’ (e.g. the police) and ‘the ideological state apparatus’ (e.g. the mass media). Interpellation denotes the process through which language constructs a social position for the individual and thereby makes him or her an ideo- logical subject: [I]deology ‘acts’ or ‘functions’ in such a way that it ‘recruits’ subjects among the individuals (it recruits them all), or ‘transforms’ the individuals into subjects (it transforms them all) by that very precise operation which I have called inter- pellation or hailing, and which can be imagined along the lines of the most com- monplace everyday police (or other) hailing: ‘Hey, you there!’ Assuming that the theoretical scene I have imagined takes place in the street, the hailed individual will turn round – […] he becomes a subject. (Althusser 1971: 174; italics in original, note omitted) Let us take as an example public information material about health in late modernity, which interpellates readers as consumers with personal responsibility for the care of their bodies through a proper choice of lifestyle. By accepting the role as addressees of the text, we affiliate our- selves to the subject position that the interpellation has created. In so doing, we reproduce the ideology of consumerism and our position as subjects in a consumer culture. By taking on the role of subject in a consumer culture, we accept that certain problems are constructed as personal problems that the individual carries the responsibility for solving, instead of as public problems that demand collective solutions. Althusser assumes that we always accept the subject positions allo- cated to us and thereby become subjects of ideology; there is no chance of resistance: 3035-01.qxd 9/6/02 6:01 PM Page 16 16 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D Experience shows that the practical telecommunications of hailings is such that they hardly ever miss their man: verbal call or whistle, the one hailed always recognizes that it is really him who is being hailed. (Althusser 1971: 174) As we are going to see in the section below, this is just one of the aspects of Althusser’s theory which has been subjected to heavy criticism by many including by the majority of discourse analytical approaches. Rejection of Determinism Althusser’s theory had a great influence on cultural studies approaches to communication studies in the 1970s. The research focus was on texts (mainly mass media texts), not on text production or reception since researchers took the ideological workings and effects of texts for granted. Meanings were treated as if they were unambiguously embed- ded in texts and passively decoded by receivers. To a large extent, cultural studies – strongly influenced by Althusser – was based on the idea that a single ideology (capitalism) was dominant in society, leaving no real scope for effective resistance (the ‘dominant ideology thesis’). But since the end of the 1970s, Althusser’s perspective has been criti- cised in several ways. First, the question was raised as to the possibilities for resistance against the ideological messages that are presented to the subject – the question of the subject’s agency or freedom of action. The media group at the Centre for Contemporary Cultural Studies in Birmingham, led by Stuart Hall, pointed, in this respect, to the complex- ity of media reception (Hall et al. 1980). According to Hall’s ‘encoding/decoding’ theory, recipients were able to interpret or ‘decode’ messages by codes other than the code which was ‘encoded’ in the text (Hall 1980). The theory was based inter alia on Gramsci’s theory of hegemony, which ascribes a degree of agency to all social groups in the production and negotiation of meaning (Gramsci 1991). Today there is a consensus in cultural studies, communication research and discourse analysis that the dominant ideology thesis underestimated people’s capa- city to offer resistance to ideologies. Some contributions to communica- tions and cultural studies may even tend to overestimate people’s ability to resist media messages (see, for example, Morley 1992 for a critique of this tendency), but usually discourse analysts take into account the role of textual features in setting limits on how the text can be interpreted by its recipients. Second, all three of the discourse analytical approaches presented in our book reject the understanding of the social as governed by one 3035-01.qxd 9/6/02 6:01 PM Page 17 T H E F I E L D O F D I S C O U R S E A N A LY S I S 17 totalising ideology. Just as they replace Foucault’s monolithic view of knowledge regimes with a more pluralistic model in which many dis- courses compete, they dismiss Althusser’s theory that one ideology con- trols all discourse. It follows from this that subjects do not become interpellated in just one subject position: different discourses give the subject different, and possibly contradictory, positions from which to speak. The different approaches have developed different concepts of the subject which we will discuss in the following chapters. But generally speaking, it can be said that all the approaches see the subject as created in discourses – and therefore as decentred – the constitution of subjects being a key focus of empirical analysis. However, the approaches differ as to the degree of emphasis given to the subject’s ‘freedom of action’ within the discourse – that is, they differ as to their position in the debate about the relationship between structure and agent. Laclau and Mouffe’s discourse theory largely follows Foucault, viewing the individual as determined by structures, whereas critical discourse analysis and discur- sive psychology to a greater extent are in line with Roland Barthes’ slogan that people are both ‘masters and slaves of language’ (Barthes 1982). Thus the latter two approaches stress that people use discourses as resources with which they create new constellations of words – sentences that have never before been uttered. In talk, language users select elements from different discourses which they draw on from mass mediated and interpersonal communication. This may result in new hybrid dis- courses. Through producing new discourses in this way, people function as agents of discursive and cultural change. As the critical discourse analyst, Fairclough, expresses it, ‘Individual creative acts cumulatively establish restructured orders of discourse’ (1989: 172). However, even in those approaches in which the subject’s agency and role in social change are brought to the foreground, discourses are seen as frameworks that limit the subject’s scope for action and possibilities for innovation. Critical discourse analysis and discursive psychology each present a theoretical foundation and specific methods for analysis of the dynamic discursive practices through which language users act as both discursive products and producers in the reproduction and transformation of discourses and thereby in social and cultural change. The third and final controversial point in Althusser’s theory is the concept of ideology itself. Most concepts of ideology, including Althusser’s, imply that access to absolute truth is attainable. Ideology distorts real social relations, and, if we liberated ourselves from ideology, we would gain access to them and to truth. As we saw, this is an understanding that Foucault rejects completely. According to Foucault, truth, subjects and 3035-01.qxd 9/6/02 6:01 PM Page 18 18 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D relations between subjects are created in discourse, and there is no possibility of getting behind the discourse to a ‘truer’ truth. Hence Foucault has no need of a concept of ideology. Laclau and Mouffe’s discourse theory has adopted this position, and its concept of ideology is practically empty. In contrast, critical discourse analysis and discursive psychology do not reject the Marxist tradition completely on this point: both approaches are interested in the ideological effects of discursive prac- tices. While they adhere to Foucault’s view of power, treating power as productive rather than as pure compulsion, they also attach importance to the patterns of dominance, whereby one social group is subordinated to another. The idea is also retained – at least, in Fairclough’s critical dis- course analysis – that one can distinguish between discourses that are ideological and discourses that are not, thus retaining the hope of find- ing a way out of ideology; a hope that Laclau and Mouffe’s discourse theory would find naïve. D I F F E R E N C E S B E T W E E N T H E A P P ROAC H E S The divergence in the way in which ideology is conceived is just one of the differences between the three approaches. In the following section, we highlight differences between the approaches with respect to, first, the role of discourse in the construction of the world and, second, analytical focus. In both these respects, the differences are matters of degree, and we will position the approaches in relation to each other on two continua to which we will refer throughout the rest of this book. The Role of Discourse in the Constitution o f t h e Wo r l d For all three approaches, the functioning of discourse – discursive prac- tice – is a social practice that shapes the social world. The concept of ‘social practice’ views actions in terms of a dual perspective: on the one hand, actions are concrete, individual and context bound; but, on the other hand, they are also institutionalised and socially anchored, and because of this tend towards patterns of regularity. Fairclough’s critical discourse analysis reserves the concept of discourse for text, talk and other semiological systems (e.g. gestures and fashion) and keeps it distinct from other dimensions of social practice. Discursive practice is viewed as 3035-01.qxd 9/6/02 6:01 PM Page 19 T H E F I E L D O F D I S C O U R S E A N A LY S I S 19 one dimension or moment of every social practice in a dialectical relationship with the other moments of a social practice. That means that some aspects of the social world function according to different logics from discourses and should be studied with tools other than those of discourse analysis. For instance, there may be economic logics at play or the institutionalisation of particular forms of social action. Discursive practice reproduces or changes other dimensions of social practice just as other social dimensions shape the discursive dimension. Together, the discursive dimension and the other dimensions of social practice constitute our world. Laclau and Mouffe’s discourse theoretical approach does not distin- guish between discursive and non-discursive dimensions of the social – practices are viewed as exclusively discursive. That does not mean that nothing but text and talk exist, but, on the contrary, that discourse itself is material and that entities such as the economy, the infrastructure and institutions are also parts of discourse. Thus, in Laclau and Mouffe’s dis- course theory there is no dialectical interaction between discourse and something else: discourse itself is fully constitutive of our world. This difference can be concretised by locating the approaches on a continuum. We have placed in brackets some of the other positions to which we refer in the book. On the left-hand side, discourse is seen as fully constitutive of the social, whereas on the right-hand side discourses are seen as mere reflections of other social mechanisms. A schematic figure like this has to be approached cautiously since the complexity of the actual theories is bound to be reduced when they are placed on a single line. This is clear, for example, in the case of the posi- tioning of discursive psychology. We have placed discursive psychology somewhat to the left on the continuum, but it is, in fact, difficult to place, as it claims both that discourse is fully constitutive and that it is embed- ded in historical and social practices, which are not fully discursive. The approaches on the far right of the continuum are not discourse analytical. If one claims, as they do, that discourse is just a mechanical reproduction of other social practices – that is, discourse is fully deter- mined by something else such as the economy – then there is no point in doing discourse analysis; instead, effort should be invested in economic analysis, for example. We have, therefore, judged the different Marxist positions on the right-hand side of the continuum according to a principle that does not quite do them justice: neither historical materialism nor cultural Marxists such as Gramsci and Althusser, have worked with ‘dis- course’ or ‘discourse analysis’, so their inclusion is based on both an interpretation and a reduction of their theories. Moreover, both Gramsci 3035-01.qxd 9/6/02 6:01 PM Page 20 20 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D Discourse is constitutive Dialectical relationship Discourse is constituted Critical discourse analysis Laclau and Mouffe’s Discursive discourse theory psychology (Foucault) (Althusser) (Gramsci) (Historical materialism) Figure 1.1 The role of discourse in the constitution of the world and Althusser leave a rather large latitude for meaning-making practices that can be interpreted as a discursive dimension. But both of them see the economy as determinant in the ‘final instance’, and that is why they have ended up so far to the right. Analytical Focus Some approaches focus on the fact that discourses are created and changed in everyday discursive practices and therefore stress the need for systematic empirical analyses of people’s talk and written language in, for instance, the mass media or research interviews. Other approaches are more concerned with general, overarching patterns and aim at a more abstract mapping of the discourses that circulate in society at a par- ticular moment in time or within a specific social domain. On a continuum, these differences can be represented as follows: Everyday discourse Abstract discourse Discursive psychology Critical discourse analysis Laclau and Mouffe’s discourse theory (Foucault) Figure 1.2 Analytical Focus On this continuum, the focus is on differences of degree rather than qualitative differences. Although discursive psychology focuses on people’s everyday practice, it constantly implicates larger societal struc- tures on which people draw, or transform, in discursive practice. And although Laclau and Mouffe’s discourse theory is mostly interested in more abstract, ‘depersonified’ discourses, the idea that these discourses are created, maintained and changed in myriads of everyday practices is implicit in the theory. 3035-01.qxd 9/6/02 6:01 PM Page 21 T H E F I E L D O F D I S C O U R S E A N A LY S I S 21 But, at the same time, the positions of the different approaches on the continuum reflect differences in theoretical emphasis: discursive psycho- logy is much more interested in people’s active and creative use of dis- course as a resource for accomplishing social actions in specific contexts of interaction than Laclau and Mouffe’s discourse theory, which instead is interested in how discourses, more generally, limit our possibilities for action. T H E R O L E O F T H E A N A LY S T For the discourse analyst, the purpose of research is not to get ‘behind’ the discourse, to find out what people really mean when they say this or that, or to discover the reality behind the discourse. The starting point is that reality can never be reached outside discourses and so it is discourse itself that has become the object of analysis. In discourse analytical research, the primary exercise is not to sort out which of the statements about the world in the research material are right and which are wrong (although a critical evaluation can be carried out at a later stage in the analysis). On the contrary, the analyst has to work with what has actually been said or written, exploring patterns in and across the state- ments and identifying the social consequences of different discursive representations of reality. In working with discourses close to oneself with which one is very familiar, it is particularly difficult to treat them as discourses, that is, as socially constructed meaning-systems that could have been different. Because analysts are often part of the culture under study, they share many of the taken-for-granted, common-sense understandings expressed in the material. The difficulty is that it is precisely the common-sense understandings that are to be investigated: analysis focuses on how some statements are accepted as true or ‘naturalised’, and others are not. Consequently, it is fruitful to try to distance oneself from one’s material and, for instance, imagine oneself as an anthropologist who is exploring a foreign universe of meaning in order to find out what makes sense there. But this suggestion to play anthropologist should just be seen as a useful starting point rather than a full response to the problem of the researcher’s role. If the research project is based on a social construc- tionist perspective, the problem of the researcher’s role goes much deeper and needs to be tackled reflexively. If we accept that ‘reality’ is socially created, that ‘truths’ are discursively produced effects and that subjects are decentred, what do we do about the ‘truth’ that we as 3035-01.qxd 9/6/02 6:01 PM Page 22 22 D I S C O U R S E A N A LY S I S A S T H E O R Y A N D M E T H O D researcher–subjects produce? This problem is intrinsic to all social constructionist approaches. Of the approaches that we present, the problem of how to deal with the contingency of truth is most pertinent in Laclau and Mouffe’s discourse theory and discursive psychology, and the two approaches solve it in different ways. The problem is largely ignored by Laclau and Mouffe, their theory and analysis being presented as if they were objective descriptions of the world and its mechanisms. In contrast, discursive psychology tries to take account of the role of the analyst through different forms of reflexivity (see Chapters 4 and 6). By comparison with Laclau and Mouffe’s discourse theory and discursive psychology, the dilemma does not at first glance seem so urgent in Fairclough’s critical discourse analysis because he makes a distinction between ideological and non-ideological discourses: in principle, the researcher ought to be able to produce non- ideological discourses. But the problem re-emerges with the question of how to distinguish between what is ideological and what is not, and the question of who is sufficiently liberated from the discursive construction of the world to make this distinction. Philosophically speaking, the problem appears insoluble, if we accept the anti-foundationalist premise, underpinning social constructionism, that it is a condition of all knowledge that it is just one representation of the world among many other possible representations. The researcher always takes a position in relation to the field of study, and that position plays a part in the determination of what he or she can see and can present as results. And there are always other positions in terms of which reality would look different. But that does not mean that all research results are equally good. In Chapter 4, we discuss how, with a social con- structionist starting point, research results can be validated and made as transparent as possible for the reader. Generally, theoretical consistency demands that discourse analysts consider and make clear their position in relation to the particular discourses under investigation and that they assess the possible consequences of their contribution to the discursive production of our world. The relativism inherent in social constructionism does not mean, either, that the analyst cannot be critical. All our approaches regard themselves as critical and in Chapter 6 we discuss at length how it is pos- sible to practise social criticism without being able to make claims to absolute truth. In brief, our position is that it is the stringent application of theory and method that legitimises scientifically produced knowledge. It is by seeing the world through a particular theory that we can distance ourselves from some of our taken-for-granted understandings and subject our 3035-01.qxd 9/6/02 6:01 PM Page 23 T H E F I E L D O F D I S C O U R S E A N A LY S I S 23 material to other questions than we would be able to do from an everyday perspective. The next three chapters can be seen as different ways to achieve this distance, and in Chapter 6 we contextualise the discussions of scientific knowledge, reflexivity and critique within the wider field of social constructionism. N OT E S 1 However, this field does not cover all uses of the label ‘discourse analysis’. The term ‘discourse analysis’ for example, is used in linguistics to denote the analysis of relations between sentences and statements on the micro level (for example, Brown and Yule 1983). Discourse analysis has also been used to denote the analysis of the ways people use mental schemata to understand narratives (van Dijk and Kintch 1983). 2 For accounts of Foucauldian forms of discourse analysis see, for example, Howarth (2000) and Mills (1997). 3 See Kellner (1995) for a call for ‘multiperspectival cultural studies’. And see Chapter 5 in this book for a discussion and illustration of multiperspectival discourse analysis. 4 What we call ‘social constructionism’ in this text is in many other connec- tions labelled ‘social constructivism’. We use the term ‘social constructionism’ to avoid confusion with Piaget’s constructivist theory (see Burr 1995: 2). 5 For discussions of the philosophical foundations of social constructionism see, for example, Collin (1997). 6 The dominance of discourse analysis is manifested in Burr’s introduction to social constructionism (Burr 1995), in which her examples of empirical research consist exclusively of forms of discourse analysis, notwithstanding the fact that she emphasises that social constructionists also use other approaches. 7 Here, we draw both on Burr (1995) and Gergen (1985). Burr’s account, as noted above, is also based on that of Gergen. 8 As authors, we have collaborated on all of the book’s chapters and have developed together many of the ideas and formulations throughout the book. However, main responsibility can be attributed in the following way: Louise Phillips for Chapters 3 and 4, and Marianne Jørgensen for Chapters 2 and 6, while both authors are equally responsible for Chapters 1 and 5. 9 Foucault’s own works from the archaeological period include both more abstract presentations of his theory and methodological tools (e.g. Foucault 1972) and empirical analyses (e.g. Foucault 1973, 1977).