Navigation – Plan du site

Discursive psychology, rhetoric and the issue of agency

[Texte anglais original]
Michael Billig
Traduction(s) :
La psychologie discursive, la rhétorique et la question de l’agentivité1

Texte intégral

1Recent years have seen the growth of a movement which has aimed to re-orientate the discipline of psychology around the study of discourse. Generally known as discursive psychology, this movement has had a large effect on social psychology, particularly in the United Kingdom, Australia, Canada, Scandinavia and South America. Its effect in the United States has been less marked. Discursive psychology is part of the general movement of critical psychology, which has been reacting against mainstream social psychology, especially the sort of experimental psychology that remains dominant in the United States (see, for instance, Gough and McFadden, 2001; Gergen, 2001; Hepburn, 2003; Sloan, 2000). Discursive psychologists have not been content merely to criticise mainstream psychology. They have proposed alternative ways of conducting psychological research, shifting the balance from the quantitative methodologies to qualitative ones, together with a commitment to studying human psychology through the use of language. Discursive psychology, however, represents more than a methodological alternative to conventional psychology. It rests upon an attempt to establish new, theoretical principles. At its heart lies a very different conception of language than that which is accepted by most mainstream psychologists, especially cognitive psychologists.

2In this paper, I will outline some of the principles of discursive psychology, especially in relation to its reaction against cognitive psychology. I will seek to show why philosophically the position of discursive psychology entails a critique of cognitivism, and I will describe attempts to use discursive principles to reformulate psychoanalytic theory. In addition, I will discuss the problem of agency, especially how it applies to discursive psychology.  I will suggest that the problem of agency is essentially a rhetorical issue rather than a theoretical or methodological one. Calling it a rhetorical problem does not imply that it is a lesser problem – as if it were ‘mere rhetoric’, and, thereby, of lesser significance than theory or methodology. To suggest this would contradict the very basis of discursive psychology, which recognizes the importance of rhetoric in social life. By calling the issue a rhetorical one, I hope to draw attention to the importance of the way that psychologists, including discursive psychologists, use language to write about the phenomena that they are studying.

1. Different flavours of Discursive Psychology

3Inevitably, this discussion will simplify some of the complexities of discursive psychology. Indeed, the very term ‘discursive psychology’ is itself a simplification. It conveys the idea of a defined theoretical position and a united intellectual movement. However, there are different types of discursive psychology. A recent book, which surveys the discursive approach to psychology, comments that “readers should be aware that the different flavours of discursive research are a bit like the different flavours of ice cream” (McKinlay and McVittie, 2008, p. 9). Although there is not space here to present all the different flavours, some brief remarks are necessary.

4Differences of approach were apparent in two early works that did much to establish the term ‘discursive psychology’ – Edwards and Potter’s Discursive Psychology (1993) and Harré and Gillett’s The Discursive Mind (1994). Neither book was the first in which the respective authors had developed their basic perspectives (see, for example, Edwards and Mercer, 1987; Harré and Secord, 1973; Harré, 1983; Potter and Wetherell, 1987). Although both Discursive Psychology and The Discursive Mind were criticising orthodox psychology from an explicitly discursive perspective, using overlapping critical arguments, their positive proposals for reconstituting the discipline of psychology were somewhat different. Edwards and Potter were advocating a methodological approach that is based on the study of conversational interaction, particularly using the techniques of conversation analysis developed by Harvey Sacks and Emanual Schegloff (e.g., Sacks, 1992; Sacks and Schegloff, 1979; Sacks, Schegloff and Jefferson, 1974). The Discursive Mind continued Harré’s ‘ethogenic’ approach, which is much more methodologically eclectic than is conversation analysis. Ethogenics involves, inter alia, applying the principles and techniques of animal ethology to the study of humans. Significantly, The Discursive Mind did not discuss conversation analysis; nor did it refer to the work of Sacks and Schegloff. On the other hand, Discursive Psychology referred to Harré, but he was principally mentioned as one of a number of critics of mainstream experimentalism. Discursive Psychology did not specifically discuss Harré’s ethogenics or the empirical work it has produced.

5The broad term ‘discursive psychology’ covers other ‘flavours’ of work, which draw on a variety of different approaches to the study of language. For example, within the broad field of discursive psychology, there is a rhetorical ‘flavour’ that self-consciously builds on the work of classical rhetoric, in order to stress the argumentative and persuasive nature of language (e.g., Billig, 1987 and 1991). The rhetorical, conversation analytic and ethogenic approaches tend to examine what language users are doing in particular discursive contexts. As such, they tend not to use the concept of ‘discourse’ in a Foucauldian sense, to indicate a system of discursive practices that is found within a particular socio-historical context. However, some discursive psychologists have specifically adopted a Foucauldian perspective (e.g., Parker, 1992; Parker, 2002; Burman et al, 1996). There is an additional factor. As discursive psychology has developed, so some of its early proponents have developed their perspectives with the consequence that their later work can take on different favours than their earlier work (see Wiggins and Potter, 2008).

6The metaphor comparing different types of discursive psychology with different flavours of ice cream may be slightly misleading. It suggests that we may easily combine flavours according to taste, choosing for example a scoop of vanilla with a scoop of strawberry. However, some of the proponents of the various forms do not see their approaches as being easily compatible. The Foucauldians have criticised those who concentrate on interpersonal interaction, while, in their turn, those concentrating on interaction have claimed that the Foucauldians tend to use an abstract, idealised conception of discourse. Within the field of discursive psychology, there is at present a vigorous debate about the extent to which the study of interaction can be based on techniques of conversation analysis or whether it should be augmented by other critical concepts from the social sciences (see, for example, Billig, 1999b; Corcoran, in press; Edley, 2001; Kitzinger, 2006; Wetherell, 1998; Wetherell, 2007).

7There is no need to go into the details of these debates which may be more meaningful to the participants than to those unfamiliar with the general field. Outsiders may well think that the various forms of discursive psychology share a great deal in common and that the internal debates represent what Freud many years ago termed ‘the narcissism of minor differences’ (1921/1985). Here, the strategy will be to emphasise the commonalities rather than the differences. Basically, the various proponents of discursive psychology share a conception of language which is materialist, in the sense that it is rooted in the practicalities of what people say and do. This is true of discursive psychologists, whether they take their inspiration from conversation analysis, the later philosophy of Wittgenstein, Bakhtin, Volosinov, ancient rhetoric, Foucault or other post-structuralist theories of language. This means that linguistics per se has contributed far less than more pragmatic approaches. In fact, many discursive psychologists have deliberately avoiding using transformational grammars. They have often viewed formal linguistics as providing idealized accounts of language rather than observing what people actually say and do (see for instance, Billig, in press b). In terms of linguists’ well-known distinction between la parole and la langue, the sympathies of discursive psychologists have been very much with la parole.

8This pragmatic approach to language is central to discursive psychology’s confrontation with cognitive psychology. The various flavours of discursive psychology can agree upon this. They all oppose the assumptions of cognitivism (see, for example, Harré, 2002, for a discussion of ‘cognitivism’). Discursive psychologists feel that structural linguistics and cognitivism – not to mention the current fashion for cognitive linguistics – are both based on idealized presuppositions, which conflict with a more materialistic, pragmatic study of the ways that people use language.

2. Anti-Cognitivism

9So what is the cognitivism that discursive psychologists reject? Edwards (1997) describes cognitivism as “a perspective that reduces all of psychological life, including discourse and social interaction, to the workings of cognitive, or even computational, mental processes” (p. 19). Within mainstream psychology, the dominant movement in the past twenty years has without doubt been cognitive psychology. The so-called ‘cognitive revolution’ started as a reaction against behaviourism, and it shifted the emphasis from the study of behaviour to the examination of the internal structures of mind. Virtually no area of psychology has remained unaffected by this shift of emphasis. Obviously, there are numerous different theories and approaches within the broad area of cognitive psychology, let alone cognitive science in general. However, there are common general assumptions. According to Branquinho (2001), one of the pervasive ideas, which cognitive scientists share, is that brains/minds are processors of information (p. xii). The brain/mind receives information from the external world and it must process such information. It is the goal of cognitive scientists to understand how this processing takes place and what mechanisms are used. A second assumption is that the brain/mind must have representations of the external world. Cognitive psychologists assume that without having mental representations of objects in the external world, an organism would not be able to recognize patterns of incoming stimuli as exemplars of such objects. For example, we would not be able to recognize a particular person’s face unless we had a stored mental representation of their face – as well as a stored representation of faces in general. For facial recognition to occur, an incoming pattern of stimuli has to be categorized as a face in general, and as a particular face.

10Cognitive scientists have expended great efforts into building models of information processing and mental representation. Donald Rumelhart’s work on schemata as mental structures for categorizing the information has been extremely influential (e.g. Rumelhart, 1980). Cognitive social psychologists have taken up the notion of ‘schema’ to re-examine traditional topics of social psychology, such as attitudes, prejudice and stereotyping. They have argued that stereotypes and attitudes are mental schemata which guide the social perceiver. Attitudes are schemata about a particular social stimulus, and stereotypes refer to the characteristics which the perceiver associates with particular social groups. Prejudice arises when the social perceiver interprets the behaviour and characteristics of individual group members in terms of negative schemata held about the social group as a whole (see, for instance, Fiske and Taylor, 2008; Leyens, Yzerbyt and Schadron, 1994; McGarty, 1999).

11In such formulations, the key term of schema refers to a supposed cognitive structure within the mind of the social perceiver. Accordingly, language-based phenomena are understood in terms of inner cognitive processes. An attitude might studied by paying attention to what someone says about a social topic, but in the cognitive approach those words are presumed to represent the inner, cognitive structure that supposedly has generated the expressed beliefs. Similarly, a stereotype has to be expressed in language: it refers to a judgement that all/most/typical members of group X possess characteristic Y. Cognitive social psychologists tend to view stereotypes as the product of a cognitive schema that associates particular traits with a particular social category.

12More generally, the cognitive approach tends to assume that language categories represent cognitive processes. This is true, for example, of the important work of Eleanor Rosch (1975). The fact that a speaker uses a particular category in a particular way is taken as evidence that they have internally processed and categorized stimuli in that way. A related cognitive assumption appears in Lakoff and Johnson’s path breaking studies of metaphor in Metaphors We Live By (1980). A metaphor, according to Lakoff and Johnson, involves understanding one concept in terms of another – for instance, understanding debate as warfare. They argue that there is a psychological component in the use of metaphors. In particular, there is an ‘experiential basis’ behind metaphors. This seems to suggest that the use of a particular metaphor involves a particular psychological state. Thus, each person who uses a given metaphor – such as the metaphor of understanding debates in terms of the language of warfare –shares a common experience based on a common cognitive structure. Critics have argued that this assumption is both implausible and unnecessary for understanding how speakers/writers actually use the language of metaphor (Billig, 2006; Glucksberg, 2001). There is no need to assume that the use of a particular metaphor indicates the presence of a particular psychological state or experience.

13Similarly, it does not make sense to suppose that we can simply read off an individual’s inner cognitive state by the categories they use in their speech, as if the categories have been produced by, and thereby represent, the inner state of mind. Nevertheless, the assumption of much cognitive psychology is that the real business of thinking is taking place in the mind of the individual. We can observe how speakers use categories in interaction, but for the cognitive scientist this outward behaviour is merely a sign of the underlying cognitive process of schematising, representing, categorizing etc. All this takes place out of sight within the mind/ brain of the language user.

14The consequence of this cognitive approach is paradoxical, as Harré (2002) has pointed out. Cognitive scientists claim to be scientific, with cognitive psychologists making much of their commitment to using the empirical procedures of experimentation. However, the objects, which these scientists study, are intrinsically unobservable. No high powered microscope or techniques of neural imaging will be able to identify a mental representation or a cognitive categorization. These are not material objects that, in principle, could be viewed if we had the right equipment. They are intrinsically unobservable entities, whose existence can only be inferred from outward actions, especially language-based actions. Thus, the science of cognition is based on examining entities that cannot be directly examined.

15Discursive psychologists, whatever their particular flavour, have reacted against the assumptions of cognitive psychology and, indeed, cognitive linguistics. It is not that they have sought to construct different models of cognitive processing, but they have argued against the very idea of treating language-behaviour as if it is a sign for the real, underlying psychological entities.

16One problem with cognitivism is that in seeking to find the supposed mental entities that stand behind social behaviour, cognitive scientists have neglected to study the intricacies of such behaviour. This charge has been important in the discursive reaction against cognitive approaches to the study of language. Here the methodological and the theoretical criticisms of cognitivism come together. According to discursive psychologists, not only do cognitivists have an inadequate view of language but their perspective distracts them from examining how people actually use language within social life.  This argument depends upon a theory of language that is different from that which is presupposed by the majority of work in cognitive psychology and experimental social psychology.

3. Wittgenstein and the pragmatic approach to language

17Philosophically many of the principles of discursive psychology can be derived from Wittgenstein’s later writings on language, although these ideas have a much longer history (Billig, 2008). In Philosophical Investigations Wittgenstein (1967) argued against a view of language that has subsequently been called the ‘telementation’ theory (Harris, 1997 and 2002). This theory, which for centuries remained the dominant philosophical approach to language, assumed that language exists to communicate internal thoughts from one mind to another. The telementation theory assumes that the categories of language represent inner mental states or ideas and, in this way, language permits the communication of these states from one person to another. Thus, if we conceive of ‘attitudes’, ‘beliefs’ or ‘thoughts’ as cognitive schemata, then the language used by a person to describe their attitudes, beliefs or thoughts provides a representation of these unobservable entities. Wittgenstein produced three streams of argument to counter this view of language, which he believed was profoundly mistaken and which he thought was leading both philosophers and psychologists into serious conceptual errors.

18First and most famously, Wittgenstein argued against the possibility of a private language. Language, he argued, is a public, social activity with a shared grammar. If we use words meaningfully, then there must be some public agreement about their meaning: there must be agreed public criteria for their use. That being so, our words cannot merely refer to internal sensations, perceptions or cognitions, as, for example, John Locke and his later followers, implied. Our language must be rooted in the external public world.

19If language is not constructed out of the need to communicate internal psychological states from one mind to another, then it must have a different basis. Wittgenstein’s second stream of argument was to insist on the pragmatic aspect of language. We do things with language: it is part of our forms of life. In this respect, words are always more than words, for, as Wittgenstein wrote, “words are deeds” (1980, p. 46).  If we want to understand the meaning of language, we should not be searching for inner cognitive structures to match against our vocabulary, but we should be examining what people are actually doing with language when they speak to each other.

20The first two arguments – against the possibility of a private language and for a pragmatic approach to meaning – have direct implications for the study of psychological topics. Again and again in Philosophical Investigations, Wittgenstein argued against the temptation to assume that when we refer to our ‘beliefs’, ‘hopes’, ‘feelings’, ‘attitudes’, we are referring to internal states. The grammar of such terms, he suggests, is misleading. Wittgenstein urged that we should look carefully at the use of such psychological words. For example, we might use ‘believe’ to perform a particular sort of act. When we say ‘I believe it is going to rain tomorrow’, the use of ‘believe’ does not indicate that a belief-structure has been cognitively activated, but that we are indicating some level of doubt that it might rain tomorrow.

21Wittgenstein was suggesting that if we observe what speakers are doing when they use psychological words, we will find that they are engaging in complex social activity. These notions were taken up by the philosopher John Austin in his book How to Do Things with Words and in his careful analysis of what speakers are doing when they use the verb ‘to promise’ (Austin, 1962 and 1979). In some circumstances, the utterance of the phrase ‘I promise to do X’ constitutes an action by which a promise, or social agreement, is made. As such, making the utterance accomplishes the action. This was the basis of what became known as ‘speech act theory’.

4. The Project of Discursive Psychology

22The ideas of Wittgenstein and Austin might have been very important in changing the dominant philosophical conception of language, but they lacked one essential ingredient: the empirical observation of talk in real life. Although Wittgenstein argued that philosophers should examine how people actually use language, neither he nor Austin was not engaged in an empirical project studying the use of language. Austin imagined hypothetical scenarios to illustrate what the utterance ‘I promise’ might mean under different circumstances, rather than studying real-life conversations. With the development of light-weight recording devices, it became possible to move from imagining hypothetical utterances to studying actual conversations. In the 1970s Sacks and Schegloff began their project of examining the organization of conversational interaction. Significantly, Schegloff in one of his earlier studies examined the use of ‘promising’. He found that ‘I promise’ often was not used to make promises in the way that Austin envisaged, but to close conversations, as participants ‘promised’ to keep in touch etc (Schegloff, 1985).

23Discursive psychologists have taken up the materialist and social view of language, proposed by Wittgenstein and Austin, combined with the empirical commitment of conversation analysis (Potter, 2001). One distinguishing feature of discursive psychology has been its project to examine how ordinary speakers use psychological words. In so doing, they have transformed the traditional questions of psychology. For example, instead of asking how does the mind remember, they have asked what are people doing when they claim to remember things (Edwards, 1997; Middleton and Edwards, 1990; Middleton and Brown, 2005).

24With respect to the study of attitudes, discursive psychologists do not assume that ‘attitudes’ really exist as cognitive schema inside the mind of individuals. This was a point made in two early books (Billig, 1987; Potter and Wetherell, 1987). Instead of ‘attitude’ being a technical concept, which is used by the analyst and which is typically studied by means of questionnaires, ‘attitude’ became a topic: discursive psychologists examined what people were doing when they claimed to be giving their opinions or attitudes in conversational discussions.  The traditional approach predicts that people will say more or less the same things when they respond to attitudinal stimuli, for their words will be produced by cognitive structures which remain more or less fixed. The discursive and rhetorical perspectives suggest that there will be great variability in what people say, because the same person may be doing different things with their talk when they speak on different occasions and in different interactional contexts. In fact, the same speaker will rarely, if ever, speak in exactly the same way on separate occasions.

25Billig (1987 and 1991) suggested that an ‘attitude’ should not be seen as an internal schema but as a stance in a controversy. When people give their opinion or say what their attitude is, they are positioning themselves on an issue where they know there is debate and difference. In this way, an opinion takes its rhetorical meaning from its counter-opinion. If giving an opinion is an entry into a matter of argument, then it must be understood in terms the opinions it is opposing. For instance, one study examined an English man who said that he had strong views on the issue of monarchy (Billig, 1991). In a discussion with his family members, this man did not endlessly repeat the same things, as if having ‘a strong view’ meant having one view that lacked variety. Instead, he adjusted his words to those positions that he was countering and to those that he was allying himself with, shifting at times from a radical to a more conservative rhetoric. In so doing, he demonstrated that the person ‘with strong views’ is frequently the person who has much to say on a given topic and much to argue about.

26Traditionally attitude researchers have assumed that their questions, particularly those delivered in the form of a questionnaire, can access respondents’ attitudes, which comprise a relatively stable cognitive structure. Discursive psychologists stress that the questions, posed by the researcher, including those within a questionnaire, are always part of an interaction. By examining in detail the nature of this interaction, one can observe how the replies of respondents relate to the ways that questions are asked (Antaki and Rapley, 1996). This is not surprising for the meaning of what speakers say in conversational interaction depends upon the sequencing of that interaction. To understand what a speaker is saying, the analyst should not abstract their utterances from the interaction, as if that utterance represents an unchanging schematic view. Utterances take their meaning from what comes before and afterwards, especially within an argumentative context.

27Accordingly, discursive analysts examine how people give their opinions in the context of the conversational turn-taking (Puchta and Potter, 2002). A speaker may downgrade their opinion, saying words to the effect ‘that’s just my opinion’, thereby constructing the issue as one where different opinions can be expected. Moreover, ‘that’s just my opinion’ can have the effect of protecting the view from challenge, for the speaker is emphasizing their entitlement to have a view different from others (Billig, 1991). Wiggins and Potter (2003), studying how opinions about food are given during meal-times, suggest that evaluations sometimes can function as justifications for actions, rather than being the determinant of those actions as traditional attitude theory suggests. Moreover, the very terms that are used to depict an issue can be rhetorically significant. A speaker may claim not to have an opinion, but they might use strongly evaluative rhetoric as they supposedly describe the ‘facts’ (Wiggins and Potter, 2008). Speakers may present themselves as unbiased, as they depict counter-positions in extreme terms and thereby as unreasonable, (Billig, 1991; Pomerantz, 1986; Potter, 1996). As such the rhetoric of opinion-giving can be subtle, complex and changing as the rhetorical context of argumentation changes.

28All this indicates a theoretical and methodological shift from the internal to the external. The discursive analyst does not search for unobservable inner schemata that determine outward utterances, but examines the details of the utterances in their interactional and rhetorical contexts. This is not to deny that internal states and inner experiences exist. It is just that these are methodologically always just out of reach. Typically when psychologists claim to be studying inner experiences they are, in fact, examining further discourse that, properly speaking, should be analysed discursively (Edwards, 1997). As Henri Bergson (1946) argued, the conventional categories of psychologists are unsuitable vehicles for describing the fleeting, fragmentary and deeply personal qualities of inner experience: the skills of novelists or poets are better equipped for such a task.

29Yet, this does not mean that thinking is so deeply hidden that it cannot be studied directly. Thinking need not be considered just as a solitary, hidden process. Within conversations thinking occurs socially. The turns of conversation are too quick for spoken utterances merely to be the outer form of a prior cognitive act. By listening to argumentative dialogue, we can actually hear and directly observe the processes of thinking, as speakers deal with the moves of others and express their own opinions in new ways within the changing rhetorical circumstances of turn and counter-turn (Billig, 1987). Internal thinking is a form of internalized dialogue, as a person debates with themselves (for a history of this idea, see Billig, 2008). This means, that developmentally we learn how to think by participating in conversation and then setting up our own internal conversations. This entails a transformation of the old view of language as a means of representing internal cognitions. The rhetorical or discursive model proposes the opposite, with conversation, not cognition, as the primary process. In short, we think because we can talk, rather than we talk because we think.

5. Reinterpreting the Freudian Unconscious

30In moving from the inner to the outer and refusing to speculate upon consciousness, discursive psychology could be criticized for producing a flattened picture of the person. There appears no depth of character or motive, just moves in conversational interaction. Nevertheless, more is possible. More is going on within conversational interaction than the speakers may be consciously aware of at the moment of their participation. Some things move too fast and too habitually to be consciously noticed or remarked upon. This leads to the possibility that the discursive approach may help us to understand some of the psychological issues of unconsciousness.

31There is an obvious problem with raising the notion of unconsciousness while advocating a sceptical position towards ‘consciousness’. The Freudian unconscious is typically conceived as a mental entity that is even more ghostly and hidden than the mental entities which cognitive psychologists seek to describe. When psychoanalytic theorists link the unconscious to the study of discourse, the results are often vague in the extreme. The work of Lacan illustrates this. Not only is his own language notoriously obscure but he employs an imprecise, unsupported and often misleading psychology (Billig, 2005a). Nevertheless, there are other ways of reinterpreting Freudian ideas to make them compatible with a critical and detailed analysis of language. The key lies in stressing that talk may be the means by which repression is accomplished and thereby the means of creating what Freud called ‘the unconscious’ (Billig, 1999a).

32Freud distinguished between two forms of unconsciousness: the preconscious and the unconscious proper. Preconscious thoughts are those, which can be consciously entertained but which at a particular moment happen not to be. When we are concentrating on one thing, we are not attending to other matters that might become the focus of attention at another moment. In conversation there is always too much occurring for participants to attend to everything. Linguistically, any utterance contains extra meanings that users do not specifically respond to. Speakers, in seeking to direct their auditors’ attention to a particular topic, must use words to point to that topic. If these pointing words themselves become the objects of focus, then the speakers will not have achieved their purposes. In pointing to the topic they will be pointing away from themselves, but if, in return, their words of pointing become the contested topic, they will have failed to direct the conversational attention where they had sought.

33On a social level, it is possible to identity features of ideology that resemble the Freudian preconscious. Many aspects of what has been called ‘banal nationalism’ operate in this way (Billig, 1995; Law, 2001; Higgins, 2004; McNeill, and Tewdwr-Jones, 2003). Reminders of nationhood fill the lives of nationals living in established nation-states, but these reminders frequently constitute the ground, not the figures, of awareness. They are the unwaved flags, hanging outside public building or on filling-station forecourts. Nationhood is reproduced by the deictic use of small words – such as ‘we’, ‘here’ and even ‘the’ – that are daily used in the media and are, as such, unnoticed for they are not the discursive focus of attention. In this way, small deictic words can function to create the nation-state as the ‘natural’ place in which ordinary life is enacted (Billig, 1995).

34To use Freudian terminology, these sorts of words function preconsciously. There is no force to stop them becoming the discursive objects of focus should the occasion arise. The taken-for-granted, but unspecified, ‘we’, that underwrites so many daily utterances in the mass media, can become an elaborated ‘we’. The unwaved flags, as it were, can become deliberately waved flags in moments of self-conscious nationalist activity. In fact, the unnoticed reminders of nationhood make possible the moments of flag-waving, self-conscious nationalism.

35Freud distinguished the ‘preconscious’ from the ‘unconscious’ per se, on the basis that the unconscious per se comprises thoughts or desires that have been repressed from conscious awareness. In this respect, the activity of repression creates what Freud loosely called ‘the unconscious’. In Freud’s terms, these are secrets that we want to keep from ourselves. The unnoticed flags are not, in this sense, repressed secrets – they are merely unnoticed most, but not all, of the time. Despite the volume and detail of Freud’s writings, there is a significant gap right at its theoretical core (Billig, 1999a). He never specified how by the individual learns to repress – what skills are involved and where these skills come from. Moreover, Freud did not specify how an individual might then habitually use such skills in order to ensure that repressed thoughts stay repressed. Much of Freud’s analysis was preoccupied with detailing how repressed thoughts sneak back into conscious awareness, rather than how they are pushed from awareness in the first place.

36It is possible to use some of the precepts of the discursive approach in order to try to rectify this omission in Freud’s theorizing. Accordingly, it has been suggested that the skills of repression are integrally related to the skills of dialogue (Billig, 1999a). It is not possible to elaborate the argument here, but the outlines can be sketched. Language is not just expressive but it is repressive. Every language has codes that must be followed. As conversation analysts have stressed, there could be no conversational interaction without shared codes for turn-taking. Proficient speakers cannot just say what they want and do so when they want. Conversation involves participation in normative social activity. There are complex codes for pitch, tone, choice of words, and timing of utterances. Young children have to learn the disciplines of speaking within interaction. They must learn that certain ways of speaking are appropriate and others are inappropriate. Moreover, there are moral values attached to inappropriate speaking. It is considered ethically wrong to interrupt at the wrong time, to ignore questions, withhold greetings etc and etc. The codes may vary from culture to culture, but no language speaking community can be devoid of such codes of politeness and impoliteness.

37There is no inborn gene that ensures that the child automatically knows how to speak appropriately. The codes of appropriateness and inappropriateness must be learnt. In becoming a competent speaker, the child will hear from adults or older children words to the effect: ‘Do not say that – it is rude’. The paradox is that the child, in learning to speak appropriately or politely, must also learn what is rude and forbidden. Freud argued that we desire to do what is forbidden, because it is forbidden. What is taboo can become an object of desire. This means that in teaching children to speak appropriately,  adults are unwittingly creating the pleasures of rudeness as forbidden objects of desire. Children can take pleasure in saying the very words that they have been told not to say; the pleasure comes from breaking the prohibition.

38However, to be a competent speaker, the temptation of rudeness must be habitually repressed. The temptations to rudeness, or to discursive selfishness, must not only be routinely resisted, they must be repressed or pushed from awareness. We must learn how to produce polite utterances spontaneously without a second thought. One might ask what evidence there is that rudeness is an object of temptation. The best evidence comes from humour. There is no culture without humour, just as there can be no culture without codes of behaviour that restrict its members from doing as they will. As Freud (1905/1991) noted in his famous book on jokes, much of the pleasure of humour comes, not from the technical aspects of jokes, but from evading social taboos. That is why people will laugh the most at jokes on topics hedged with taboos, such as sex, violence and lavatories. Rudeness, too, is regularly a matter of laughter. People may be outraged by social infractions, but they can also laugh at, and take pleasure in, episodes when someone says something inappropriate. The pleasure that can be derived from observing a social infraction is sign of  a desire to break the social codes of politeness that restrain social actors. In the case of speech, this would be evidence of the enduring, but habitually repressed, desire for the pleasures of rudeness (Billig, 2005b; see chapter nine for a re-analysis of an incident in Freud’s case of Little Hans, where Hans’s parents laugh at Hans’s inappropriate behaviour in a public place, indicating simultaneously pleasure and a wish to censure).

39If language creates the necessity for repression, then it also provides the means for repressing. In acquiring the skills of conversation, the child witnesses adults changing topics of conversation, using rhetorical skills to do so. If inner thought is a dialogic process, modelled upon outer conversation, then the skills of changing topics permit the possibility of routine internal repression. As the topic of internal dialogue drifts one way, so the thinker might re-direct the topic in another direction. If this becomes a habitual response, then the thinker will have learnt how to mis-direct themselves without conscious awareness. In this way, outward dialogic activity, not some unseen and thereby unteachable cognitive process, provides the basis of repression, for repressing is essentially a discursive skill. This means that if the child is to learn how to repress then he or she must primarily have witnessed repression as an observable social activity.

40The consequence is that repression can be studied as an outward, ideological and discursive activity that is habitually practised. The well-known defence mechanisms of psychoanalytic theory are patterns of discourse. All, according to Freudian theory, involve an element of denial; denial characteristically is accomplished by acts of language, as the speaker denies something about themselves. Projection is a type of explanation by which the self blames, or criticizes, others for things for which the self should properly be blamed, or criticized. Patterns of projection can be ideologically shared. Such a pattern, for instance, appears in the so-called ‘third person effect’, whereby people claim to be able to resist the persuasive messages of the media by ‘reading between the lines’, while claiming that others, including friends, are easily persuaded (e.g., Billig, 1992; Duck and Mullin, 1995; Hoorens and Ruiter, 1996). This common pattern of attribution constitutes a defensive solution to a major ideological dilemma of mass society: people are dependent for their information about the social world upon media but they know from the media themselves that the media are not to be trusted (for ideological dilemmas, see Billig et al, 1988). The ‘solution’ is not merely defensive but is part of an ideological view that constantly identifies others as ‘inferior’, as well as creating the conditions for continual personal self-doubt, insecurity and denial.

41By such a discursive reinterpretation of Freud it is possible to resolve the seeming contradiction between translating internal cognitive processes into outward discursive activity, while still retaining a critical view that accords importance to unconscious factors. The solution involves recasting ‘the unconscious’. Instead of being seen as a hidden mental entity, it is seen as an activity – that of repressing – that itself is at root a discursive activity. As such, the habitual business of repressing is part of the activity of language and thereby is rooted within the social activity of life.

6. Issue of agency

42Lastly, there is the issue of agency. The problem can be easily stated: Does discursive psychology depict social actors as the agents of their own actions or does it attribute such actions to other causal forces? Traditionally, psychology has sought to find the causes of people’s thoughts and behaviours. This has typically involved depicting people in ways that deny agency. It is not the person who is said to be the agent of their actions but the causal variables that the psychologists uncover, whether these variables are biological, personality or cognitive variables. For example, cognitive psychologists might talk about cognitive systems processing information and feeding it back to the action-centres of the brain. It is as if the hidden cognitive processes are the true agents of activity: behind the scenes these processes are activating and directing things.

43The problem is compounded by the practices of experimentalism. It has been argued that experimental psychologists typically adopt a Newtonian model of action, which assumes that one thing causes another (Markova, in press). Experimental psychologists devise experimental situations to test the effects of identified variables. They allot subjects to an experimental condition that differs from a control condition solely in terms of one particular variable. If the experimenters then find that subjects in the experimental group on average respond differently than subjects in the control group, they will claim that the manipulated variable has caused those differences. The underlying model is that variable X causes action Y.

44The model seems to deny agency to the actor, ascribing it to the variable rather than the person. Habitually, experimentalists use styles of writing that reflect this model. Just to give one example from thousands, two leading social psychologists summarise an experimental investigation of social categorization in the following way: “Arbitrary social categorization produced robust differences between in-group and out-group projection”, supporting the anchoring hypothesis by which “projection is selectively engaged whenever a target group includes the self” (Clement and Kruger, 2002, p. 225). The summary describes what social categorization does – not what people do. But social categorization cannot do anything unless people (both experimental subjects and experimenters) engage in complex social actions and interactions within the experimental situation.

45Discursive psychologists do not conduct experiments as such. In consequence, they face little theoretical temptation to translate action into the cause and effect of variables. Their goal typically is to show the complex patterns of interaction, rather than to search out hidden causal determinants. However, they too are faced with the rhetorical temptation of writing in ways that attribute action to technical entities rather than people. Again, an example can be given. Schegloff (2000) has examined how speakers resolve situations in which they are competing within a conversation to speak at the same time. He demonstrates the subtle moves, signs and reactions. He shows how one speaker may withdraw from the floor, leaving another to complete their turn. However, he does not go further, as a traditional psychologist might, to search for the causal variable– whether it is gender, power, personality etc – which enables one speaker to triumph in the competitive situation and the other to retire defeated. Instead, Schegloff details the phenomenon in painstaking detail.

46Yet, the problem of agency remains. The temptation facing discursive psychologists is not to find hidden causes, but to present features of discourse as if they are the determinants of action. Thus, conversation analysts and discursive psychologists isolate what particular interactional moves do or accomplish, just as Foucauldian analysts describe whole patterns of discourse. In such accounts, discourses and rhetorical devices can be seen to be the agents of action. For example, Schegloff (2000) discusses a phenomenon that he calls “a repeat or partial-repeat” of the trouble source in a situation of overlapping talk. The repeat or partial repeal “helps to pinpoint just what was problematic” and “it invites confirmation” (p. 36). In such phrases, the conversation move (or what Schegloff and other conversation analysts often term a ‘device’) is presented as accomplishing the actions of helping to identify problems and inviting confirmations. It is as if the speakers have given their agency over to the devices that then accomplish the discursive actions. Of course, not all of conversation analysis or discursive psychology comprises this sort of writing. Schegloff, in the article just mentioned, also describes what individual speakers do. However, he tends to do this when starting to analyse particular episodes of talk. As the analysis develops, he moves from the particular to the general – from talking about the actions of speakers to general statements about what devices do.

47Recently, a number of critical psychologists have been troubled by the issues of agency and subjectivity. They have not called for empirical investigations to distinguish between situations where speakers might seem to be the agents of their utterances and those where they are, by contrast, helpless victims of a situation. Rather than examining such matters empirically, such critical psychologists have stressed the need to ‘theorise’ subjectivity and/or agency (e.g., Blackman et al, 2008; Wetherell, 2008). The call to ‘theorise’ is interesting. In this context, theorising does not mean producing testable theories that will predict what will happen in conversational exchanges – i.e. theories that will predict which sort of speaker will emerge triumphant from a competitive conversational overlap. ‘To theorise’ means, in this context, to write in general terms, using a heavily philosophical vocabulary.

48However, the problem of agency will not be resolved in this way. The problem is not to probe further about what the concepts of ‘agency’ or ‘subjectivity’ might mean and how these concepts relate to the works of Derrida, Nietzsche or Lacan. The problem, at least in the form that it confronts discursive psychologists, is neither theoretical nor methodological. It is essentially rhetorical. How should we write discursive psychology? How can we depict people as agents of their own actions? This comes down to a matter of determining to write in ways that emphasise human action.

49The analyses of some linguists, especially critical discourse analysts, are relevant here. Some linguists have made a distinction between processes and entities. Processes are essentially events that extend over time and involve changes, while entities are things. Some linguists have claimed that the most ‘congruent’ ways of describing processes are with phrases using verbs, such as a noun subject plus verb, and entities are most congruently described with nominals or nouns (Halliday, 1985 and 2003; see also Goatly, 2007). Human actions are processes produced by humans. This means that the most congruent way of describing human actions is with an active sentence that uses the agent as subject and the action denoted by a verb: e.g. ‘she went’, ‘he arrived’, ‘they will return’ etc.

50Much scientific writing does not use active verbs to describe processes. Processes are identified and given nominal labels in a style of writing that gives nouns greater emphasis than verbs (Halliday and Martin, 1993). In examples, which have already been cited, analysts use nominal terms that describe processes of human action: ‘social categorization’, ‘a repeat or partial repeat’, ‘in-group-projection’. It is as if social scientists have re-packaged complex actions, which take place over time, as enduring entities. They give technical terms to terms to these entities, and then they write as if these entities can cause other things to occur or to be the agents of other actions (for details of this analysis, see Billig, in press a and in press b). The result can be a reified language, in which speakers/writers rhetorically depict processes-as-things as the agents of activities (see Schafer, 1976, for a discussion of ‘reified’ language in relation to psycho-analytic theory).

51Interestingly, two of the most distinctive features of this sort of ‘scientific’ writing are the two features that critical discourse analysts have identified as being associated with ideological discourse: nominalization or the turning of verbs into nominals; and the use of passives (Fowler et al, 1979; Fowler, 1979; Fairclough, 1992 and 2003). This is no coincidence but illustrates is a problem with ‘scientific’ language, when analysts use it to describe human actions. Generally, it conveys less information than ordinary, non-technical language. Speakers/writers can use passive sentences and nominalizations to refer to processes/actions without specifying the agent of these actions. Headline writers of newspapers can take advantage of these forms for ideological purposes. A newspaper that supports the government might headline an event with a nominal form such as ‘Attack on protestors’ or a passive form ‘Protestors attacked’. Both forms refer to the event without specifying the agents of the event. Use of an active verb would require identifying an agent: ‘Government troops attack protestors’. In this respect, speakers/writers can use nominalizations or passives to avoid describing exactly who did what to whom – and in using a nominalization, a speaker can omit placing the action within time.

52The problem for the social sciences is that there is a rhetorical preference for using technical terms that depict actions as things. The use of technical terms not only distinguishes social scientific writing from journalism or common sense writing but it also conveys a sense of precision.   Nevertheless, this precision may on occasions be illusory. Social psychologists can use a term such as ‘social categorization’ without specifying exactly what sorts of actions are being described – who is doing the categorizing, when it is being done and, most importantly how precisely actors accomplish the activity of categorizing. The very term ‘nominalization’ is the sort of term whose ideological implications critical discourse analysts warn against. And when  critical discourse analysts use ‘nominalization’, they are often very imprecise about what they mean, using it to refer to linguistic entities, unspecified processes and even pseudo-processes without distinguishing between these  very different meanings of ‘nominalization’ (Billig, in press a; in press b). Although analysts might claim technical terms to be more precise than those of ordinary language, analysts, by using technical nominals to describe processes/actions, can actually be more imprecise than if they had used ordinary language. The technical terms may appear impressive and analysts can appear to have solved problems when they have actually avoided them.

53The remedy then is not to devise more theory, with yet more technical terms; nor is it to insist on one set of methodological practices over another. The remedy is simultaneously simpler and harder. It is to write clearly, and to be suspicious of technical jargon, especially when describing human actions. The prime concern of discursive psychologists should be with people, not abstract entities such as ‘discourses’, ‘the organizational structures of conversation’ or ‘social categorization’. That being so, discursive analysts should not see their task to move from the language of action to the technical language of processes-as-entities. They should seek to describe what speakers are doing, when they interact. Ultimately we should not lose sight of the fact that the subject matter of any human psychology – whether discursive, behavioural, cognitive or whatever – is people.

54To produce psychology, we need to write psychology. No academic discipline can exist without written texts, or without rhetorical practices. As far as possible, we should endeavour to ensure that our rhetoric matches our subject matter. And this means trying to write as simply and clearly as possible about what people do. Sadly too much psychological writing continues to treat people as things. After the cognitive and discursive revolutions, we still stand in need of a rhetorical revolution.

Haut de page


Antaki, C. and Rapley, M. (1996). ‘Quality of life’ talk: the liberal paradox of psychological testing. Discourse & Society, 7, 293-316.

Austin, J.L. (1962). How to Do Things with Words. Oxford: Clarendon Press.

Austin, J.L. (1979). Philosophical Papers. Oxford: Clarendon Press.

Bergson, H. (1946). Creative Mind. New York: Citadel Press.

Billig, M. (1987). Arguing and Thinking. Cambridge: Cambridge University Press.

Billig, M. (1991). Ideology and Opinions. London: Sage.

Billig, M. (1992). Talking of the Royal Family. London: Routledge.

Billig, M. (1995). Banal Nationalism. London: Sage.

Billig, M. (1999a). Freudian Repression: conversation creating the unconscious. Cambridge: Cambridge University Press.

Billig, M. (1999b). Whose terms? Whose ordinariness? Rhetoric and ideology in conversation analysis. Discourse and Society, 10, 543-558.

Billig, M. (2006). Lacan’s misuse of psychology: evidence, rhetoric and the mirror stage. Theory, Culture and Society, 23(4), 1-26.

Billig, M. (2005b). Laughter and Ridicule: towards a social critique of humour. London: Sage.

Billig, M. (2006). A psychoanalytic discursive psychology: from consciousness to unconsciousness. Discourse Studies, 8, 17-24.

Billig, M. (2008). The Hidden Roots of Critical Psychology: understanding the impact of Locke, Shaftesbury and Reid. London: Sage.

Billig, M. (in press a). The language of critical discourse analysis: the case of nominalization. Discourse & Society.

Billig, M. (in press b). Nominalizing and de-nominalizing: a reply. Discourse & Society.

Billig, M., Condor, S., Edwards, D., Gane, M., Middleton, D. and Radley, A.R. (1988). Ideological Dilemmas. Sage: London.

Blackman, L. Cromby, J., Hook, D., Papadopoulos, D. and Walkerdine, V. (2008). Creating subjectivities. Subjectivity, 22, 1-27.

Branquinho, J. (2001). Introduction. In J. Branquinho (ed.), The Foundations of Cognitive Science. Oxford: Oxford University Press.

Burman, E. et al. (1996). Psychology, Discourse, Practice. London: Taylor and Francis.

Clement, R.W. and Krueger, J. (2002). Social categorization moderates social projection. Journal of Experimental Social Psychology, 38, 219-231.

Corcoran, T. (in press). Second nature. British Journal of Social Psychology.

Duck, J.M. and Mullin, B.A. (1995). The perceived impact of the mass media: reconsidering the 3rd person effect. European Journal of Social Psychology, 25, 77-93.

Edley, N. (2001). Unravelling social constructionism. Theory and Psychology, 15, 305-324.

Edwards, D. (1997). Discourse and Cognition. London: Sage

Edwards, D. and Mercer, N.M. (1987). Common Knowledge. London: Routledge.

Edwards, D. and Potter, J. (1993). Discursive Psychology. London: Sage.

Fairclough, N. (1992) Discourse and Social Change. Cambridge: Polity Press.

Fairclough, N. (2003) Analyzing Discourse. London: Routledge.

Fiske, S.T. and Taylor, S. E. (2008). Social Cognition. New York: McGraw-Hill.

Fowler, R., Hodge, B., Kress, G. and Trew, T. (1979) Language and Social Control. London: Routledge.

Fowler, R. (1991) Language in the News. London: Routledge.

Freud, S. (1905/1991) Jokes and their Relation to the Unconscious. Harmondsworth: Penguin.

Freud, S. (1921/1985). Group psychology and the analysis of the ego. In S. Freud, Civilization, Society and Religion, Penguin Freud Library, Volume 12. Harmondsworth: Penguin

Gergen, K.J. (2001). Social Construction in Context. London: Sage.

Glucksberg, S. (2001). Understanding Figurative Language: from metaphors to idioms. Oxford: Oxford University Press.

Goatly, A. (2007) Washing the Brain. Amsterdam/Philadelphia: John Benjamins.

Gough, B. and McFadden, M. (2001). Critical Social Psychology. Basingstoke: Palgrave.

Halliday, M.A.K. (1985) An Introduction to Functional Grammar. London: Edward Arnold.

Halliday, M.A.K. (2003) On Language and Linguistics. London: Continuum.

Halliday, M.A.K. and Martin, J.R. (1993) Writing Science. London: Falmer Press.

Harré, R. (1983). Personal Being. Oxford: Blackwell.

Harré, R. (2002). Cognitive Science: a philosophical introduction. London: Sage.

Harré, R. and Gillett, G. (1994). The Discursive Mind. London: Sage.

Harré, R. and Secord, P.F. (1973). The Explanation of Social Behaviour. Oxford: Blackwell.

Harris, R. (2002). The rise of the language myth. In R. Harris (ed.), The Language Myth in Western Culture. Richmond: Curzon Press

Harris, R. (1997). Landmarks in Linguistic Thought. London: Routledge

Hepburn, A. (2003). An Introduction to Critical Social Psychology. London: Sage.

Higgins, M. (2004). Putting the nation in the news. Discourse and Society, 15, 633-648.

Hoorens, V. and Ruiter, S. (1996). The optimal impact effect: beyond the 3rd person effect. European Journal of Social Psychology, 26, 599-610.

Kitzinger, C. (2006). After post-cognitivism. Discourse Studies, 8, 67-83.

Lakoff, G.  and Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press.

Law, A. (2001). Near and far: banal national identity and the press in Scotland. Media, Culture and Society, 23, 299-317.

Leyens, J.-P., Yzerbyt, V.Y.A. and Schadron, G. (1994). Stereoytpes and Social Cognition. London: Sage.

Marková, I. (in press). The epistemological significance of the theory of social representations. Journal for the Theory of Social Behaviour.

McGarty, C. (1999). Categorization in Social Psychology. London: Sage.

McKinlay, A. and McVittie, C. (2008). Social Psychology and Discourse. Chichester: Wiley-Blackwell.

McNeill, D. and Tewdwr-Jones, M. (2003). Architecture, banal nationalism and re-territorialization. International Journal of Urban and Regional Research, 27, 738-743.

Middleton, D. and Brown, S.D. (2005). The Social Psychology of Experience. London: Sage.

Middleton, D. and Edwards, D. (1990). Collective Remembering. London: Sage.
Parker, I. (1992). Discourse Dynamics. London: Routledge.

Parker, I. (2002). Critical Discursive Psychology. London: Palgrave Macmillan.

Pomerantz, A. (1986). Extreme case formulations: a new way of legitimating claims. Human Studies, 9, 219-230.

Potter, J. (1996). Representing Reality. London: Sage.

Potter, J. (2001). Wittgenstein and Austin. In M. Wetherell, S. Taylor and S.J. Yates (eds), Discourse Theory and Practice. London: Sage.

Potter, J. and Wetherell, M. (1987). Discourse and Social Psychology. London: Sage

Puchta, C. and Potter, J. (2002). Manufacturing individual opinions: market research focus groups and the discursive psychology of attitudes. British Journal of Social Psychology, 41, 345-363.

Rosch, E.R. (1975). Cognitive representations of semantic categories. Journal of Experimental Psychology: General, 104, 192-233.

Rumelhart, D. (1980). Schemata, the building blocks of cognition. In R.J. Spiro, B.C. Bruce and W.F. Brewer (eds), Theoretical Issues in Reading Comprehension. Hillsdale, N.J.: Lawrence Erlbaum.

Sacks, H. (1992). Lectures on Conversation. Oxford: Blackwell.

Sacks, H. and Schegloff, E.A. (1979). Two preferences in the organization of reference to persons in conversation and their interaction. In G. Psathas (ed.), Everyday Language. New York: Irvington.

Sacks, H., Schegloff, E.A. and Jefferson, G. (1974). A simplest systematics for the organization of turn-taking for conversation. Language, 50, 696-735.

Schegloff, E.A. (1985). On some questions and ambiguities in conversation. In J.M. Atkinson and J. Heritage (eds), Structures of Social Action. Cambridge: Cambridge University Press.

Schegloff, E.A. (2000). Overlapping talk and the organization of turn-taking for conversation. Language in Society, 29, 1-63.

Sloan, T. (ed.) (2000). Critical Psychology: voices for change. London: Macmillan

Wetherell, M. (1998). Positioning and interpretative repertoires: conversation analysis and post-structuralism in dialogue. Discourse & Society, 9, 387-412.

Wetherell, M. (2007). A step too far: discursive psychology, linguistic identity and questions of identity. Journal of Sociolinguistics, 11, 661-681.

Wetherell, M. (2008). Subjectivity or psycho-discusive practices? Investigating complex intersectional identities. Subjectivity, 22, 73-81.

Wiggins, S. and Potter, J. (2003). Attitudes and evaluative practices: category vs. item and subjective vs. objective constructions in everyday food assessments. British Journal of Social Psychology, 42, 513-531.

Wiggins, S. and Potter, J. (2008). Discursive psychology. In C. Willig and W. Stainton-Rogers (eds), The Sage Handbook of Qualitative Research in Psychology. London: Sage.

Wittgenstein, L. (1967). Philosophical Investigations. Oxford: Blackwell.

Wittgenstein, L. (1980). Culture and Value. Oxford: Blackwell.

Haut de page

Pour citer cet article

Référence électronique

Michael Billig, « Discursive psychology, rhetoric and the issue of agency », Semen [En ligne], 27 | 2009, mis en ligne le 10 décembre 2010, consulté le 21 janvier 2017. URL :

Haut de page


Michael Billig

Loughborough UniversityLoughborough, Leicestershire, LE11 3TU, United Kingdom

Haut de page

Droits d’auteur

© Presses universitaires de Franche-Comté

Haut de page