The Dialogic Unconscious: psycho-analysis, discursive psychology and the nature of repression *

Michael Billig

Dept of Social Sciences Loughborough University Loughborough
LE11 3TU

To appear: Theory, Culture and Society


This paper explores possible links between discursive psychology and psycho-analytic theory. At first sight, the two approaches would seem to be incompatible. Discursive psychology, in keeping with its Wittgensteinian and conversation analytic background, concentrates upon the social and discursive constitution of psychological phenomena, rather than on supposed inner motivations. However, the notion of the Dialogic Unconscious is introduced in order to suggest how processes of repression can be studied discursively. The argument for the Dialogic Unconscious suggests that conversational interaction can have repressive functions, as well as expressive ones. It is suggested that discursive psychology has tended to overlook this repressive dimension, concentrating upon the presences rather than absences in discourse. However, the conversational devices, which conversation analysts have revealed to be vital for politeness and everyday morality, can also be seen to repress the temptation of rudeness. That being so, repression can be observed to be routinely accomplished by discursive interaction. Moreover, the notion of the Dialogic Unconscious not only suggests that dialogue can be repressive, but also that repression is itself a dialogic, or discursive, process. The implications for both discursive psychology and Freudian psycho-analytic theory are discussed.

This paper will discuss the possibility of making links between discursive psychology and psycho-analytic theory. At first sight, these two ways of understanding psychological phenomena seem utterly incompatible. The discursive psychology, which is being developed as a challenge to orthodox social psychology, discourages speculation about 'inner' processes (e.g., Antaki, 1994; Billig, in press a; Edwards, 1991 and 1994; Edwards and Potter, 1992 and 1993; Harré, 1995; Harré and Gillett, 1994; Parker, 1992; Potter, 1996; Potter and Wetherell, 1987 and 1995). Discursive psychology takes inspiration from the philosophical tradition of Wittgenstein's later philosophy and from the development of ethnomethodology and conversation analysis. These traditions of analysis stress the need to examine in detail the outward accomplishment of social life, showing how social order is reproduced through discursive interaction. Discursive psychology applies this project to psychological phenomena. It argues that phenomena, which traditional psychological theories have treated as 'inner processes', are, in fact, constituted through social, discursive activity. Accordingly, discursive psychologists argue that psychology should be based on the study of this outward activity rather than upon hypothetical, and essentially unobservable, inner states. In this respect, discursive psychology seems inimical with psycho-analytic theory, which presumes that hidden, unconscious motive-forces lie behind the surface of social life. Psycho-analytic theorists often treat outward social activity as a cipher for unobservable, inner motivational processes. The apparent incompatibility between the new discursive psychology and psycho-analysis might be expressed crudely: discursive psychologists turn the person inside-out, converting inner mental life into outward social activity, while psycho-analysts move in the opposite theoretical direction by turning social life outside-in.

It will be suggested that matters are not quite so simple, especially if one takes seriously the notion that conversational interaction is not only a mode of expression, but is also a mode of repression. There is a case, it will be argued, for considering that there is a 'Dialogic Unconscious', which is constituted in dialogue and which functions as a form of repression. If the repressive aspect of conversation is considered, then discursive psychology and psycho-analytic theory may not be utterly incompatible, at least if theoretical adjustments are made on both sides.

In developing these themes, the present paper will attempt to argue within the basic assumptions of discursive psychology, especially the Wittgensteinian notion that psychological language stands in need of outward, rather than inner, criteria. It will be suggested, this position does not preclude psycho-analytic forms of explanation, but, under certain circumstances, encourages them. Nevertheless, to develop psycho-analytic themes based on the idea of a Dialogic Unconscious, it is necessary to expand the image of social interaction, which discursive psychology has inherited from conversation analysis and ethnomethodology. At present, conversation analysts have revealed in depth how everyday morality is routinely accomplished in conversational interaction. If the unconscious, as well as the conscious, is constituted in interaction, then the everyday interaction is not only reproducing moral norms, but it also reproduces immoral temptations, which are routinely resisted and repressed. To investigate these issues, so it will be argued, it is necessary to examine the absences, rather than presences, in dialogue.

If this is done, then one should be able to investigate not only how everyday morality is accomplished within talk-in-interaction, as conversation analysts and ethnomethodologists have demonstrated; also it will be possible to see how repression is routinely accomplished. In so doing, discursive psychology will be attending to some of the traditional concerns of psycho-analysis. Most importantly, the basic psycho-analytic process of repression, whose routine operation even Freud considered to be mysterious, might be observed directly as it is constituted through social, discursive interaction.

Before discussing these themes, it is necessary to say a few words about the use of the terms 'psycho-analysis' and 'discourse analysis', for both terms cover a variety of different intellectual positions. In discussing psycho-analysis, the emphasis will be on the writings of Freud, rather than on later schools of thought. Of course, it should be possible to link the notion of a Dialogic Unconscious with concepts derived from other psycho-analytic schools. Indeed, this needs to be done, if the idea of a Dialogic Unconscious is to be fully developed. However, this initial paper concentrates upon Freud's writings. Not only have Freud's writings a unique historic and cultural significance, but arguably they still constitute the clearest presentation of psycho-analysis's central, enduring concepts.

Just as there are a variety of psycho-analytic schools, so very different approaches to the study of language are to be found under the broad heading of 'discourse analysis' (see, for example, the compendia edited by van Dijk, 1985 and 1996). The type of discourse analysis to be explored in this paper is that which has been heavily influenced by conversation analysis, ethnomethodology and the later philosophy of Wittgenstein. This form of analysis concentrates on studying actual utterances in their particular contexts. In this respect, it differs substantially from Foucauldian and Lacanian approaches. Foucauldian studies tend to view discourses as cultural totalities, and they seek to identify these totalities rather than examine the ways in which people use utterances in specific circumstances. Foucault's notion of a 'discourse' is, to use Saussure's distinction, more langue than parole - concerned with the hypoothetical total structure rather than particular usage. Similarly, Lacan, who takes seriously the link between language and the unconscious, operates with a somewhat abstract theory of language, which explicitly gives priority to langue, or the total linguistic structure, rather than to parole (see, for example, Lacan, 1979, pp. 20-1; for more details, see Billig, 1996 and in press c). By contrast, the present approach, in common with much discursive psychology, is firmly rooted in the study of the utterance and the particular, occasioned use of language. The present approach, therefore, differs from those psychologists who have been recently investigating the links between psycho-analysis and discourse in terms of Foucauldian and Lacanian ideas (see, inter alia, Burman and Parker, 1993; Henriques et al, 1984; Hollway, 1989; Parker, 1990 and 1992; Wilkinson and Kitzinger, 1995). See footnote 1

In consequence, the notion of the Dialogic Unconscious, as outlined in the present paper, is neither Lacanian nor Foucauldian. It does not refer to a presumed 'fallen' structure of signfication. Nor does it refer to utterances, which cannot be spoken because the speaker lacks the cultural and discursive resources for composing such an utterance. On the contrary, the Dialogic Unconscious comprises utterances which could well have been spoken, but which remain unspoken. These are utterances which in some sense have been dialogically repressed in particular conversational settings.

Discourse Analysis and the Unconscious

Discursive psychology has set itself the project of investigating how psychological states are constituted through discursive activity. This means paying attention to the use of psychological language in conversation. The point is not to treat psychological language (such as 'I feel happy' or 'my thoughts on this matter are...') as symptoms of inner processes. The analysis of psychological language, for discursive psychologists, is not a second-best strategy, as if the 'real' topics of psychology were internal, and thereby unobservable, phenomena. Quite the contrary, discursive psychologists assume that the traditional topics of psychology refer to phenomena which are outwardly observable. If one wishes to study 'feelings', psychologists should be paying attention to what people are doing when they claim to have feelings.

The principles behind this approach can be found in the later writings of Wittgenstein, who stressed that, because language is socially-shared, there must be public criteria for the use of psychological words. If there were not such criteria, speakers would not be able to speak so readily of feelings, emotions or states of mind (see Shotter, 1993a, 1993b and 1995, for discussions of Wittgenstein and his contribution to discursive psychology). As Wittgenstein stated in Philosophical Investigations, "an 'inner process' stands in need of outward criteria" (1953, remark 580). He claimed that "the characteristic mark of all 'feelings' is that there is expression of them, i.e. facial expression, gestures" (1967, remark 513). The implications are profound. Psychologists should not search for the ghostly, unobservable essences, which are presumed to lie behind the use of feeling-words. They should be examining in detail the ways people make claims about psychological states and what they are doing when they make, or dispute, such psychological claims: they should be asking "what is the purpose of this language, how is it being used?" (Wittgenstein, 1967, remark 716).

The techniques of conversation analysis and ethnomethodology enable discursive psychologists to answer such questions, for they permit analysts to investigate the micro-processes of speech-acts in which psychological language is used. Analysts can study how, for example, claims to have particular emotions or psychological states are discursively accomplished. In this way, emotions and psychological claims are seen to be socially constituted and accomplished (Edwards, 1995a). For example, traditional psychologists assume that 'remembering' is something which takes place within the cognitive system of the isolated individual. By contrast, discursive psychologists treat remembering as a social and collective activity (Middleton and Edwards, 1990; Edwards and Potter, 1992; Potter and Edwards, 1990). They examine in detail the speech-acts involved in making claims about remembering and forgetting, and they ask what such claims are accomplishing socially (e.g. Edwards and Middleton, 1986; Billig and Edwards, 1994). Consequently, remembering, far from being treated as an unobservable, internal process, becomes a directly observable, social activity based upon speech-acts.

Discursive psychology's stance of exploring observable social action, rather than searching for hidden, inner states, is reflected in the analysts' own theoretical language. Their concepts tend to emphasise action and activity, rather than inner process. Following the example of Harvey Sacks, the word 'do' features prominently in discursive analysis (Edwards, 1995b). 'Feelings' are activities which have to be 'done' (or accomplished socially and discursively). Discursive psychologists do not search for the feelings behind observable actions (or doings), but concentrate upon the doings, without which no-one, including analysts, could talk of feelings.

It should be little surprise that an approach, which discourages phenomenological language as a tool of analysis, has little theoretical space for the vocabulary of psycho-analysis. If inner consciousness is not an explanatory concept, then there would appear to be even less reason to hypothesize an inner unconsciousness. The main texts of discursive or rhetorical psychology contain hardly a mention of the 'unconscious', 'repression' or Freud. Occasionally, and somewhat briefly, psycho-analysis is discussed, but typically in ways which discount its significance (see, for example, Gergen, 1994, p. 169 or Shotter, 1993a, pp. 138f; but see also Shotter, 1990, pp. 129f, for an argument that cognitive psychology has forgotten its Freudian heritage).

Discursive psychology, if it turned its attention to psycho-analysis, might easily treat psycho-analytic language as a topic of investigation, rather than as a theoretical resource for analysing other forms of discourse. This was a tactic which Volosinov/Bakhtin recommended years ago in his critique of Freudianism (Volosinov, 1994; see also Daelemans and Maranhao, 1990). Just as discursive analysts presently study how people use psychological language in everyday conversation, so they could examine how the language of unconscious motives is used in both lay and professional discourses. The distinctiion between the lay and the professional is not firm. As Moscovici (1976) has shown, psycho-analytic concepts are no longer confined to professional, psycho-analytic circles, but have been re-presented, and altered, in ordinary language. In this respect psychology, including psycho-analysis, has a wide cultural and political significance in contemporary Western society (Parker, 1994; Rose, 1990).

Discursive psychologists could specifically examine the current use of psycho-analytic concepts, thereby examining in discursive detail the phenomenon identified in broad outline by Moscovici. Such an examination of psycho-analytic discourse would not necessarily demand special methodological or theoretical treatment. Whether analysts were studying the language of consciousness or unconsciousness, the methodological techniques could be the same. The analysts would not themselves be using psycho-analytic concepts as analytic tools. They would see what is being 'done' with talk about 'hidden, unconscious motives'. They would not claim that 'hidden, unconscious motives' stand behind, or cause, this talk. For example, they would study how people talk about 'repression', but they would not assume that the talk itself was based upon repression. In this respect, such analyses would not be examining the repressive aspects of discourse; nor would they be probing the discursive aspects of repression, as understood by psycho-analysts.

Return of the Repressed Repression

Although the unconscious may be largely absent from the topics studied by discursive psychology, it is not to be easily eliminated. As psycho-analysts well understand, the repressed returns to haunt consciousness. Discursive psychologists, not to mention conversation analysts, might exclude the concept of repression from their theoretical vocabulary, but, occasionally, the repressed repression can be detected, lurking on the edges of analysis. Two examples can be briefly mentioned. Wetherell and Potter (1993), in their analysis of racist discourse, present an example of a 'benevolent racist', who disclaims being racist. This person, they suggest, fits Karen Horney's image of the individual resolving a neurotic conflict through fantasies about the ideal self (p. 54). Billig (1992), analysing how English families talk about the British Royal Family, at one point has recourse to the notion of projection (pp. 106-9; see also Billig, in press b). Speakers are discussing whether the heir to the throne would be allowed to marry a non-white. The speakers disclaim their own racism, and they also deny that the Royal Family, with whom they claimed to 'identify', is racist. Billig suggests that the 'racism' is projected onto others: the speakers oppose the idea of a non-white marrying the heir to the throne, but attribute this position to others - to 'the public' in general. The details of the analysis are not what matter for the moment, only the analyst's use of psycho-analytic terminology.

What, one might ask, are the analysts 'doing', when they use psycho-analytic terms? In both examples, analysts are considering the utterances of speakers, who are denying that they are racist, while expressing views which can easily be heard as racist. The analysts, in presenting this material, distance themselves from the respondents' denials of racism (see also van Dijk, 1992 and 1993a). If the respondents disclaim any racist motives, the analyst, in using psycho-analytic terminology, alerts readers not to accept those disclaimers: the possibility of racist motives, unacknowledged by the holder of the motives, is left open. In this respect, as Parker (1992) has suggested, psycho-analytic concepts can be useful for a discursive psychology which has an acknowledged critical, political stance.

The analytic move of using psycho-analytic terminology does not represent an abandonment of the basic tenets of speech-act theory, but is, in some respects, derived from them. The possibility of using psycho-analytic descriptions follows from the assumption that there are outward criteria for the use psychological concepts. As Wittgenstein continually argued, because there are public criteria for the use of psychological language, we do not describe our own psychological states directly, whilst being forced to infer the states of others indirectly and with less certainty. On the contrary, as Wittgenstein wrote, "every day we hear one man saying of another that he is in pain, is sad, is merry, etc. without a trace of doubt" (Wittgenstein, 1980, remark 138). Thus, we can legitimately claim confidence in our descriptions of another's psychological state. We may be confident, even if the other person disagrees with our assessment; in this case, we can argue that the other's description of themself conflicts with the outward criteria by which such descriptions must be made.

Wittgenstein, at one point, commented that "it also happens that one says 'I know that you felt like this then, even if you won't admit it now" (1980, remark 138). In this sort of situation - where one person claims to know the feelings of another better than the person themself - the possibility of psycho-analytic explanations is raised. Such an explanation signals that the other's self-descriptions are unconvincing. Not only is a more convincing account of their motives offered, but also an account why the self-descriptions are unconvincing - why the person is unable to admit something about themself which is apparent to another and which should be, but is not, apparent to themself. In these circumstances, unconscious motives may be cited convincingly. Wittgenstein gives the example of someone who cannot explain why they suddenly climbed on a chair: "Perhaps in such a case" he suggested, "we should say that he acted with unconscious intention" (1988, remark no. 222). In another example, Wittgenstein wrote about someone, attempting to draw the picture of one individual but actually producing a portrait of another, being "guided by forces in his unconscious" (1967, remark no. 262).

There are two points which can be noted about these usages of unconscious explanations. In the first place, such explanations depend upon a network of more 'normal' explanations. The psycho-analytic language-game can never replace in its entirety the language-games of ordinary psychological explanations, for they are to be used when the ordinary games appear to break-down (Brearley, 1991; Cioffi, 1991; Hopkins, 1995). Second, the possibility for making explanations, which cite unconscious motivations, is not prohibited by the assumptions of speech-act theory, on which discursive psychology is based. Indeed, the possibility follows from the Wittgensteinian assumption that emotions are social constructs and that it is not inevitable that the individual actor is privileged in producing more convincing self-descriptions than any outside observer can. Under some circumstances, others may doubt a person's self-ascriptions, or may wonder at the failure to produce any self-descriptions; then they may speculate why the person cannot admit to themself what appears apparent to outsiders. This sort of discursive move is to be expected in a society where psycho-analytic conceptions have passed into common-sense (Moscovici, 1976). As is illustrated by the two examples, taken from the analysis of speakers denying racism, discursive analysts can make the move too.

Conversation and Constraint

An important question needs to be addressed: if the discursive approach does not theoretically prohibit psycho-analytic speculations, at least in the way that it prohibits speculation about inner cognitive processes, then why are psycho-analytic notions so rarely employed? One might answer such a question historically, pointing out how psychology in general has repressed its psycho-analytic heritage (Parker, 1994; Shotter, 1990). In this respect, discursive psychology would merely be continuing a long established intellectual habit. On the other hand, one might attempt to answer more specifically by showing how discursive psychology's routine procedures of theory and methodology combine to inhibit discussion of the unconscious. It will be suggested that discursive psychology has adopted practices from conversation analysis, which tend to draw attention to the presences, rather than the absences, in discursive utterances. In particular, there are two principles of conversation analysis, which inform discursive psychology and which militate against psycho-analytic concerns with absences: (a) analysts should try to conduct their analyses from the participants' perspectives, as revealed in what the participants say; (b) analysts should use what participants say in order to reveal the structural organization of conversation.

a) Conversation analysts, as well as ethnomethodologists, frequently claim that one of the strengths of their approach is that they focus "on how participants themselves produce and interpret each other's actions" (Pomerantz, 1988, pp. 360-1). Conversation analysts do not seek to explain interaction in terms of sociological structures, which lie 'behind the backs of the participants' (Heritage, 1984; Boden, 1994). Instead, they observe how participants make sense of, and account for, the social world. In this respect, conversation analysis, to quote Deirdre Boden, assumes that "human beings are knowledgeable agents in the production (and reproduction) of their lives and their history" (Boden, 1994, p. 13). This assumption of knowledgeability forms the core of the ethnomethodological enterprise: "By giving back to social agents their knowlegeability of their own social actions, it was then possible to sit back and observe the structuring quality of the world as it happens" (Boden, 1994, p. 74). In this way, ethnomethodologists, to quote Boden again, consider social order as "a member's order than an analyst's construct" (p. 65).

The assumption of actors' knowledgeability is both a methodological facilitation and a restriction. On the one hand, analysts are bidden to take seriously what people actually say and do, rather than assume actors are simply re-enacting given roles within a hypothetical social structure. There is no doubt that this principle has productively concentrated attention on the fine detail of social life. On the other hand, the assumption is theoretically restricting in that the analyst is discouraged from taking a stance, which observers of others often do take and which is central to the whole project of psycho-analysis. This is the stance, which claims that 'members', or social actors, sometimes are unaware of the reasons for their actions and that their accounts, far from indicating knowledgeability, express a lack of knowledge and, indeed, repress self-knowledge. By assuming knowledgeability at the outset, the analyst cannot easily investigate how the repression of knowlegeability might be accomplished.

b) The second principle aims to find within the utterances of social actors the structural organization of everyday life (Boden, 1994; Heritage, 1984). According to Buttny, much social interaction, such as adjacency pairings in conversation, has a "normative organization" (1993, p. 39). Conversation analysts seek to discover this normative organization, which is presumed to lie within, and not behind, what participants say. This is why turn-taking and sequential organization have been so central to conversation analysis. According to Boden (1994), "turn-taking and the sequenced structuring of action" lie "at the heart" of social interaction (p. 53). The organizational pattern is presumed to be discoverable through understanding the sequential constraints on speakers. Drew (1995) claims that 'turns' in conversation are treated "as the product of the sequential organization of talk" (p. 70). The organization is presumed to be present in the accounts which speakers themselves give. Thus, the analyst seeks to note conversational devices, which "are demonstrably or observably relevant to the participants themselves" (Drew, 1995, p. 76; emphasis in original). By so doing, analysts hope to discover the "stable and organized properties of conversational structures" (p. 76).

These structures are so normatively powerful, that they shape to a large extent what is said in conversational interaction. Drew (1995) claims that, according to the

perspective of conversation analysis, participants' inner cognitive states are autonomous from their talk. What people say "is not determined by, or the automatic product of, the processes of the mind"; instead, utterances in conversation "are shaped most proximately by the sequential position in which the turn is produced" (Drew, 1995, p. 70). It is as if speakers find themselves inhabiting a normative structure which is more powerful than their individual feelings and to which they have to conform for interaction to proceed. This structure is present within the interaction and the analyst, therefore, seeks to examine "connectedness among participants' actions to uncover the sequential ordering of accounting" (p. 59, emphasis in original). When analysts have located an instance of accounting, they should search for the conversational antecedents and consequences of these accounts. Most importantly, they should "examine what connects these parts together as a coherent discursive practice of accounting" (p. 60, emphasis in original).

The upshot is that analysts tend not to search for absences in conversation, for the given conversation is to be seen as coherent in itself. It is presumed that participants are knowledgeable about the conversational structure, which is determining their "conduct in talk-in-interaction" (Drew, 1995, p. 78). Similarly, it is presumed that this knowledge is expressed in the utterances themselves, and, thus, the structure is knowable through these expressions. In consequence, the determination of particular utterances is to be understood in terms of other utterances. Again, the emphasis is upon the uttered presences of conversation. Analysts work with the text (or transcription) of the interaction, rather than using the text as a sign of something else, which is itself absent in the text, but which leaves traces. The meaning of a particular conversational turn is not to be sought in the inner states of mind of the speaker; nor is it to be sought in a hypothetical 'social structure'; it is to be found in the relations with the preceding interventions, and checked by the subsequent ones. What was, and what will be spoken, provide the means for understanding the function and meaning of what is said. In this respect, the analysts put themselves in the same position as the participants. What this means is that analysts search for the connections between utterances which are present in the conversation. What is absent from the conversation tends to be absent from the analysis.

Conversational Morality and Its Flaws

By following the theoretical and methodological presuppositions of conversation analysis, researchers have provided an impressive corpus of knowledge about the intricate structures of contemporary conversations. They have demonstrated in detail the shared practices by which participants manage everyday interactions and, thereby, reproduce the normative structures which permit conversation to progress. For example, conversation analysts have claimed that there is a 'preference for agreement', which facilitates expressions of agreement and mitigates against expressions of outright disagreement (Pomerantz, 1984; Bilmes, 1987). There are ways of 'doing politeness', which hedge requests in indirectness, so that they are not heard as conversation-threatening commands (Brown and Levinson, 1987; McHoul, 1987). Conversationalists, in asking questions, offer possible answers and "in so doing, they help to promote a friendly or cooperative atmosphere" (Pomerantz, 1988, p. 371). There is the structure of turn-taking, which "gives talk its rather syncopated and agreeably collaborative quality" (Boden, 1994, p. 71). Above all, there is politeness. As Chilton (1990) has written: "Politeness formulations are universal and pervasive in everyday intercourse largely conventionalized and often benign. The preferred response is reciprocal politeness" (p. 222). See footnote 2

All this might be taken as suggesting that conversation analysis reveals little more than codes of etiquette. However, the theoretical importance of ethnomethodology and conversation analysis is that they demonstrate how social order is reproduced (Heritage, 1984). At root, the codes of conversation and politeness are codes of morality. Thus, Garfinkel (1967) claimed that ethnomethodology was revealing the accomplishment of 'practical morality'. In this respect, the micro-processes of conversational interaction reveal how everyday life is infused with moral concerns. Even the momentary pauses and fractional hesitations in routine conversations can be filled with moral import. For example, turn-taking is not merely a structure which permits a conversation to proceed agreeably: it is a moral matter. As Boden (1994) suggests, by learning the routines of turn-taking "we take on that reciprocal moral code that guides human action" (p. 54).

If conversation analysis provides an investigation of practical morality, it also conveys an image of the nature of this morality. This is a morality which is seen to be routinely and successfully accomplished: it is not a morality which demands greater standards of ethics than can be practically realized. Often, but by no means always, a somewhat up-beat, positive rhetoric is used. Analysts, in quotations already presented, use terms such as 'benign', 'cooperative', and 'agreeable'. 'Members' as seen to 'collaborate', in order to 'accomplish' this morality; even the term 'member' suggests a belongingness. Because internal speculation is discouraged, the characteristics of the 'members' appear to be constituted through the talk-in-interaction. And this talk is seen to be a collaborative accomplishment, which is structured by the normative demands which are shared by the members. The result is that it is easy for 'members' to appear analytically as cooperative turn-takers, who do not possess a clouded hinterland.

This positive image of social morality is very different from Freud's view. In the Freudian account, another, more dangerous, side of human nature is visible: immorality always lurks on the edge of over-demanding morality. For a Freudian, polite codes of civilized behaviour function to restrict impulses, which continually threaten social order. The more that the demands of politeness are heard to fill conversational spaces, the more the Freudian would suspect the repression of immoral impulses. The Freudian would not ask how the conversational structures of morality are accomplished, but what temptations these structures are they trying to prevent. For example, if agreement is the socially preferred norm, then perhaps disagreement and disruption constitute a hidden temptation, whose delights must be repressed from consciousness.

Such questions tend not to be asked by conversation analysts, whose perspective to date has emphasised the achievement of conversational morality, rather than the repression of immorality. This can be illustrated by considering the conversational device of 'repair', which has been much studied by conversation analysts (i.e. Schegloff et al, 1977; Pomerantz, 1984). Introducing a discussion of conversational repair, Nofsinger (1991) states that "conversation, like other forms of human behaviour, is not perfect". People can start to say the wrong things, or they can forget a word. According to Nofsinger, "it is important for our understanding of each other that we fix these flaws before they lead to more serious and fundamental problems" (p. 124). Nofsinger, gives an example, which is taken from Pomerantz (1984; the same example is also analysed by Potter and Wetherell, 1995). One speaker suggests to another that the latter had not brought a great deal of fruit-cake. The silence, with which the remark is greeted, leads the original speaker to 'backdown' and reverse the original statement, suggesting that, after all, a little piece of cake can go a long way. The silence and the backing-down indicated that the original statement was being heard as a possible criticism, which would disrupt the collaborative interaction. And so a repair was called for.

Much effort has been spent in detailing what Boden (1994, p. 75) calls the "impressive range of repair", such as 'self-repair' versus 'other-repair' etc. For present purposes, two points are worth noting. In the first place, repairs are seen to be necessary in order to rectify 'flaws', which threaten the structure of conversation. Second, the analytic attention is upon the repair rather than the so-called flaw. The flaws, which threaten the structure of conversation, cannot be explained by that structure. Analysts tend to demonstrate how and why the repair-structure operates. Why the flaw should have arisen in the first place is typically left unexplained, even though flaws must occur as often, if not more often, then the repair work which follows them. Significantly, there is no technical terminology for describing how and why 'flaws' are 'accomplished'. The technical vocabulary is reserved for the repair-work, which is seen as the accomplishment. Indeed, the very use of the word 'flaw', as used by Nofsinger, suggests that such utterances pose a threat to the conversational accomplishment.

Given the intricacy of speech, especially in relation to the codes of indirectness and mitigation by which politeness is routinely 'done', one can see the normative structures of conversation as highly constraining. Speakers are constrained to wait their turn, express demands indirectly, mitigate refusals and so on. In these ways, rudeness, or conversational immorality, is kept at bay. What is not asked - indeed, the standard ways of doing conversation analysis discourage the question - is why rudeness so frequently appears, or almost appears, before being quickly repaired. By contrast, the interruptions of social order - or, to use Freud's phrase, the 'parapraxes of everyday life' - have often constituted the starting-point for psycho-analytic enquiry. The flaws indicate a hidden realm of desire, rather than chance accidents which happen because humans are not perfect beings. The moral codes constrain; their structures inhibit; but they do not totally dominate. Their intricacy of restraint can be read as a sign that the temptation to immorality stalks the practice of conversational morality, awaiting the opportunity to colonize vacated spaces.

Temptations of Rudeness and the Dialogic Unconscious

According to psycho-analytic theory what is prohibited beckons as temptation. If the forbidden were not tempting, there would be little need for complex moral codes. For temptation to be successfully curbed, it must not only be resisted but also denied. Social actors must chase the immoral desire from the conscious mind, lest the surface of everyday life be disrupted. All this is relevant to the practice of everyday morality, which can be witnessed in the course of even the most casual conversations. This morality suggests a realm of discursive immorality: self-repairs, successfully accomplished, indicate a temptation resisted. This can be the temptation to act directly in ways which rudely throw off the complex restraints of discursive indirection.

One might stipulate that the possibility for 'doing politeness' depends upon the capability for 'doing rudeness', and, thus, that there is a permanent flaw in the apparently agreeable, cooperative structure. Developmentally, politeness and rudeness are linked. If politeness is routinely taught, as it must be for the skills of conversational turn-taking to be acquired, then so must be rudeness. Each time the adult says 'you must say please...don't speak when others are speaking...don't use that word - it's rude', the adult is telling the child what not to say, when speaking politely. In addition, the adult is telling the child about rudeness; in pointing out what must not be said the adult is equipping the child with dangerous, disruptive tools. Moreover, the parent, in instructing the child about these matters, tends to speak in direct ways which would not classed as polite in adult company (Aronsson, 1991). In this way, the teaching of politeness, which is itself constituted through language, provides models of rudeness.

If all forbidden acts promise the possibility of pleasure, because they are forbidden, then so rudeness and disagreement might be thought to haunts all those routine, banal enactments of discursive morality. Each conversational exchange, which successfully unfolds a normative structure without breakdown, might be said to express that structure of discursive politeness and to repress the corresponding one of rudeness. However, the notion of repression implies more than that something is hidden: it implies that what is hidden is desired as a source of pleasure. What evidence, one might ask, is there for supposing that rudeness (or politeness's repressed Other) is a hidden, tempting pleasure? In Pomerantz's example of successful repair, one might ask why first speaker broke the codes which prohibit direct criticism. Was the second speaker then tempted to challenge the first speaker, instead of remaining silent? Was the spontaneous self-repair a sign of temptation avoided? These questions are discouraged in conventional conversation analysis, whose theoretical vocabulary does not provide the words for describing a 'dialogical id' in battle with the 'conversational superego'.

Thomas Scheff (1990) has suggested that the fear of 'shame' holds social interaction in play. Should the normative structure be broken, an unrepaired conversational breakdown may ensue. The outcome is likely to be recognized as 'shameful' and the speaker feels shame, which, according to Scheff, "is probably the most intensely painful of all feelings" (1970, p. 169). The speaker in Pomerantz's example corrects herself, lest she be heard as making an inappropriate criticism of her fellow conversationalist. Had she not done so, a scene - or conversational breakdown - might have ensued with all its attendant embarrassment and shame.

The use of psycho-analytic terminology is appropriate if there are reasons for thinking that what is shameful is also tempting. One might suggest that all social restrictions are accompanied by the desire - whether or not conscious, and whether or not constant or intermittent - to transcend the restriction. If this is so, then one should expect impoliteness to be an object of temptation and to represent a desired freedom. This temptation is both inhibited and created by the fear of shame and by the routines of practice, which accomplish politeness. Moreover, these routines, by their very accomplishment, regularly succeed in driving the disruptive thought from the mind. Sometimes, the rude insult might just slip out, despite ourselves. And regularly, too, and again without conscious afore-thought, the words of repair will spontaneously slip out.

However tempting this line of argument, with its suggestion of dialogical id and superego, nevertheless it is necessary to offer some signs that the rudeness is a temptation, which must be routinely repressed. One criterion for the existence of unconscious desires is that they regularly, but unpredictably, leak into conscious awareness and into the flow of social activity. Freud, especially in The Psychopathology of Everyday Life, claimed that the 'parapraxes' of social life, such as lapses of memory, indicated the workings of repressed desires. However, the case of humour is particularly interesting. According to psycho-analytic theory, in humour the socially repressed is not only expressed but is enjoyed. Freud claimed that "our enjoyment of the joke" indicates what is being repressed in more serious talk (Jokes and their Relation to the Unconscious, p. 138). Mulkay (1988) has argued that humorous discourse, by being marked off from 'normal' serious discourse, sustains "one of the basic requisites of the serious mode: namely that disagreement should be avoided as far as possible" (p. 79). The point is not merely that humorous discourse expresses disagreement and impoliteness, but that the latter, which are repressed in normal discourse, can be enjoyed in the format of humour. Comic heroes are frequently figures who disrupt the patterns of restrictive politeness. From Diogenes of Sinope, living in his tub, through to Groucho Marx and John Cleese there is a common thread. Audiences find pleasure in the character who sweeps shame and the social rules aside. Sometimes, as with John Cleese's character of Basil Fawlty, the constraints of indirection are turned into weapons of rudeness: 'thank-you-so-much', he proclaims in a parody of politeness (see Brown and Levinson, 1987, pp. 229-230 for a discussion of the way that overpoliteness can become a means of insult).

It might seem curious that displays of rudeness, which ordinarily will be treated as unethical, are greeted with pleasure in the context of humour. Freud, as is well known, claimed that the pleasure of a tendentious joke expressed a repressed desire: "To the human psyche all renunciation is exceedingly difficult, and so we find that tendentious jokes provide a means of undoing the renunciation and retrieving what was lost" (1991, p. 145). This is true of the urge to be impolite. According to Freud, "the prevention of invective or of insulting rejoinders" occurs so commonly that "tendentious jokes are especially favoured in order to make aggressiveness or criticism possible against persons in exalted positions who claim to exercise authority" (p. 149). The comic hero becomes a figure for identification, to be admired for doing precisely that which we would fear to do. Our laughter betrays the repressed wish for disinhibition, whilst, because the joke is 'only' a joke it confirms that inhibition. Peter Sloterdijk (1988), discussing Diogenes, who insulted Alexander the Great with impunity, talks of the "dialectic of disinhibition" (p. 103). Inhibitions are released within the joke; pleasure is performed safely, especially if we, the inhibited, can enjoy vicariously the insults. The joker, recognized as such, reaffirms the boundary between normal and humorous discourse, thereby strengthening the inhibition. Even Alexander the Great could join in the laughter at Diogenes. The disinhibition is localized, contained within the joke. Yet, the whole business of laughter and delight only makes sense if there is a wish which stretches beyond the moment of humour: otherwise why would the comic figure be greeted with delighted rather than outrage?

For Freud, the forces of desire were universal and biologically based; as such they threatened the very nature of social order. Polite dialogue, thus, depended upon the sublimation of forces, which owed their origin to non-civilized, non-dialogical factors. The opposition between desire and civilization was, for instance, clearly expressed in later works such as Civilization and its Discontents and The Future of an Illusion. However, repression, and thereby the unconscious, must be formed in dialogue. Without moral requirements, which must be consciously expressible, there would be no reason to repress desires. One can go further. Language, which is necessary for the production of morality, is not merely a vehicle for transmitting moral prescriptions. Language constructs and requires its own morality of talk, and thereby its own repressed desires. In consequence, the location of desire outside the processes of dialogue and social order is not necessary. The temptations of impoliteness do not stand outside the dialogic process, but are constituted within it. The desires to be rude, to contradict, 'to speak one's mind', to have done with the constraints of politeness are formed within dialogue. Such desires cannot antedate, nor stand outside, the constrictions of politeness. In this respect, it makes sense to talk of the unconscious being dialogically constituted.

If the impulses for rudeness are formed in dialogue, then, so it may be argued, are their repression. Dialogue must create its own ways to push the temptations of rudeness aside from routine consciousness, and to permit the conscious aims of dialogue to continue. The person's sense of their own self will be involved, investing this repression with greater force than a social misdemeanour which can be lightly broken. In talking politely and agreeably - in doing consideration of others' feelings, through repair, indirection and a thousand other conversational devices - we are reproducing ourselves as tolerant, moral, polite selves. The intolerant, rude, hurtful possibilities (and desires) have to be routinely pushed aside from conscious awareness. The mere thought of being rude - of criticising directly, snapping back, discarding the constraints of discursive etiquette - raises temptation in ways, which should not even be thought if interaction is to proceed routinely in its structured, reflexively monitored smoothness.

It might be said that the routine conversational intricacies, which are revealed by conversational analysts, depend upon a repression which is not routinely revealed. The pushing aside of temptation must be itself pushed aside, for, as Freud stressed, the process of repression is itself to be repressed. One might suggest that the theoretical stance of conversation analysis, as practised in the main, expresses this repression. Its image of the speaker, acting to agree, avoiding disruption, and sustaining the normative structures of dialogue, reflects the self-image of the caring, considerate, tolerant speaker with no further interior life. At the same time, residing within us is the capacity to hoot with pleasure at uncaring, inconsiderate and intolerant actions.

Dialogic Unconscious in Action

If the foregoing arguments are accepted, then it should be possible to analyse discourse, in order to examine the processes of repression in action. The Dialogic Unconscious will be constituted within talk, and the analyst, seeking to understand how repression is being discursively accomplished, will need to examine the absences of discourse. What is not said, but could easily have been, and, indeed, on occasions is almost said but then removed from the conversation, becomes of prime significance. Such analyses would extend the scope of conversation analysis, while remaining within discursive psychology's project to show how psychological phenomena are constituted within social activity. However, such analyses would not add a 'psycho-analytic dimension' in any simple way. The notion of the Dialogic Unconscious should leave neither discursive psychology nor psycho-analysis untouched.

In the first place, an analysis of the Dialogic Unconscious in discursive action lends itself to ideological analysis. The focus can be shifted from the individual to the social and to the construction of the unconscious. Repression, then, is not a universally static process, but something which is part of ideological and socio-historical currents (Frosh, 1989). Particular codes of politeness differ culturally and historically (Brown and Levinson, 1987; Katriel, 1986). So do the topics on which people are expected to converse and remain silent. Freud's patients found it difficult, even in the privacy of the psycho-analytic consulting-room, to talk of personal, sexual matters, which are now the common themes of television discussion shows (Carbaugh, 1988; Livingstone and Lunt, 1994). On the other hand, his patients would unlikely to have been as troubled by some of the themes of 'race' as were the speakers studied by Wetherell and Potter (1993) and Billig (1992). The latter two examples illustrate the conditions of ideological analysis. Analysts not only have to stand back from what the particular speakers are saying; analysts also must attempt to distance themselves from, and reflect upon, the social conventions and ideological assumptions which enable such accounts to be given (Parker, 1992 and 1994; van Dijk, 1993b).

The analyst of the Dialogic Unconscious needs to investigate how routines of talk can prevent the utterance of themes/accounts/questionings, which might seem reasonable to outsiders but which are collaboratively avoided by the particular speakers as a localized form of politeness. To the outsider, including the analyst, it can look as if there is a joint conspiracy to achieve a collective refusal of knowledgeability. If, as Boden (1994) suggests, rationality is localized, then the local rationality can also be seen as an irrationality, with its collaborative talk accomplishing discursive repression.

The traditional procedures of psycho-analysis are not, by themselves, suited for observing the repressions of the Dialogic Unconscious. In traditional psycho-analysis, one form of dialogue - the psycho-analytic conversation - is privileged as a means for discovering repressions which are presumed to have taken place elsewhere. As will be suggested, Freud was uncertain how repression was actually accomplished, being more concerned to show its incompleteness. What he did not take seriously was the possibility that conversations - including those between the psycho-analyst and patient - repressed through their expression. As conversationalists talk about one set of topics, they are keeping others from discussion. How possible topics routinely remain undiscussed - and how possible questions remain unasked - requires the sort of micro-analysis pioneered by conversation analysts, applied to discursive absences as well as presences.

As will be suggested, the analysis of discursive absences might provide clues about the operation of repression. In this respect, the notion of the Dialogic Unconscious might help to illuminate issues right at the heart of psycho-analytic theory. After all, Freud himself believed that the idea of repression, rather than that of the unconscious, was the real discovery of psycho-analysis: he claimed that "nothing like it had ever before been recognized in mental life" (1995c, p. 18). However, Freud was loth to conceive of repression in dialogic terms.

A brief example can be given of the sort of routine dialogic repression, which is overlooked in classic psycho-analytic theory. The example is taken from the famous 'Dora' case, presented in Freud's 'Fragment of an Analysis of a Case of Hysteria'. The 'Fragment' is a classic piece of psycho-analytic writing, notable for the detail of Freud's exposition and his imaginative interpretations (Marcus, 1986). In recent years, this case has become an object of intense reinterpretation. Feminist critics, in particular, have claimed that the case illustrates the dynamics of male power, which Freud overlooked and which, by his overlooking, he exemplified (for critiques of Freud in relation to Dora, see inter alia: Cixous and Clémont, 1986; Gallop, 1986; Gearhart, 1986; Irigaray, 1989; Jacobus, 1987; Lacan, 1986; Masson, 1990; Moi, 1986; Ramas, 1983; Rose, 1986). The background can be briefly given (for details of the historical situation of Dora herself, see Appignanesi and Forrester, 1993; Decker, 1991; Loewenberg, 1985; Rogow, 1978 and 1979). Dora, as a young woman of nineteen, was brought to Freud for treatment by her wealthy, domineering father. She was suffering symptoms of hysteria and had recently threatened suicide. The life of Dora's family was intertwined with that of another family in a complex pattern of deceit. Dora's father was having a protracted affair with a young married woman, Frau K, although he denied this to his own family. To facilitate the affair, he encouraged increasing contact between the two families. Frau K befriended Dora, often leaving her young children in Dora's care. Herr K had been propositioning Dora since she was fourteen, on occasion grabbing her, trying to kiss her and, once, on holiday entering her bedroom while she slept. Dora told her father about K's advances, but he refused to believe her, claiming that she, in her hysteria, was engaging in sexual fantasies. This was one of the symptoms, which Dora's father hoped Freud would cure.

It was clear that Freud did not trust Dora's father, yet he followed the remit to restore Dora to psychological health. To this end, Freud sought to uncover repressed desires, which he assumed lay behind her hysterical symptoms. Dora was, according to Freud, really in love with Herr K, and her ready denials, when Freud put the idea to her, were further proof of the desire and its repression. Freud, after the analysis was terminated, came to the further conclusion that Dora also had erotic desires for Frau K, and that she also repressed these from conscious awareness.

Critics have claimed that Freud fell into the trap of pathologising Dora, the victim of a pathological, patriarchal family. Freud's methodology was to use the psycho-analytic dialogue as a means of uncovering repressions, which were presumed to be individual and non-dialogic. His evidence for Dora's repressed wishes came from his interpretations of what Dora was telling him. In a later essay

'Repression', Freud specifically was to suggest that repression was "individual in its operation" (1995b, p. 572, emphasis in original). In searching for the individual aspects of repression, Freud's attention was directed towards Dora's individual inner life - to her dreams, fantasies, private thoughts and so on. He paid little attention to the conversations in which she would have routinely participated. One might argue that these conversations were crucial for the situation, in which Dora was caught. Moreover, if one takes the notion of the Dialogic Unconscious seriously, they might have provided models for the individual repression, in which Freud was interested.

According to the account which Freud presented, Herr K sent Dora flowers every day for a whole year, whenever he was in the neighbourhood. He constantly bought her presents and spent "all his spare time in her company" (1995a, p. 188). Freud adds an astounding, but unelaborated, remark: Herr K pursued Dora "without her parents noticing anything in his behaviour that was characteristic of love-making" (p. 188). How, one might ask, was this failure to notice achieved. It was, of course, a convenient failure for all concerned, except for Dora, the youngest and most vulnerable of the participants. What was said by Dora's parents when the flowers were delivered yet again, or when Herr K arrived with another wrapped present for the adolescent? As he stood in the hallway, with all the conventional demeanour of a lover, what routines of greeting were developed? Did they make jokes about his being a suitor?

Here was a process of repression being socially produced in overt family activity, as all refused to draw the obvious conclusion or to ask the dangerous question. K's visits and flowers, by virtue of their frequency, must have taken on the character of a habit. If jokes were uttered, did these become family jokes, which by their familiarity repressed the truth they expressed? Or did a silence, filled by other words, become routine? One might predict that a pattern was established, so that the repression was reinforced through routine, especially dialogic routine. In short, one would want to know how the routines of greeting, as well as the routine ways of talking about the visits when the topic occurred in conversation, functioned to repress other, more obvious interpretations.

Similar discursive processes of avoidance can be seen in the two examples of racist discourse, discussed earlier. The denial of prejudice is not challenged. Speakers use conventional expressions, which polite white speakers are not expected to dispute: 'I'm not prejudiced but...' is rhetorically structured for agreement (van Dijk, 1992 and 1993a; Billig, 1991; Potter and Wetherell, 1988). Questioning must be absent, for the conversation to proceed without 'flaw'. For instance, the English families, talking about royalty, were collectively denying their own racism, while arguing for the impossibility of a non-white monarchy. A discursive projection was constructed: it was others - 'public opinion' or an amorphous 'them' - who would prevent any interracial marriage - not the Queen, with whom the speakers were claiming to identify as the figure of nationhood. In this way, the speakers were distancing themselves from the racist practices, which they were tacitly supporting but attributing to others. Complex discursive strategies were used to maintain the position. Speakers avoided the word 'white', with its connotations of racial purity, and, thereby, of racism (for details, see Billig, in press a). The discursive balance was preserved by the avoidance of awkward questioning, as speakers collaborated to accomplish the collective projection.

If repression is dialogically and socially constructed then topics of repression will vary culturally and historically. The bourgeois citizenry of Freud's Vienna might have developed complex discursive codes to repress the topic of sexuality. However, the Viennese, whose elected mayor at the time of Freud's conversations with Dora was an openly racist demagogue, would be surprised, perhaps shocked, by today's inhibitions about 'race'. They would have difficulty adjusting to today's complex discursive codes of utterance and absence, by which racism gains entry to respectable homes with the demeanour of denial, rather like Herr K being welcomed through the hallway with his bunch of flowers.

Dialogic Nature of Repression

Finally, some implications from the notion of the Dialogic Unconscious can be drawn in relation Freudian ideas about the nature of repression. Necessarily these implications can only be outlined briefly. Discursive psychologists have suggested that public discourse often provides the model for individual functioning. For example, Billig (1996) proposed a rhetorical model of thinking, suggesting that internal debate is based upon public argumentation; children learn to think inwardly on the basis of their participation in the outward rhetoric of debate (see also Shotter, 1993a and 1993b; Wertsch, 1991). The same approach can be applied to repression. Through participation in dialogue, which represses as it expresses, people acquire the skills of repression. These dialogic and social skills, when internalized, provide the basis for the sort of repression, which Freud called individual in its operation.

Although Freud might have seen repression as the key concept of psycho-analysis, his writings are somewhat sketchy about describing the routine operation of repression. Routines are to be expected for, as Freud claimed, repression "demands a persistent expenditure of force", for the repression of any particular thought is not "an event which takes place once" (1995b, p. 572). When Freud attempted to describe how repression operates, he often lapsed into conceptual difficulty or mechanical metaphor (see Soyland, 1993 and Gellner, 1985 for discussion of Freud and metaphor).In 'The Ego and the Id', Freud envisaged the conscious part of the mind (the ego) repulsing unconscious forces from the id. Yet, the repulsion, accomplished by the ego, must itself be repressed. Thus, the ego must be unaware of itself and its accomplishments. Conceptually, therefore, the ego cannot be identified with consciousness for in some respects it "behaves just like the unconscious" (1995d, p. 631). But how can consciousness be just like the unconscious, pre-linguistic id? Freud recognized that there were conceptual problems: "We land ourselves in endless obscurities and difficulties if we keep to our habitual forms of expression" (p. 631).

Part of the problem lay in Freud's insistence that repression was, at root, a non-linguistic process, which was as mysterious as the non-verbal, instinctual id. There were, he admitted, linguistically constituted forms of repression, but these were more superficial, less basic forms. Freud, in his essay on 'Fetishism', specifically distinguished between repression proper (Verdrangung), which was non-linguistic, and disavowal (Verleugnung), which operated through language (Freud, 1977, pp. 352ff). Repression keeps affect away from consciousness, whereas disavowal is aimed at only the idea of the affect, rather than the affective forces themselves (see Laplanche and Pontalis, 1983, for further references to Freud's distinction). What is clear is that Freud believed that the basic process, by which the ego deflects the desires of the id, was the non-linguistic Verdrangung, rather than Verleugnung. And about the ways Verdrangung operated he could say little except by way of the mechanical metaphor.

It is possible to challenge the primacy of the non-linguistic forms of repression, while still keeping within the internal logic of Freudian theory. As Freud stressed, the necessity for repression only arises as a consequence of social morality. In 'On Narcissism', he wrote that "for the ego the formation of an ideal would be the conditioning factor of repression" (1995e, p. 558). Without a sense of conscience - and a punitive ideal to live up to - the ego would not need to censor desires. It can be argued that, from a developmental point of view, the notions of morality only arise within language: indeed, as conversation analysis suggests, talk-in-interaction forms the basis of routine, practical morality. If language-in-interaction creates moral imperatives, then it may also provide the means for their routine accomplishment. This would include the routines of dialogic repression.

Once the issues are put like this, then a step is made towards clearing some of the mysteries of repression. If repression is thought to be individual, unconscious and a-social, then the individual cannot learn (let alone be taught) how to repress. Something mysteriously innate, and non-cultural, must spontaneously occur, when cultural norms make their presence felt. On the other hand, if language-in-interaction itself depends upon repression, then in learning how to use language the child is learning lessons of repression. Polite talk will provide the models for individual repression, just as argumentation provides the models for internal deliberation (Billig, 1996). The child who learns the codes of discursive politeness, which are integral to learning to talk, learns how to change the subject, to avoid questioning ('it's rude to ask'), to turn threatening topics into humour, and so on. In learning these dialogic skills, the child is acquiring routines, which can be applied to their own individual mental life. Shockingly shameful thoughts may flash into consciousness; the individual knows how to change the subject, attribute thoughts to others, or tag them with humour. Above all, the person is acquiring routines of thought - or the buzz of internal chatter - which can keep the shocking waiting outside, pressing the bell and only gaining admission if it can appear, like Herr K, as a friendly, respectable guest.

If this line of argument is accepted, then the routine conversations in Dora's household were not irrelevant to her 'condition'. They, and other conversations, provide the models for any self-disavowals enacted within the interior mental life. The theoretical implications are considerable. On the one hand, the notion of a Dialogic Unconscious serves to make the basic processes of repression appear less mysterious and more openly observable than is posited in Freudian thinking. On the other hand, the Dialogic Unconscious depicts ordinary conversations as more psychologically and ideologically charged than they typically appear to conversation analysts. In consequence, the idea of the Dialogic Unconscious opens the way to an unlikely combination of psycho-analytic and conversation analytic ideas in ways which could broaden and deepen the project of discursive psychology.


Antaki, C. (1994). Explaining and Arguing. London: Sage.

Appignanesi, L. and Forrester, J. (1993). Freud's Women. London: Virago.

Aronsson, K. (1991). Facework and control in multi-party talk: a pediatric case study. In I. Markova and K. Foppa (Ed), Asymetries in Dialogue. Hemel Hempstead: Harvester/Wheatsheaf.

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

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

Billig, M. (1996). Arguing and Thinking, Second Edition. Cambridge: Cambridge University Press.

Billig, M. (in press a). Discursive, rhetorical and ideological messages. In C. McGarty and A. Haslam (Ed.), The Message of Social Psychology: Perspectives on Mind in Society. Oxford: Blackwell.

Billig, M. (in press b). Keeping the white queen in play. In M. Fine, L. Weis, L. Powell and M. Wong (Ed.), Off-White. London: Routledge.

Billig, M. (in press c). From codes to utterances: cultural studies, discourse and psychology. In M. Ferguson and P. Golding (Ed.), Beyond Cultural Studies. London: Sage.

Billig, M. and Edwards, D. (1994). La construction sociale de la mémoire. La Recherche, 25, 742-745.

Bilmes, J. (1987). The concept of preference in conversation analysis. Language in Society, 17, 161-181.

Boden, M. (1994). The Business of Talk. Cambridge: Polity Press.

Brearley, M. (1991). Psychoanalysis: a form of life?. In A.P. Griffiths (Ed.), Wittgenstein's Centenary Essays. Cambridge: Cambridge University Press.

Brown, P. and Levinson, S.C. (1987). Politeness: some universals in language use. Cambridge: Cambridge University Press.

Burman, E. and Parker, I. (Ed.). (1993). Discourse Analytic Research. London: Sage.

Buttny, R. (1993). Social Accountability in Communication. London: Sage.

Carbaugh, D. (1988). Talking American. Norwood, NJ: Ablex.

Chilton, P. (1990). Politeness, politics and diplomacy. Discourse and Society, 1, 201-224.

Cioffi, F. (1991). Wittgenstein on Freud's 'abominable mess'. In A.P. Griffiths (Ed.), Wittgenstein's Centenary Essays. Cambridge: Cambridge University Press.

Cixous, H. and Clémont, C. (1986). The Newly Born Woman. Manchester: Manchester University Press.

Daelemans, S. and Maranhao, T. (1990). Psychoanalytic dialogue and the dialogic principle. In T. Maranhao (Ed.), The Interpretation of Dialogue. Chicago: University of Chicago.

Decker, H.S. (1991). Freud, Dora and Vienna 1900. New York: Free Press.

Drew, P. (1995). Conversation analysis. In J.A. Smith, R. Harré and L. van Langenhove (Ed.), Rethinking Methods in Psychology. London: Sage.

Edwards, D. (1991). Categories are for talking. Theory and Psychology, 1, 515-542.

Edwards, D. (1994). Script formulations: an analysis of event descriptions in conversation. Journal of Language and Social Psychology, 13, 211-247.

Edwards, D. (1995a). Two to tango: script formulations, dispositions and rhetorical symmetry in relationship troubles talk. Research on Language and Social Interaction, 28, 319-350.

Edwards, D. (1995b). Sacks and psychology. Theory and Psychology, 5, 579-596.

Edwards, D. and Middleton, D. (1986). Conversational remembering and family relationships: how children learn to remember. Journal of Social and Personal Relationships, 5, 3-25.

Edwards, D. and Potter, J. (1992). Discursive Psychology. London: Sage.

Edwards, D. and Potter, J. (1993). Language and causation: a discursive action model of description and attribution. Psychological Review, 100, 23-41.

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

Freud, S. (1977). Fetishism (1927). In On Sexuality. Harmondsworth: Penguin.

Freud, S. (1995a) Fragment of an Analysis of a Case of Hysteria (1905). In P. Gay (Ed.), The Freud Reader. London: Vintage.

Freud, S. (1995b) Repression (1915). In P. Gay (Ed.), The Freud Reader. London: Vintage.

Freud, S. (1995c) An autobiographical study (1925). In P. Gay (Ed.), The Freud Reader. London: Vintage.

Freud, S. (1995d). The Ego and the Id (1923). In P. Gay (Ed.), The Freud Reader. London: Vintage.

Freud, S. (1995e) On Narcissism (1914). In P. Gay (Ed.), The Freud Reader. London: Vintage.

Frosh, S. (1989). Psychoanalysis and Psychology. London: Macmillan.

Gallop, J. (1986). Keys to Dora. In C. Bernheimer and C. Kahane (Ed.), In Dora's Case. London: Virago.

Garfinkel, H. (1967). Studies in Ethnomethodology. Englewood Cliffs: Prentice Hall.

Gearhart, S. (1986). The scene of psychoanalysis: the unanswered questions of Dora. In C. Bernheimer and C. Kahane (Ed.), In Dora's Case. London: Virago.

Gellner, E. (1985). The Psychoanalytic Movement. London: Paladin.

Gergen, K. (1994). Realities and Relationships. Cambridge, Mass.: Harvard University Press.

Gilbert, G.N. and Mulkay, M. (1984). Opening Pandora's Box: a sociological analysis of scientists' discourse. Cambridge: Cambridge University Press

Grimshaw, A.D. (Ed) (1990). Conflict Talk. Cambridge: Cambridge University Press.

Harré, R. (1995). Discursive psychology. In J.A. Smith, R. Harré and L. van Langenhove (Ed.), Rethinking Psychology. London: Sage.

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

Henriques, J., Hollway, W., Urwin, C., Venn, C. and Walkerdine, V. (1984). Changing the Subject. London: Methuen.

Heritage, J. (1984). Garfinkel and Ethnomethodology. Cambridge: Polity Press.

Hollway, W. (1989). Subjectivity and Method in Psychology. London: Sage.

Hopkins, J. (1995). Wittgenstein, interpretation and the foundations of psychoanalysis. New Formations, 26, 54-73.

Horowitz, A.D. (1996). 'A good old argument': the discursive construction of family and research through argumentation. Unpub. Ph.D. thesis, Loughborough University.

Irigaray, L. (1989). The gesture in psychoanalysis. In T. Brennan (Ed.), Between Feminism and Psychoanalysis. London: Routledge.

Jacobus, M (1987). Reading Woman: essays in feminist criticism. London: Methuen.

Katriel, T. (1986). Talking Straight. Cambridge: Cambridge University Press.

Lacan, J. (1979). The Four Fundamental Concepts of Psycho-Analysis. Penguin: Harmondsworth.

Lacan, J. (1986). Intervention on transference. In C. Bernheimer and C. Kahane (Ed.), In Dora's Case. London: Virago.

Laplanche, J. and Pontalis, J.-B. (1983). The Language of Psycho-Analysis. London: Hogarth Press.

Livingstone, S. and Lunt, P. (1994). Talk on Television. London: Routledge.

Loewenberg, P. (1985). Decoding the Past: the psychohistorical approach. Berkeley: University of California Press.

Marcus, S. (1986). Freud and Dora: story, history, case-history. In C. Bernheimer and C. Kahane (Ed.), In Dora's Case. London: Virago.

Masson, J. (1990). Against Therapy. London: Fontana.

McHoul, A.W. (1987). Why there are no guarantees for interrogators. Journal of Pragmatics, 11, 455-471.

Middleton, D. and Edwards, D. (Ed.) (1990). Collective Remembering. London: Sage.

Moi, T. (1986). Representation of patriarchy: sexuality and patriarchy in Freud's Dora. In C. Bernheimer and C. Kahane (Ed.), In Dora's Case. London: Virago.

Moscovici, S. (1976). La Psychanalyse: son image et son public, second edition. Paris: Presses Universitaires de France.

Mulkay, M. (1988). On Humour. Cambridge: Cambridge University Press.

Nofsinger, R.E. (1991). Everyday Conversation. Newbury Park: Sage.

Parker, I. (1990). The abstraction and representation of social psychology. In I. Parker and J. Shotter (Ed.), Deconstructing Social Psychology. London: Routledge.

Parker, I. (1992). Discourse Dynamics. Routledge: London.

Parker, I. (1994). Reflexive research and the grounding of analysis: social psychology and the psy-complex. Journal of Community and Applied Psychology, 4, 239-252.

Pomerantz, A. (1984). Agreeing and disagreeing with assessments: some features of preferred/dispreferred turn shapes. In J.M. Atkinson and J. Heritage (Ed.), Structures of Social Action. Cambridge: Cambridge University Press.

Pomerantz, A. (1988). Offering a candidate answer: an information seeking strategy. Communication Monographs, 55, 360-373.

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

Potter, J. and Edwards, D. (1990). Nigel Lawson's tent: attribution theory, discourse analysis and the social psychology of factual discourse. European Journal of Social Psychology, 20, 405-424.

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

Potter, J. and Wetherell, M. (1995). Discourse analysis. In J.A. Smith, R. Harré and L. van Langenhove (Ed.), Rethinking Methods in Psychology. London: Sage.

Ramas, M. (1983). Freud's Dora, Dora's hysteria. In J.L. Newton, M.P. Ryan and J.R. Walkowitz (Ed.), Sex and Class in Women's History. London: Routledge and Kegan Paul.

Rogow, A.A. (1978). A further footnote to Freud's 'Fragment of an analysis of a case of hysteria'. Journal of the American Psychoanalytic Association, 26, 331-356.

Rogow, A.A. (1979). Dora's brother. International Review of Psychoanalysis, 6, 239-259.

Rose, J. (1986). Dora: fragment of an analysis. In C. Bernheimer and C. Kahane (Ed.), In Dora's Case. London: Virago.

Rose, N. (1990). Psychology as a 'social' science. In I. Parker and J. Shotter (Ed.), Deconstructing Social Psychology. London: Routledge.

Scheff, T.J. (1990). Microsociology. Chicago: Chicago University Press.

Schegloff, E.A., Jefferson, G. and Sacks, H. (1977). The preference for self-correction in the organization of repair in conversation. Language, 53, 361-382.

Schiffrin, D. (1984). Jewish argument as sociability. Language in Society, 13, 311-335.

Schiffrin, D. (1987). Discourse Markers. Cambridge: Cambridge University Press.

Shotter, J. (1990). The social construction of remembering and forgetting. In D. Middleton and D. Edwards (Ed.), Collective Remembering. London: Sage.

Shotter, J. (1993a). Conversational Realities: studies in social constructionism. London: Sage.

Shotter, J. (1993b). The Cultural Politics of Everyday Life. Milton Keynes: Open University.

Shotter, J. (1995). Dialogical psychology. In J.A. Smith, R. Harré and L. van Langenhove (Ed.), Rethinking Psychology. London: Sage.

Sloterdijk, P. (1988). Critique of Cynical Reason. London: Verso.

Soyland, A.J. (1993). Psychology as Metaphor. London: Sage.

van Dijk, T.A. (Ed.) (1985). Handbook of Discourse Analysis, volumes 1-4. London: Academic Press.

van Dijk, T.A. (1992). Discourse and the denial of racism. Discourse and Society, 3, 87-118.

van Dijk, T.A. (1993a). Elite Discourse and Racism. Newbury Park: Sage.

van Dijk, T.A. (1993b). Principles of critical discourse analysis. Discourse and Society, 4, 249-283.

van Dijk, T.A. (Ed.) (1996). Discourse Studies: a multidisciplinary introduction. London: Sage.

Volosinov, V.N. (1994). Critique of Freudianism. In P. Morris (Ed.), The Bakhtin Reader. London: Edward Arnold.

Wertsch, J.V. (1991). Voices of the Mind. Sussex: Harvester Wheatsheaf.

Wetherell, M. and Potter, J. (1992). Mapping the Language of Racism. Hemel Hempstead: Harvester/Wheatsheaf.

Wilkinson, S. and Kitzinger, C. (Ed.) (1995). Discourse and Feminism. London: Sage

Wittgenstein, L. (1953). Philosophical Investigations. Oxford: Blackwell.

Wittgenstein, L. (1967). Zettel. Oxford: Blackwell.

Wittgenstein, L. (1980). Remarks on the Philosophy of Psychology. Oxford: Blackwell..

Wittgenstein, L. (1988). Lectures on Philosophical Psychology, vol 1. Sussex: Harvester.


* The author is grateful for the continuing support from members of the Discourse and Rhetoric Group at Loughborough University, and, in the present case, is particularly grateful to Derek Edwards and Jonathan Potter for their unrepressed, productive disagreement. The author also recognizes the helpful comments of Hélène Joffe, John Shotter, Teun van Dijk and the anonymous reviewers.

Footnote: 11. Hollway (1989), in an interesting and original investigation, uses psycho-analytic ideas to investigate dialogue. In certain respects her work parallels some of the present arguments. However, she does not use conversation analytic notions, which stress the occasioned character of discourse. She tends to use Lacanian ideas of signification which, from a conversation analytic point of view, do not tie in closely with the details of the utterances, which she studies.

Footnote: 2. There has been a tendency for conversation analysts to study agreement rather than disagreement, or to see disagreement as a disruption to the joint accomplishment of social life. However, more analysts are beginning to see the importance of argumentation in social life. To quote Schiffrin (1984), argument can be a 'form of sociability' (see, also, Billig, 1991 and 1996; Grimshaw, 1990; Horowitz, 1996; Schiffrin, 1987; and see Gilbert and Mulkay, 1984, for analyses of argument in science). The effect of this stance on the theoretical vocabulary of conversation analysis remains to be seen. Already, some analysts are counterposing the supposed 'preference for agreement' with an occasioned 'preference for disagreement' (Billig, 1991; Horowitz, 1996).

home Return to Virtual Faculty Homepage