University of Stirling
ABSTRACT. Although internal dialogue and inner speech form essential part of communication, it has not been given much attention in studies of interaction and language. This article suggests that the basic unit of dialogical epistemology, the triad Ego-Alter-Object/representation could be conceptually relevant to dialogical subjectivity. It proposes that internal dialogue involves different kinds of symbolically and socially represented inner Alter. Giving examples of internal dialogue and its features, this article argues for the study of dialogical phenomena in their complexity and multiplicity, whether these are involved in language, communication, subjectivity or in social representations.
Concepts like dialogicality1 and dialogical subjectivity can be conceived in various ways but usually, their definitions presuppose some kinds of interdependencies between the Ego and Alter and a multiplicity of positions that they can take with respect to one another. Two further characteristics of these interdependencies should be noted. First, dialogicality and dialogical subjectivity are not concerned with the Ego and Alter as abstract or schematic notions but with their concrete manifestations, for example, with the self versus another self, the self versus group, the group versus another group, the self versus culture and so on. In each case, one component of the dyad is interdependent with the other one. And second, dialogical subjectivity is not reducible to the Ego versus Alter in the sense of the Ego’s ‘taking the role of the other’ or the Ego being solely an actor in that interdependent relation. Instead, it is conceived in terms of multiple symbolic social representations that the Ego takes in relation to the Alter and viceversa (Moscovici, 2005). Hermans (2001) analyses this phenomenon in terms of ‘collective voices in the self’. Salgado and Ferreira (2004) and Salgado et al (in press) speak about dialogical subjectivity as involving intersubjectivity, which these authors coin as ‘the other-in-the-self’. Bearing on this, potentially, the concept of dialogicality provides a forceful alternative to more traditional approaches studying dialogue primarily in terms of interactions as exchanges of gestures and symbols or as the participants’ speech actions.
AUTHOR NOTE. Please address all correspondence regarding this article to Prof. Ivana Marková, Department of Psychology, University of Stirling, Stirling FK9 4LA Scotland, UK. Email: firstname.lastname@example.org
Nevertheless, despite considerable interest that the dialogical alternative of subjectivity and intersubjectivity has been generating during the last two decades, this alternative is still no more than a potential requiring conceptual precision so that a coherent dialogical theory may develop and become of practical benefit both in the social science research and in professional communication. With this in mind, in this article I shall take a closer look at the basic unit of dialogical epistemology, the Ego-Alter-Object/representation triad (Moscovici, 1984; Marková, 2003) and I shall suggest how it can become conceptually relevant – or extended to – dialogical subjectivity as manifested in and through the self’s inner dialogue. In doing this conceptual work, I shall draw on Salgado’s et al. (e.g. Salgado and Ferreira, 2004; Salgado et al. in press) concept of ‘the other-in-the-self’.
1. Some basic presuppositions of dialogical interdependencies
As dialogicality and dialogism are receiving more and more attention among human and social scientists, we must face a challenging question. Can we identify any specific features that differentiate dialogicality and dialogism from other approaches studying conversation and dialogue, for example, ethnomethodological, interactionist, cognitive, structuralist or constructivist? Quite understandably, these latter approaches cannot be bundled together as an undistinguished bloc. They differ from one another in terms of their theoretical and empirical priorities, some emphasising the role of the speaker, others stressing the sequentiality of contributions, still others focusing on verbal or non-verbal interactions, exchanges, the mutual construction of meaning, speech acts, and so on. Yet the question holds: can we identify some distinguishing features between the latter approaches on the one hand and dialogicality on the other? This question cannot be answered without posing another one: what are the basic presuppositions of dialogicality2?
In concurrence with my previous writing (Marková, 2003), I shall presuppose that dialogicality of the Ego-Alter is of ontological nature. This means that in and through communication the Ego-Alter intersubjectively co-constitutes one another: one does not exist without the other. We can find an explicit display of this position in neo-Kantian philosophers and linguists of the early years of the twentieth century, who actually coined the term ‘dialogism’, or even ‘existential dialogism’, like Buber, Rosenstock, Rosenzweig and Cohen, among others (Marková, 2003). This position was further developed in the nineteen twenties by the Bakhtinian Circle in Russia which included, in addition to Michail Bakhtin himself, scholars like Voloshinov and Medvedev, as well as others of that period like Yakubinski, Shpet and Russian formalists. After the disappearance of the Circle during the Stalinist persecution in the Soviet Union, Michail Bakhtin continued this work in his quiet isolation until his death in 1975.
But the idea of the Ego–Alter interdependence as a point of departure for the study of human phenomena and specifically, for the concept of language, has been also pursued more generally by other researchers. For example, for the French linguist Emile Benveniste, the interdependence of the I and you formed an essential feature of language: ‘language provides the very definition of man’ (Benveniste, 1971, p. 224). Since language is ‘in the nature of man, and he did not fabricate it’ (Beneniste, 1971, p. 224), Benveniste argued, it is wrong to describe language as an instrument of communication or as a tool that the mankind invented. Instead, language is the human condition. This clearly stated position of Benveniste challenges the often cited image of language attributed to Vygotsky (1962), according to which speech and the written word is a cultural and symbolic tool given to the individual by society. In contrast, for Benveniste, the polarity of the Ego–Alter is such that ‘neither of the terms can be conceived of without the other…the condition of man in language is unique’ (Benveniste, 1971, p. 225). Therefore, the image of language as a tool is a misconception. Bearing on this formulation, we can conclude that for Benveniste, too, the Ego–Alter interdependence in language is ontologically given. This further entails that meaning-making and thematisation of contents are joint communicative activities of the Ego and Alter. The dyad is mutually responsible for the meaning-making, the Ego being an addressee for the Alter and vice versa. Their interdependence, nevertheless, does not imply that, dialogically speaking, positions of the Ego and Alter merge into one another. On the contrary, their subjectivities, rather than being stripped off their independence, are enriched in and through their interdependence. Both the Ego and Alter seek visibility and recognition by one another, as each subject actualises his or her potential through interaction and communication. Since communication is never a ‘neutral’ transmission of information, understanding and acknowledgement of the Ego and Alter is judgemental and evaluative. Their thoughts are communicable ‘in and through language’ (to use Benveniste’s (1971) expression) and equally, dialogue is shaped by participants’ concepts and ideas.
Social environment, in which the Ego–Alter dialogue takes place, rather than being a kind of a stage for the performance, is itself dynamically interdependent with dialogue. The dialogue shapes its social environment which, in turn, has an effect on dialogue. For example, the dialogue between political dissidents and the totalitarian government takes place in the public sphere and alters public opinions shaped in that public sphere. The public opinion in turn affects the nature of the dialogue between dissidents and their opponents.
1.2 The multifaceted nature of dialogicality
The above claims amply point to the multifaceted nature of human thinking and dialogue. In referring to different forms of thinking in a daily social encounter and in science, Serge Moscovici (1961/1971) uses the term ‘polyphasic thought’. Conceptual and polyphasic thought is communicable or better, it is dialogical. Michail Bakhtin (1979/1986a; 1979/1986b; 1984), discussing the multifaceted nature of dialogue, introduces the term ‘heteroglossia’. The nature of dialogue expresses itself in multiple ways, for example in the subject’s positioning (Salazar Orvig, 2006), meaning potentialities (Linell, 2005), communicative and speech genres (Marková, 2001) and so on. Focusing specifically on the dynamic multiplicity of relations into which the self can enter, Hermans (2001) is developing a theory of personal and cultural positioning of the dialogical self. His theory presents the self as moving in multiple inter- and intra-psychological positions that are mutually intertwined. In any dialogical situation the self naturally changes its positions, for example, from being a father to being a researcher or a colleague. Moreover, the self also speaks from various cultural positions, expressing different ‘collective voices’ and using different ‘social languages’, e.g. national languages, speech genres, languages of specific groups like adolescents, professions, age groups and so on. Hermans’s theory assumes that dialogical relations are ‘embodied, spatialized and temporalized processes’ and it illustrates ‘how individual voices coexist and are interwoven with collective voices’ (Hermans, 2001, p. 266).
1.3 Implications for a dialogical theory of knowledge
Dialogicality as an ontological position implies a dialogical theory of knowledge or a dialogical epistemology. It further presupposes that ontological interdependence of the Ego-Alter entails joint communicative and meaning-making activities; moreover, that the Ego–Alter is interdependent with social, historical and cultural environments; and that conceptual thought is communicable ‘in and through language’.
Dialogicality of the Ego-Alter is a fundamental feature of the dialogically conceived theory of social representations. In contrast to the position that knowledge is generated by the individual cognition à la Descartes or by collectivity à la Durkheim, dialogicality assumes that knowledge is communicatively generated by the Ego –Alter and that it is captured as a triadic relation Ego-Alter-Object. This triadic relation through which knowledge is generated has been recently elaborated in a number of different contexts (e.g. Bauer and Gaskell, 1999; Jovchelovich, in press; Marková, 2006).
The concept of triadic relation in the formation of knowledge is relevant not only with respect to external objects that are part of social reality but also with respect to the self as an object of knowledge. This is already implied in George Herbert Mead’s notion of the interdependence between I and Me, both co-constructing self-identity and self-esteem, etc. This position is also implicit in Mark James Baldwin’s work as well as in that of other pioneers of the concept of the social self, arising from the idea of the self-other interdependence, e.g. Brooks and Lewis, Papoušek and Papoušek, and Selman, among many others.
But where do we go from theoretical positions expressed by Mead’s (1934) ‘taking the role of the other’ and ‘conversation of gestures’ or Baldwin’s (1895) ‘dialectic of social growth’ and from empirical demonstrations that self and other are interdependent? Despite these significant contributions and despite the genuinely ‘social’ character of these approaches, the ‘other’ and the ‘social’ still remain abstract notions. Can we get any closer view of the dialogical interdependence of the Ego-Alter?
2. From interpersonal interaction to ‘the third party’
2.1 The problem of complexity
Humans involved in day-to-day conversation, in reading novels, watching television or negotiating their points of view are well aware that dialogue takes place simultaneously at different levels and in different forms. Hiding some ideas and strategically exposing others, carrying internal dialogues, changing their points of view, abandoning or contradicting ideas that they defended earlier – all these processes belong to the art of conversation as well as to the established common-sense practices in communication. For great novelists and writers this complexity of language and dialogue provides infinite resources for exploring the creative nature of conversation, dialogical cognition, emotions and, we can say, for exploring the human drama in its entirety. But these multifaceted features of dialogical communication create tremendous difficulties for social science researchers trying to empirically combat such complexity.
Let us consider some examples to reflect on the extent of this problem. Generally speaking, studies of world views in social psychology, public enquiries, surveys, opinions, social attitudes and otherwise are, above all, the studies of individual world views and they are linked with the notion of the individual attitude. This notion was explicitly introduced into social psychology in 1935 by Gordon Allport.3 He repeated the same definition two decades later, defining attitude as ‘a mental and neural state of readiness, organised through experience, exerting a directive or dynamic influence upon the individual’s response to all objects and situations with which it is related’ (1954, p.45). This definition still dominates empirical studies of attitudes and public enquiries based on questionnaires, scales and interviews. It is a perspective ignoring that individuals belong to various groups, that they are committed to particular collective positions and that they defend, sometimes simultaneously, different kinds of socially shared knowledge. Instead, Allport’s perspective treats groups as something unreal. What is considered to be real is only cognition of the individual. If we transpose this perspective into, say, communication in focus groups, we find that much social research treats a focus group as an aggregate of individuals who argue their individual opinions against one another (for criticism of this position, see Grossen, 2006).
But even if we casually inspect social interaction approaches like conversation analysis, talk-in-interaction and various kinds of discourse analysis that presuppose the I – you interdependence, we find that in their theoretical and empirical priorities they close windows to interactions in their complexity. For example, conversation analysis presupposes that dialogical participants are involved in joint meaning-making activities; some discourse analytic approaches emphasise the interdependence between dialogue and context; and so on. Theoretically, they acknowledge that speakers mutually construct their meanings, that their speech is filled with others’ speech, world views and collective opinions. They emphasise that dialogue takes place in specific contexts, that speakers perform different and multiple communicative activities, and so on. Empirically, however, they use extracts from conversations that are usually brief in terms of a number of dialogical contributions. Since their aim is to study the organisation of talk, accountable patterns of meaning and cultural contextualisation of actions in talk, such procedure enables them to focus on turn-taking, sequencing and actions like repetitions, repairs, adjacent pairs and so on (e.g. Drew and Heritage, 1992).
If we turn again to social psychology, the studies of interpersonal interaction have been avoiding strategies examining interactions in their complexity. Instead, they have been traditionally concerned with verbal and non-verbal interactions, e.g. frequencies of gaze, gestures, seating arrangements and so on, using established statistical or qualitative methods, e.g. coding, content analysis, description of interactions and so on. Such analyses are being performed in the name of science: in order to understand complex processes, the researcher must first examine elementary interactions, rules of sequencing and meanings of elementary communicative phenomena. After all, there are good historical reasons for that. The aim of social psychology should be to create science and not journalism, as the British social psychologist Michael Argyle used to emphasise. But such perspective holds only if we presuppose that language and communication is constructed from elements which, if put together, can reveal the complex message. Yet there is little evidence that multifaceted communication is a phenomenon constructed from simple elements.
Thus it appears that while a daily dialogue explodes from the richness of ideas involving complex social phenomena in which interactants are implicated, texts or conversations examined by means of selected brief extracts from big research corpuses of data, are likely to lose all or most of its human relevance. In this way the student of dialogue and conversation deprives him- or herself by own choice of the possibility of exploring social reality. This is not to argue in principle against using brief and selected extracts from discourse in order to analyse certain interactional phenomena, e.g. the structure of interaction and organisation of talk. Nevertheless, it is not clear whether information so obtained makes, or can make, a contribution to the study of knowledge of language in communication, grasping contents, feelings and in general, to the meaning-making of talk.
We can conclude that current studies of public opinions, of social interaction and dialogue share a characteristic militating against the study of the multifaceted nature of thought and communication.
They assume that ‘the other’ is a kind of an ‘objective entity’ interacting with the self. For example, such a brilliant scholar as Fritz Heider, in characterizing ‘the other as perceiver’ and the self ‘perceiving the other person’ assumes that the self and the other are ‘”objects” [that] have color and occupy certain positions in the environment’ (Heider, 1958, p. 21). Of course, he claims that in contrast to physical objects, persons are not manipulanda; they are centres of action that have intentions, motives and desires, which either can be accurately perceived or misperceived. Nevertheless, he defines the Alter by his or her qualities that the Ego may or may not perceive correctly. But Moscovici (2005) draws attention to the fact that the Alter is not an ‘objective’ entity, but a social representation generated jointly by the Ego and the Alter. Heider’s focus on social perception, i.e. the focus on the person’s ability to perceive the other correctly detracts attention from the fact that ‘perception’ requires a dialogical and culturally based theoretical analysis that is to be captured not by means of a social perception but by a social representation. The former conception ignores the Ego’s dialogical subjectivity.
Just like Fritz Heider, so Gustav Ichheiser and Erwin Goffman and much of the extensive research on self-management, self-perception and self-control, all focus on perceiving correctly the qualities of the Alter that the self attempts to manage, control and understand, and so on. Thus we can say that both classic as well as contemporary interactional approaches share fear of addressing the complexity of human dialogicality. It is qualities of the Alter that are supposed to determine strategies of the Ego, for example, to hide, to make invisible and otherwise, the Ego’s characteristics relevant to the interaction in question.
In contrast to the prevailing research trend directing attention to the study of relatively simple interactions or on social perception as defined by qualities of the other, dialogically orientated researchers have started extending the concept of Ego-Alter interaction to ‘third parties’, ‘third person’, ‘virtual others’, ‘other others’ or the ‘positioning’ of the self with respect to physically or symbolically co-present ‘others’. They argue that interaction between the Ego and Alter necessarily involves or refers to some other Alter that is not immediately present in dialogue. For example, Wibeck et al. (2004) show that focus group discussions often invoke references to ‘virtual’ parties. Since the detailed analysis of notions like ‘third person’, ‘virtual others’, ‘other others’ and so on would go far beyond the scope of this article, I shall consider here one conceptual characteristic in relation to these different notions of ‘the third’. I shall refer to ‘the third’ as being either ‘outside’ or ‘inside’ of the Ego-Alter dialogue.
2.2 The third party entering dialogue from the ‘outside’.
Let us first consider the meaning of ‘the third party’ entering dialogue from the ‘outside’. The term ‘the third party’ is commonly used in sociology, politics, rhetoric, in the media and otherwise. Since holding a dialogue does not mean that participants are always in harmonious relations, that they understand and agree with one another, it is not uncommon that the opposing parties cannot resolve their disagreements and face a conflict. In such cases they may call on ‘the third party’, a mediator who enters into the conflict from the outside and offers a ‘neutral’ position. The term refers to someone or something that suggests a solution to the problem in question through searching for a common ground, which the two existing main parties cannot find or cannot manage on their own. In a more general sense, the third party may offer products that the two main competitors cannot provide. The third party can also benefit from its position as a negotiator and consequently, its interest might be to take an advantage from the conflict or competition between the two contesting parties.
Georg Simmel’s (1950) classic analyses of triads in terms of the non-partisan and the mediator, the tertius gaudens and divide et impera, draw attention to subtle and specific psychological characteristics of interdependencies in triads that do not exist either in dyadic interactions or in more than three-party interactions. In the case of the non-partisan and the mediator, the third party represents an impartial element, an arbitrator between the two who cannot find a solution to their controversy on their own. An example of this could be the conciliatory party trying to solve the conflict, like that between employees and the employer, unions and the management, and so on. In contrast to this case, in which the arbitrator and mediator saves the dyad from splitting up, in the case of tertius gaudens, i.e. ‘the third who enjoys’,the third party promotes its egoistic interests. Rather than functioning as an independent arbitrator, the third party may grant support to one of the parties in conflict or may make the two parties in conflict compete for favour of the third thus taking advantage of the conflict. Finally, the third party may bring about the conflict between the two participants while succeeding in making itself invisible, leaving the two in hostilities. In this way it dominates the situation: divide et impera.
From my point of view, the important characteristic of ‘the third party’ discussed here is that it enters dialogue from the ‘outside’. It is not a dialogical party at the beginning of the encounter but it enters as a mediator of positions that are not reconcilable by the two original dialogical partners. In this way the third party then becomes physically present in a dialogue, although it may intend to remain invisible either in negotiation or in the outcome.