Representations, Text and Practice: a Discourse-Analytic Model of the Consumer

Darryn Mitussis, University of Oxford
Richard Elliott, University of Exeter
ABSTRACT - This paper develops a dynamic, language-centric model that links consumers to their social environment. The model is developed by integrating insights from Serge MoscoviciBa French social psychologistBand Norman FaircloughBan English critical linguist. The model is used to examine briefly ways in which companies compete for consumer mind-share in and over language. Examples from the marketing of high technology are used to illustrate the theoretical.
[ to cite ]:
Darryn Mitussis and Richard Elliott (1999) ,"Representations, Text and Practice: a Discourse-Analytic Model of the Consumer", in NA - Advances in Consumer Research Volume 26, eds. Eric J. Arnould and Linda M. Scott, Provo, UT : Association for Consumer Research, Pages: 312-319.

Advances in Consumer Research Volume 26, 1999      Pages 312-319


Darryn Mitussis, University of Oxford

Richard Elliott, University of Exeter


This paper develops a dynamic, language-centric model that links consumers to their social environment. The model is developed by integrating insights from Serge MoscoviciBa French social psychologistBand Norman FaircloughBan English critical linguist. The model is used to examine briefly ways in which companies compete for consumer mind-share in and over language. Examples from the marketing of high technology are used to illustrate the theoretical.


The purpose of this discussion is to build a conceptual model in which language mediates the link between consumers as individuals and their social environment. In this model situated language use is an integral part of nderstanding consumers.

The marketing literature contains much useful theorizing on the way that alternate methods may be used to understand consumers (see, for example, Thompson, 1997). Unfortunately, we have yet to build a model of the interplay between individuals and the social structures that give order and rules to human life. Particularly important is to build a model that includes a role for the way in which firms can affect these social structures in and through language in their competition for consumer mind-share. The battle for consumer mind-share in language is obviousBadvertisers are constantly trying to convince consumers of the benefits of their products. The battle over language is a little more subtle but involves restructuring the links between words, images and thoughts. Changes in the way that words, images and thoughts are tied together in consumers’ minds can work to define the bounds of competition. For example if 'PC’ becomes synonymous with the Wintel platform then a consumer’s search for a new 'PC’ is limited to computers that meet this criterion. In such a case, personal computers running other operating systems on other hardware platforms are automatically excluded from the decision making process. [Wintel refers to a combination of Microsoft Windows operating system running on computer hardware designed and built around Intel computer chips.] In this sense, language has served to limit commensurability. Language therefore should be seen no longer as a neutral system of signs with permanent one-to-one correspondence with phenomena. Rather language becomes a more complex dynamic socially constructed system which competitors may seek to influence in order to change or entrench individual consumers’ constructions of social (market) structures.

The interaction between individuals and the social is particularly important for the introduction of new productsBespecially new technology. This is so because existing social structures constrain the way that new technology might be implemented and simultaneously new technologies impact on individuals in ways that might lead to the changing of social structures. For example, new communications technology facilitates a more mobile work-force, enabling employees to work on the road, from home or from shared regional offices. However, typical work practices, particularly established supervision and monitoring systems, limit the scope for the implementation of such technology to its fullest potential. On the other hand, even partial implementation of the technology induces changes in our understanding of the nature of work and the social structures that guide it. This change then affects the way that technology will be implemented and so on.

One of the difficulties is that all socio-structural constructs (such as established structures of authority at work) are generated individually but only through social interaction, though the exact nature of this relation is the subject of debate (Pettit, 1995). These social constructs are usually quite subconscious in their form but made manifest in human action, often as ideology, tacit knowledge, common sense or unstated assumptions. In this context, we mean ideology to be the logic that arises from a particular set of circumstances. In the past ideology may have been considered as arising from a particular set of material circumstances. However, in a post-structuralist world replete with symbols and symbolic consumption, it would seem more appropriate to view ideology as the logic that arises out of a particular set of shared symbolic circumstances (see Thompson, 1990; Fairclough, 1992). In general, it is fair to say that the term 'ideology’ is more weighted with concepts of power and dominance that the term 'common sense’. In order to minimize a long and tortured review of the definitions of these terms, and to avoid confusion with the lay meaning of 'ideology’ for those less familiar with the critical studies literature, the term 'unstated assumption’ will be used throughout this paper. Because, at least for the purpose of the following discussion, the distinction between these concepts is not relevant, the term 'unstated assumption’ should be taken to mean those naturalized, common sense assumptions and social facts that are required for tets to be understood, be they ideology, tacit knowledge or common sense.

The subconscious form of unstated assumptions makes them difficult to identify. By extension, the way in which they constrain human action is also difficult to identify and, more difficult still, how they might change and to what effect. Their importance though is without question. For example, many years ago Kuhn (1970) demonstrated the importance and magnitude of change in what we might call scientific common sense as paradigm changes occur. Any approaches or perspectives, therefore, that help us understand the way that unstated assumptions manifest themselves in consumers’ perceptions of the world in generalBor in their perceptions of product categories and the social settings in which they will be usedBwould be most useful in the analysis of consumers. It is worth repeating that this is particularly the case with the adoption of new technology which may either be seriously hindered by shared unstated assumptions or which may lead to paradigmatic change.

It is important at this juncture to illustrate the importance of language in this process and it might be best to quote Hodge and Kress (1993: 5-6) to highlight its link to thought and, hence, the importance of studying language:

"Languages are systems of categories and rules based on fundamental principles and assumptions about the world. These principles and assumptions are not related to or determined by thought: the are thought ... Such assumptions are embodied in language, learnt through language, and reinforced in language use."

Critical discourse analysis is an approach that offers much to the study of the issues and phenomena introduced above. Critical discourse analysis is a technique that has its origins in linguistics, especially sociolinguistics and cognitive linguistics, and critical social and political studies (see Fairclough, 1992: 3). It is worth noting that critical discourse analysis varies from grammar-centric approaches to discourse analysis (see, for example, Brown & Yule, 1983) by integrating a critical social dimension to the study of language (Fowler, 1996: 3-4). The critical approach is worthy of consideration in business research for two important reasons. First, business is about social structuresBcompanies, markets, consumption communities and the like are all social phenomena that provide structures that enable and limit behavior, especially the exercise of power. Second, the approach is different enough from the conventional business research paradigms to provide a useful basis for critical evaluations of current methods.

The techniques of critical discourse analysis are used to identify unstated assumptions that are required to be accepted in order for the reader to make sense of the text. The relationship between the unstated assumptions and social power relations is also an important part of the approach. However, in order to understand how this might occur a theory linking individual cognitions to social processes is needed.


The theory of social representations has developed in France since the 1960s (Farr, 1987), although similarities can be found in the work of Durkhiem, in France (Deutscher, 1984), and of Thomas and Znaniecki (1918-20), in the United States (Jaspars & Fraser, 1984; Farr, 1987). The theory has been operationalized in the social psychology domain with a wide variety of methods.

Social representations are constructs that fit between concepts and precepts (Moscovici, 1984) and may help marketers better understand culture (Farr, 1987), attitudes (Jaspars & Fraser, 1984), opinions and image (Moscovici, 1963, as quoted by Farr, 1987: 344), innovation diffusion (Farr, 1981; Farr, 1984), communications (Runkel, 1965, as cited by Jaspars & Fraser, 1984: 119) and group dynamics (Galam Moscovici, 1991). Social representations are shared images that permit us to give objects, persons and events "a definite form, locate them in a given category and gradually establish them as a model of a certain type, distinct and shared by a group of people" (Moscovici, 1984: 7). Because they provide models for categorization and evaluation, social representations are behaviorally prescriptive (Moscovici: 1984). Social representations function on a psychological level but are social in the sense that they are only constructed through social interaction (Moscovici, 1984). Two psychological processes are introduced by Moscovici (1984) to explain the working and creation of social representations: anchoringBthe process of classifyingBand objectifyingBthe process of taking something unfamiliar and making it familiar.

Anchoring (Moscovici, 1984: 29-37) is a

"... process which draws something foreign and disturbing that intrigues us into our particular system of categories and compares it to the paradigm of a category which we think to be suitable. It is like anchoring a stray boat to one of the buoys in our social space ... To anchor is, thus, to classify and to name something."

The importance of this is that each object is tied to others and assumes its place in a hierarchy of objects named and classified. Moscovici (1984: 32) considers the buoys to be prototypes representative of a class. By generalizing we take a feature and classify objects according to which prototype has that feature. When particularizing the differences between an object and the prototype are examined. Moscovici (1984) found that by exaggerating a specific feature of the prototypical psychoanalyst, an image of an American psychoanalyst could be produced. Consider the following marketing example. We might generalize a new type of computerBsay the Network ComputerBto be a PC because it consists of a monitor, standard form CPU box, keyboard and mouse. However, it may be unfavorably evaluated because upon particularizing it lacks the standard user interface (MS Windows), has no hard-drive and no Pentium CPU. In this case, the promoter of the Network Computer should try to create a new social representationBof a network-centric computerBby providing the material for individuals to objectify the new device.

Objectifying (Moscovici, 1984: 37-43) happens in two stages. The first stage is "... to discover the iconic quality of an imprecise idea or being, to reproduce a concept in an image." This iconic quality becomes the way we consider something visually whether they are objects, actions or even qualities. In the example above, this might be by designing a different box, borrowing different metaphors to describe the operating system and placing the new device in a different environmentBwhere PCs are not commonBor marketing it to different individualsBfinance managers or schools because of its low cost. The second stage occurs when, according to Moscovici (1984: xx, original italics),

"... the distinction between image and reality is obliterated ... what is perceived replaces what is conceived [but since] they must have a reality we find one for them, not matter what. Thus, by a sort of logical imperative, images become elements of reality rather than elements of thought."

Moscovici (1984) provides three reasons why we might create social representations: the first, the hypothesis of desirability, covers the pragmatic need for communications; the second, the hypothesis of imbalance, suggests social representations are methods of resolving tensions caused by a lack of social integration; and, the third, the hypothesis of control, suggests that groups create representations to "filter information ... and thus control individual behavior." The following sections will deal with the hypothesesof desirability and control as the two hypotheses that are most relevant to marketing.


The hypothesis of desirability is recognition that social representations are fundamental to effective communication because words and social representations are inextricably intertwined. If the representations entangled with the words we use are not shared, meaning will not be shared. Quine (1960) has claimed as much when arguing that radical translationBfrom one language to anotherBis impossible because individuals from two cultures could not possibly share the same histories and, therefore, by extension, representations. This is, of course, contrary to the idea of unequivocal literal meaning (see Chomsky, 1972). Moscovici (1984: 11) goes further, claiming that the structure of the language system and its underlying representation system can be changed with the introduction of new words or the alteration of their context. For example,

"[if] the word 'neurosis’ were to disappear, and to be replaced by the word 'disorder’, such an event would have consequences far beyond its mere significance in a sentence, or in psychiatry. It is our inter-relations, and our collective thought, which are involved and transformed."

Research by Runkel (1965, as reported by Jaspars & Fraser, 1984: 119) found that more effective communication occurred between instructors and students with similar underlying cognitive representations, but independent of attitudes to those representations. Jaspars and Fraser (1984) believe that if Runkel’s results can be generalized, larger changes in attitudes could be effected, because information transmission is more effective, if both parties shared the same underlying representations. Citing research by Rogers and Shoemaker (1971) in support of this claim, Jaspars and Fraser (1984: 120, original italics) state:

"Research on the diffusion of innovation ... in which source and receiver have by definition different social representations shows convincingly, though sometimes only in anecdotal form, how ineffective attempts to change attitudes can be if differences in social representations are not taken into account."

The creation or modification of social representations would, therefore, seem to be a powerful tool for marketers attempting to understand, or effect, attitude change, manage commensurability, or encourage product diffusion. [The possibility to manage commensurability is enabled because both the contents of social representations and the links between them are the basis for decomposing a representation into its attributes. The attributes might then used as the basis for attribute-by-attribute comparison. An example here might be of some of the new generation of communication technologies that put small computers into a mobile telephone form factor. If these new products are closely linked to mobile telephones then the new products will be unfavorably compared to mobile telephones with respect to size. On the other hand, if a new shared representation of the product can be developed, the link to the shared representation of mobile telephones may be so weakened that size becomes a secondary or tertiary basis for comparison, after particular attributes of functionality.] The link between language and social representations is the subject of discourse analysis and is of essential interest to the marketers of new technology because the first experiences of these products by consumers is always through languageBin product announcements, press reports, advertising, and word of mouth communication. Insight into how consumers view the world can be derived from analysis of their texts that present and re-present their unstated assumptions.


The hypothesis of control (Moscovici, 1984: 23) suggests that groups could create representations to "filter information ... and thus control individual behavior." For Moscovici (1984: 35) the social interaction that creates social representations is an exercise of power, that is we are "forced into an identity matrix [we] have not chosen and over which [we] have no control." By contrast, Rommetveit (1984) conceives of more autonomous individuals who, during normal discourse, transcend thir private worlds and contract to use particular categorizations of concepts and use specified images and words to represent them. Again within the marketing domain, the former view may explain the adoption of new representations of computers as an exercise of a CEOs influence over the media leading to articles in print which simply recite the press releases and thus to an audience which has no option but to accept the representations or fail to make sense of the material. The second view may help to explain the reframing of particular symbols (representations) by subcultures (see, for example, Ritson, Elliott & Eccles, 1996). At play here is a structuralist / post-structuralist tension. Fairclough (1992) builds a model that integrates these two philosophical extremes by not overstating the constructionist position inherent in the structuralist view and building a model where there is reciprocity between individuals and the social systems they perceive. In this model we are, as Giddens (1984) also understands, each a product of our social interactions and the social systems that structure them but these only exist because we are actively constructing our own models of the world based on our interactions. This process leads to both a re-presentation of the representations we have adopted and a shared struggle to create new ones.


Shared unstated assumptions are the facts, rules and logic that arise out of a particular set of shared significationsBparticularly the signification of the physical world, social relations and social identities. These significations and constructions of reality are built into the frameworks that provide the rules for the production of texts and which serve to produce, reproduce or transform power relations (Fairclough, 1992). The frameworks that provide these rules for the production of texts are referred to as orders of discourse. These orders of discourse are the subject of struggle as they determine what may and may not be said and even what is thought in particular settings. A similar position is taken by van Dijk (1997) in his development of the context model. An obvious example is the struggle over the information that should and should not be included in advertising or on product packaging. This struggle extended to how both laws and conventions structure how information should be expressed and displayed and both consumer groups and manufacturers attempt to alter perceptions of the acceptable information and its form.

The term power relations refers to both power relations between the corporation and consumers and between corporations. Marketers of new technology, for example, exercise the power they haveBby way of access to media and advertising, standards committees, wholesalers and retailersBto change the common sense notions about the nature of and uses for the current technology and attempt to naturalize representations about the new technology. The extreme in success would be to transform the way that the relevant parties signify and construct the world with respect to the new technology so that the choice of that product is naturalized.


The discussion above has implications for the way that we conceptualize competition. It helps to perhaps re-conceptualize competition as a struggle to limit commensurability that occurs in and over language. As suggested in the introduction, the battle in language occurs as advertisers and sales-reps are constantly trying to convince potential consumers of the benefits of their products. The battle over language is a little more subtle. Another example that comes to mind is the battle to propagate words and phrases for new conceptsBsay, the attempt by Network Computer Inc. (an alliance of Oracle, IBM, Apple, Sun and Motorola) and Inte Corporation to establish either 'Network Computer’ or 'Connected PC’ as the dominant phrase to describe the $500 internet appliance. Even while the concepts underlying the details of the two competing views are still being finalized, and even though the functionality of the equipment would seem similar to the end user, each phrase has different links to information within our existing knowledge structures. If initially the phrase 'network computer’ becomes dominant then consumers’ expectations and evaluations about the technology will be different to if the phase 'connected PC’ becomes dominant. Furthermore there exist existing links to the respective companies that may be the source of market powerBhere we might consider how 'windows’ has come to suggest Microsoft even though operating systems by Apple Computer, Sun, Silicon Graphics, Be and others have windows. In both cases though, to win the battle companies need the consent of the commentators, analysts and consumers. Given our acceptance of a reciprocity between the social and the psychological, as discussed above, competition between companies, when viewed this way, begins to look like a hegemonic struggle.



The concept of hegemony comes from the Italian theorist Gramsci who suggested that the dominant groups within society maintained influence not only by threat of force but also by consent. Developing and maintaining power requires the leading groups to make compromises that hold various groups together thus creating the critical mass required for favorable representations to develop. These representations are then presented and are re-presented so that they become part of our subconscious, forming that which might be called 'common sense’ and which we have generically labeled 'unstated assumptions’. Fairclough (1992: 92) identifies in Gramsci "a view of 'common sense’ as both a repository of the diverse effects of past ideological struggles, and a constant target for restructuring in ongoing struggles." As discussed above, if unstated assumptions are produced and reproduced in social practice then power is maintained. In technology industries this happens through the use of standards setting bodies and through strategic alliances. We may consider standards setting and strategic alliances as rhetorical tools that help build a critical mass of people with shared representations of an industry or issue, especially when these individuals are in the position to disseminate their world views in the media or by implementing policies and systems that entrench them. Once proposals become standards not to use them is nonsensical and thus those who are able to gain the consent of the standards committees have their power in the standards committees reproduced in the larger market place. We can see similar struggles taking place in hearings by mergers and monopolies commissions, as producers attempt to secure supply chains and so on. After time, though, that a distributor or retail carries a particular brand or product category becomes not a subject for struggle but a matter of common sense. The example of standards setting is one of the most obvious. Most of the struggle over language is far more subtleBas the 'Network Computer’ and 'Connected PC’ examples suggest.

The struggle to influence the nature of our underlying representations of a particular technology and its applications is one that we would expect to see carried out through the entire length of the supply chain. We would expect that the most successful firms are those which manage to create similar or complementary representations of their products and uses thoughout the supply chain so that the manufacturer, distributors, retailers and consumers are able to effectively communicate. More than this, propagating particular sets of unstated assumptions affects the way in which people define themselves as consumers of certain products; helps consumers define themselves and the products; and helps consumers and others construct the relationships that need to be acted out in the purchase and use of the product.


Fairclough (1992) examines three practices as part of three intertwined perspectives that can be used to identify the structure of representations implicit in text production and consumption. These are social practice (content, relations and subjects), their associated structural constructs (knowledge and belief, social relationships and social identities) and their associated textual values (experiential, relative and expressive). This structure is represented in Table 1.

Social practices are "the real instances of people doing or saying or writing things" (Fairclough, 1992: 57) and are separated from the structural aspect because Fairclough believes that it is questionable to assume, as Foucault appears to do, that one can extrapolate from structure to practice. Taking the implied Foucault position would require no separation of structure from practice because one could infer the latter from the former. This criticism is further supported by Elliott and Wattanasuwan (1998) who found that lived meaning of brandsBor as Fairclough put it above, doing thingsBwas a potent source in the construction of identity. This introduces the important reciprocity of structure and action. Furthermore, because it is only in behavior that we interact with others, this is the way that representations of social structure are formed and so to work in only one way ignores the reciprocity between practice and structure. The structural dimension is then the knowledge and beliefs, social relations and social identities that are formed by social practice but which also guide and constrain social practice (see Giddens, 1984). Analysis in the structural dimension involves the analysis of social representations. Analyzing the textual dimension requires an examination of the way in which the knowledge and beliefs, social relations and social identities are expressed and negotiated in the texts and is in part a search of the underlying social representations that guide text production and behavior and a search for evidence of uncertainty of or struggle over their form.

When analyzing advertising texts, Fairclough uses the content, relations and subjects analytical framework to examine the ways in which advertising texts, in the case of contents, draw upon ideological assumptions of the world to build a picture of the product being advertised; in the case of relations, to examine the relationship built between the producer of the text and its audience (see also Scott, 1994); and, in the case of subjects, the way in which consumers are constructed as part of consumption communities. Similar approaches could be used for the analysis of different types of commercial interaction. Telesales and consumer response lines are particularly obvious examples.

At the heart of the analysis is the notion that there exist both implied authors and implied readers of texts, a position adopted in the marketing literature. In the marketing literature though, instead of the terms implied author and implied reader, the terms persona and implied consumer are used (Stern, 1994), though, in the case of advertising, ideal consumer would seem to be an even better description of the implied consumer. This implied authorBideal consumer view of text creation and interpretation is consistent with the theory of social representations if particular contexts or texts cue particular sets of social representations.

The implied consumer is the imaginary recipient whom the persona addresses. The implied consumer has beliefs about the world, a social identity and relationships with the persona and others. These beliefs, identities and relationships must be adopted to some degree in order for the text to be successfully reconstructed; that is to say, particular sets of social representations need to be activated. There is a degree of philosophical disagreement in the literature regarding the extent to which the reader is created by the text. The marketing literature errs on the less structuralist side with the view that the reader must play the role in which the author has cast him s it is only by agreeing to play the role of this created audience for the duration of his/her reading can a actual reader correctly understand the work (see Stern, 1994). In this view, readers are expected to play the assigned within-text role, as does an appreciative audience viewing a play or film. In the discourse analysis literature some authors (e.g. Fairclough, 1992) go further, arguing in fact that the audience is partly created by the text, rather than just agreeing, or not, to play the role.


Whether or not the reader merely agrees to play the assigned role or is created by the text, what is important is that, as readers are created or adopt the position of implied reader, subtle cognitive effects occur, changing the way that knowledge is linked within readers’ minds and altering the way in which they perceive the world about them. When viewed from this perspective, the degree of ease with which the actual reader can take the position of the implied reader, and the extent to which they are willing, depends on the initial cueing and internal consistency of the text. We would expect that texts cue particular implied readers just as phenomena cue particular schemata, frames or categories. Similarly, texts should evidence particular contents, relations and identities created or adopted in text production. For example, different greetings by in-store personnel should establish different relations and identities for the consumer and employee, resulting in different discourse, questions and frames of reference. These can be used not only to help explain buyer behavior but, perhaps, also to create a more pliable consumer. Fairclough (1992) provides interesting examples of the way that business can adapt and modify existing discourse types to bring about new ones with a more instrumental purpose. Fairclough calls this process of actively seeking to change the structure of discourseBand hence perceptions of contents as well as identities and relationsBthe "technologization of discourse." Within the social representations framework we might consider the technologization of discourse as a process to develop texts and situations that cue particular sets of social representations which, in turn, incline consumers to interpret phenomena and behave in particular ways.

Of course, attempts to induce consumers to adopt particular sets of social representations may not be met without resistance. Should there be too much of a difference between the set of social representations active when a consumer comes to a text and the set needed to interpret satisfactorily the text a tension will develop. We call this the tension between the self and the textually constructed self.


In this section, the theories discussed above are integrated into a single conceptual model. Figure 1: "A Dynamic, Contextual Socio-Semiotic Model of the Consumer" is a graphical representation of the model. The model has five basic objects and four processes. The five objects are social practice, the location of text production, the location of text interpretation, unstated assumptions and social representations. These objects are placed on the figure running from the outside inward representing the move from the most social objects to the most individualBthough all have both individual and social aspects. The four processes are anchoring, objectifying, the resolution of a dialectic tension between the self and the socially constructed self and hegemonic struggle. The processes of anchoring and objectifying are the most cognitive (individual) processes. The dialectic tension and its resolution are a more social process and hegemonic struggle more social still.

Social represetations were introduced in the discussion above as cognitive elements that provide models for categorization and evaluation. Social representations are formed by the two processes of anchoring and objectifying, which were discussed in some depth in the section "Social Representations". Because most categorization appears to take place subconsciously by some sort of gestalt processingBat least with familiar objects (Churchland, 1995; Sujan, 1985; Sujan & Dekleva, 1987)Bthe nature and relationship between social representations leads to automatic categorization and evaluation of phenomena in particular ways. Because this process is automaticBand very similar amongst individuals sharing representational structuresBthe results of the categorization and evaluation manifest themselves as unstated assumptions.

Social representations are not, however, linked together in fixed ways. Social representations appear to be linked together by themes, which might be called cognitive models or cultural models. Different themes contain different sets of social representationsBwith some social representations being part of more than one theme. This view explains our ability to perceive different meanings for the same phenomena in different contexts (Ungerer & Schmid, 1996) and applies just as much to the interpretation of texts as it does to the perception of different phenomena. Some things are more likely to be said in certain media and in certain contexts than others because different contexts activate different sets of social representations. It follows then that certain symbolic phenomena (texts) will be interpreted in different ways in different contextsBthe interpretation of advertisements provides a useful example of the importance of this context of text effect (see, for example, Wooten & Galvin, 1993). The set of social representations active in the context of text production and the set active in the context of text interpretation may be different. The dialectic tension between self and textually constructed self arises because readers must make shifts between their currently active set of social representations and the set needed to interpret the text. If each individual has a 'real’ self, this implies a set of representations that are more likely to be active and connected. The part transition from the currently active set of social representations (which might wholly or partly constitute an individuals 'real’ self) to the set required for the interpretation of a text therefore results in the activation of a set of social representations that constitute a self, part 'real’ and part the ideal consumer constructed by the text. Miscommunication occurs if this transition to 'ideal’ consumer does not go far enough to activate the complete set of social representations that are required for successful communication.



Social practice is the last object in the model. Social practice is treated as on object rather than a process because we can only observe each instance of behavior as an artifact of cognitive, social and physical structures. Social practice is the textually mediated behavioral manifestation of social representations and also the basis for their creation because it is only in social practice that representations become shared social representations. The relationship between social practice and the structure of social representations is also the ground for a hegemonic struggle for consumer mind-share.

Hegemonic struggle is the last of the processes included in the model. Hegemonic struggle is the battle to lead consumers to particular behavior. This battle must be textually mediated for two reasons. First, most of the battle takes part in language, through advertising, the sales process, media reporting, and word-of-mouth communication. Second, because of the intertwined nature of representations and language, all experiences are textually mediated even more the case as symbolic consumption becomes more pervasive. This battle for consumer mind share takes place both in language and over language. The battle in language is related to the first aspect of textual mediation of experieces as language is used as a tool in a given socio-cognitive environment. That is, texts are produced to be interpreted in given contexts with a particular set of unstated assumptions. The battle over language attempts to change the nature of ideology and tacit knowledge to induce different behavior. The struggle is shown as a process linking social practices and ideology because this accords more closely with the existing literature (see Fairclough, 1992) though, of course, any change to the nature of unstated assumptions necessarily requires a change in the nature of, and relationship between, social representations. There is, however, a reciprocity in this process because although certain sets of active social representations suggest certain behavioral outcomes, behavior plays an important part in the creation of social representations and hence the naturalization or revelation of unstated assumptions.


The model described above suggests both generic research questions and approaches as well as application specific questions and approaches. The generic research questions could be classified into static and dynamic. Static questions include identifying the nature of particular social representations. In the case of the marketing of high technology, we might ask about representations of the information super-highway, and the iconic qualities of generic terms such as 'PC’ and 'server’. These questions should then aid the design process. Also it is appropriate to identify the context specific manifestation of unstated assumptionsBfor example, those evidenced in consumers’ selection processes and the relationship between them and the attributes that are automatically included in evaluation exercises.

Dynamic questions examine process rather than content. The four processes of the model are all candidates for research. The twin processes of anchoring and objectifying present opportunities for examining how social representations are created. These should provide insight into the nature of information links within memory and, in particular, the structure of categories. An especially useful application of this is the analysis of new brands and how they are integrated with pre-existing social representations. The dialectic between self and textually constructed self is also an important process for study. In particular the way the viewers of advertisements and participants in sales negotiations are constructed and construct themselves in the various roles. This is instructive with respect to both those roles and the power relations created by them. Furthermore, the nature of the struggle to accept or fight against the textually constructed self may suggest ways in which that process can be used to alter outcomes. Finally, the dynamic between social practice and unstated assumptions and the suggested field of hegemonic struggle, is perhaps the most interesting dynamic for study as it is perhaps the one where the stakes are highest. In terms of marketing of high-technology, the way that common sense and social practice exist in reciprocity suggests an explanation for path dependence and a more than simply financial explanation for the development of platform lock-in and monopolies.

We can identify three important aspects to research involving the model introduced above. First, an identification of the social representations implicit in any textual interchange. This involves an analysis of both the linguistic and iconic aspects of social representation and the way in which they are context-dependently interlinked. The reciprocity between structure and action is also of particular importance. The critical approach to the study of language requires an examination of the social structures that constrain social action. The example used above involved an examination of how the social perception of the nature of work affects the implementation of new technology and how the implementation of new technology afects social perceptions of the nature of work. Finally, there is an important role for the socially constructed self and a realization that humans as social beings can be highly sensitive and adaptive to different social situations. The result of this understanding is an approach that leans toward continuous theory building and an avoidance of claims of generalizability.


This discussion has attempted to place consumers in a dynamic conceptual model that links congnitions, language and behavior in social context. Tools that might be used to analyze consumers within the theoretical framework were introduced. An important part of this model is the conceptualization of the consumer as an individual who is simultaneously created by and creating their world. Furthermore, most of this circular process occurs inBand is mediated throughBlanguage. The model also permits examination of firms seeking to develop and exercise market power. Firms are seen as not only engaged in an hegemonic struggle for consumer mind share that takes place in language but also in an hegemonic struggle over language. Techniques of critical discourse analysis were posited as appropriate tools to develop to operationalize the model.

We are still some way off a critical theory of the consumer that includes an appropriately significant role for language. However, the theory and examples outlined above serve to provide a useful basis for its continued development. Useful insight and development should stem from the work of Giddens and other sociologists who examine the internal / external dialectic and also from specialists in linguistics, especially in grammar, such as Haliday. Particularly important is to build the theory to include a role for competition. This we covered briefly above where the idea that companies competed in a hegemonic struggle in and over language for mind-share of the consumer. Finally, perhaps the most important aspect is to put the tool into practice to determine its uses and usefulness.


Brown, Gillian and George Yule (1983), Discourse Analysis. Cambridge, UK: Cambridge University Press.

Chomsky, Noam (1972), Studies on Semantics in Generative Grammar. The Hague: Mouton.

Churchland, Paul M. (1995), The Engine of Reason, the Seat of the Soul: A Philosophical Journey into the Brain. Cambridge, Mass: MIT Press.

Deutscher, Irwin (1984), "Choosing Ancestors: Some Consequences of the Selection from Intellectual Traditions," In Social Representations, Robert M. Farr & Serge Moscovici, eds. Cambridge, UK: Cambridge University Press, 71-100.

Elliott, Richard & Kritsadarat Wattanasuwan (1998) "Brands as Symbolic Resources for the Construction of Identity," International Journal of Advertising, 17, 131-144.

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

Farr, Robert M. (1981), "On the Nature of Human Nature and the Science of Behaviour," In Indigenous Psychologies: The Anthropology of the Self, P. Heelas & A. Lock, eds. London: UK: Academic Press, 303-317.

Farr, Robert M. (1984), "Social Representations: Their Role in the Design and Execution of Laboratory Experiments," In Social Representations, Robert M. Farr & Serge Moscovici, eds. Cambridge, UK: Cambridge University Press, 125-148.

Farr, Robert M (1987), "Social Representations: A French Tradition of Research," Journal for the Theory of Social Behaviour, 17 (December), 343-369.

Fowler, Roger (1996), "On Critical Linguistics," In Texts and Practices, Carmen Rosa Caldas-Coulthard & Malcolm Coulthard, eds. London, UK: Routledge, 3-14.

Galam, Serge and Serge Moscovici (1991), "Towards a Theory of Collective Phenomena: Consensus and Attitude Changes in Groups," European Journal of Social Psychology, 21 (January/February), 49-74.

Giddens, Anthony (1984), The Constitution of Society: Outline of the Theory of Structuration. Cambridge, UK: Polity Press.

Gramsci, Antonio (1992), The Prison Notebooks, Volume1. New York, NY: Columbia University Press.

Gramsci, Antonio (1996), The Prison Notebooks, Volume2. New York, NY: Columbia University Press.

Halliday, M. A. K. (1994), An Introduction to Functional Grammar: Second Edition. London, UK: Arnold.

Hodge, Robert and Gunther Kress (1993), Language as Ideology, Second Edition. London, UK: Routledge.

Jaspars, J. and C. Fraser (1984), "Attitudes and Social Representatinos," In Social Representations, Robert M. Farr & Serge Moscovici, eds. Cambridge, UK: Cambridge University Press, 101-124.

Kuhn, Thomas (1970), The Structure of Scientific Revolutions, Second Edition. Chicago, Ill: Univresity of Chicago Press.

Moscovici, Serge (1963), "Attitudes and Opinions," Annual Review of Psychology, 14, 231-260.

Moscovici, Serge (1984), "The Phenomenon of Social Representations," In Social Representations, Robert M. Farr & Serge Moscovici, eds. Cambridge, UK: Cambridge University Press, 3-69.

Pettit, Philip (1995), "Methodological Holism and Individualism," In The Oxford Companion to Philosophy, Ted Honderich, ed. Oxford, UK: Oxford University Press, 564-565.

Quine, Willard van Orman (1960), Word and Object. Cambridge, Mass: The MIT Press.

Ritson, Mark, Richard Elliott and Sue Eccles (1996), "Reframing Ikea: Commodity-Signs, Consumer Creativity and the Social/Self Dialectic," Advances in Consumer Research, 23, 127-131.

Rogers, E. M. and F. F. Shoemaker (1971), Communication of Innovations: A Cross-Cultural Approach, 2nd Edition. New York, NY: The Free Press.

Rommetveit, Ragnar (1984), "The Role of Language in the Creation and Transmission of Social Representations," In Social Representations, Robert M. Farr & Serge Moscovici, eds. Cambridge, UK: Cambridge University Press, 331-359.

Runkel, P. J. (1956), "Cognitive Similarity in Facilitating Communication," Sociometry, 19, 178-191.

Scott, Linda M. (1994), "The Bridge from Text to Mind: Adapting Reader-Response Theory to Consumer Research,"Journal of Consumer Research, 21 (December), 461-480.

Stern, Barbara B. (1994), "A Revised Communication Model for Advertising: Multiple Dimensions of the Source, the Message, and the Recipient," Journal of Advertising, XXIII (June), 5-15.

Sujan, Mita (1985), "Consumer Knowledge: Effects on Evaluation Strategies Mediating Consumrr Judgments," Journal of Consumer Research, 12 (June), 31-46.

Sujan, Mita and Christine Dekleva (1987), "Product Categorisation and Inference Making: Some Implications for Comparative Advertising," Journal of Consumer Research, 14 (December), 372-378.

Thomas, W. I. and F. Znaniecki (1918-1920), The Polish Peasant in Europe and America. Boston, Mass: Badger.

Thompson, Craig J. (1997), "Interpreting Consumers: A Hermeneutical Framework for Deriving Marketing Insights from the Texts of Consumers’ Consumption Stories,"Journal of Marketing Research, XXXIV (November), 438-455.

Thompson, John B. (199), Ideology and Modern Culture. Cambridge, UK: Polity Press.

Ungerer, F. and H.-J. Schmid (1996), An Introduction to Cognitive Linguistics. Harlow, UK: Addison Wesley Longman.

van Dijk, Teun A. (Forthcoming), "Cognitive Context Models and Discourse," In Cognition and Consciousness, Maxm Stamenov, ed. Amsterdam: Benjamins, .

Wooten, David B. and Tiffany Galvin (1993), "A Preliminary Examination of the Effect of Context-Induced Felt Ethnicity on Advertising Effectiveness," In Advances in Consumer Research: Volume XX, Leigh McAlister & Michael L. Rothschild, eds. Provo, UT: Association for Consumer Research, 253-256.