1. Setting up the stage

It is well known that social psychological thinking emerged and developed, with an impressive strength, in several European countries (Austria, France, Germany, Italy, United Kingdom...) from the middle of 19th century until the first years of 20th century. Strangely enough a long silence followed this highly productive period, interrupted only by a few exciting but lonely voices like Bartlett's (Bartlett, 1932) or Vygotsky's (Vygotsky, 1931). It was not until the end of the second world war that Social Psychology begun to spread again through Europe, directly imported from the United States, as if it had been one more of the Plan Marshall's pre-packed stuffs sent to this devastated continent. Logically, the imported Social Psychology came from U.S. mainstream orientations with all their neopositivistic and individualistic flavour. Needless to say that this kind of Social Psychology is still strongly influential in Europe, but since the early sixties until nowadays some European original contributions to the discipline have opened a new line of development. Roughly speaking this new line can be defined by a stronger concern with the "social dimension" of social psychological phenomena, and by a critical position towards neopositivistic trends in research methodologies (Israel & Tajfel, 1972). Some of the most outstanding European contributions are undoubtedly the Theory of Social Representations (Moscovici, 1961; Jodelet, 1989), the Theory of Conversion Behaviour (Moscovici, 1980; Moscovici & Mugny, 1987), the Social Categorisation Theory and Social Identity Theory (Tajfel, 1981; Turner, 1987), and the impressive development, mainly in United Kingdom, of the Discourse Analysis orientation (Potter & Wetherell, 1987; Parker, 1992), the focus on Rhetoric (Billig, 1987), and the concern with Everyday Explanations (Antaki, 1981).

I will deal here with one of the most influential of these European contributions namely the Theory of Social Representations. The formulation of this theory has been strongly beneficial for Social Psychology, widening its perspectives and emphasizing its social dimension (Ibáñez, 1990). No doubt that the "research program" (Lakatos, 1978) constituted around Moscovici's theoretical work is far from being exhausted and will continue to be "progressive" during some more years. But, to acknowledge the virtues of a theory does not exempt from keeping a critical look upon it. This is precisely what has been done by a number of colleagues belonging to different theoretical orientations. Some of them have stressed the vagueness and the ambiguities of the theory (Jahoda, 1988). Others have questioned the extent to which the term "social" refers effectively to a social dimension instead of a mere individual attribute (Harré, 1984). The main corpus of critical writings stems  from researchers who are working in the Discourse Analysis and the Rhetorical Orientations (Potter and Litton, 1985; Potter and Wetherell, 1987; McKinlay and Potter, 1987; Parker, 1987; Billig, 1988). These authors argue that the Theory of Social Representations is not so much opposed to cognitivist trends as it pretends to be and that it lays upon a set of cognitivist assumptions and concepts which explain why the theory slides so easily towards cognitive reductionism. My own criticisms go in the same direction, but two preliminary remarks may help to grasp better my argument.

The first is related to the proper heading of this paper: "Constructing a Representation or Representing a Construction?" and to what may appear as a strange paradox when we consider the responses which may be given to the interrogation it conveys.

Serge Moscovici is not categorized usually as a straight constructionist, he has even formulated some reservations about constructionism, and nonetheless I bet that his response would be: both. We both construct representations and we represent constructions. That seems to reflect indeed a strong constructionist commitment...

I am strongly influenced by the social constructionist turn (Gergen, 1982, 1985) and nonetheless my response would be: neither. We neither construct representations nor do we represent constructions. Of course I will try to show that this paradox is only apparent and my argument will be based on the assumption that ordinary people and Moscovici are fairly similar in their way of making sense of reality: neither of them engages in the construction of representations.

The second preliminary remark is about social sciences in general, social psychology in particular, and the theory of social representations more specifically.

In so far as Social Sciences are articulated through social processes, are embedded in social networks, and are, in short, social objects, it is hardly surprising that some of the theories produced by these sciences to account for social phenomena can also be used to account for some aspects of the proper sciences which have generated the theories. In other words, they are sciences which produce theories which can explain some aspects of what has produced them. Some reflexivity is here at stake, and this is for sure an interesting characteristic of these sciences.

But the case of Social Psychology is even more interesting.  Effectively, it is a discipline where some theories are able to generate the proper phenomena they pretend to account for.  Maybe we could speak, in this case, of "Pigmalion Theories", of "Self-Fulfilling Theories" or even of "Self-Validating Theories".  Anyway, there is no doubt that Serge Moscovici's Theory of Social Representations is an impressive example of these kinds of theories.  If we have a look at contemporary European Social Psychology we will grasp immediately that one of its current characteristics is the existence of a quite large group of social psychologists (... which is recognizable as a group precisely because its members share a similar representation...!) which have a representation of society as being constituted by a dynamic and constant flow of Social Representations... and this is an effect due to Moscovici's creation of the Theory of Social Representations...!

I am not just doing an easy bad joke about Social Representations; I am convinced, indeed, that it would be quite easy to use the proper terms of the theory to show, with some precision, how the Theory of Social Representations has spread in a community by communicating a set of images and concepts which are used to give sense to social reality, or to show how this theory has been turned familiar by being anchored in former prevailing conceptions; or how some of its abstract theoretical entities have been objectified by some of the followers of Moscovici; or even to show how emotional factors are embedded in the Social Representation of society which has been generated by the Theory of Social Representations..., and so on.  But I am not going to give evidences about all that here because that would imply that I assume the explicative relevance of the Theory of Social Representations. Instead of that, I will try to use reflexively the Theory of Social Representations to illustrate what seems to me some of its shortcomings. This is not an easy task because the theory has developed an efficient "protective belt" against criticism. Curiously enough, "La Psychanalyse: son image et son public" (Moscovici, 1961 -and 1976 for the revised second edition) remains the only book written by Serge Moscovici on Social Representations, but since 1961 Moscovici has published a considerable amount of articles, book chapters, responses to criticisms, comments to empirical works done by his colleagues, textbook chapters, conference papers...The result is that so many debates and theoretical reflections have been devoted during the past decades to Social Representations that different versions of the Theory can be found depending on which set of texts is taken into consideration. In this situation any criticism which seems to do justice to one version of the Theory has a high probability of being dismissed when other versions of the Theory are called to the arena. This is for sure a strong handicap for formulating any criticism at all. So I will not concentrate on any specific version of the Theory but on two general statements: first, the asertion that the Theory of Social Representations deals especifically  with Social Thinking (... on the understanding that this expression does not imply the existance of a "non social" type of thinking); second, the claim  contemporary social thinking is constituted by some kind of representational stuff. These two points are present in all the versions of the Theory insofar as they constitute the very basements of the Theory itself. A third point is present in all the attempts done by Moscovici to separate clearly the construct of "representation" from the construct of "attitude". Attitudes mediate responses to stimuli while Social Representations act also on the encoding process, that is, they shape the input, not only the output when people deal with social objects.  If this is point is taken seriously (this is not the case in all the versions of the Theory) then it injects into the Theory a strange "constructive loop" which I will refer to when starting the critical analysis of this Theory.

2. The implications of the "constructive loop" asserted by the Theory of Social Representations"

What exactly did Moscovici do when he constructed his theory through theoretical work and empirical research?  Did he draw a more or less accurate scientific representation of a social phenomenon called "Social Representation"?  Obviously enough the answer is "no".  But to make clear the implications of this answer let me sketch two short stories, or two brief narrative accounts, about the way the Theory of Social Representations came to life. The first story could be entitled: "The discovery of Social Representations". The second account could be entitled: "The Invention of Social Representations". Both accounts are reflexive in the sense that they are constructed from inside the proper theoretical assumptions of the theory they pretend to account for.

The first narrative, "The discovery of Social Representations" could run something like this:

Insofar as the Theory of Social Representations did not exist, obviously enough, before Moscovici formulated it, this theory can be dated with precision.  But if we give credit to this theory then it appears that the objects it accounts for can also be dated with more or less precision.  Effectively, Moscovici tell us that a Society like ours which is submitted to an accelerated process of change and which is pervaded by scientific productions, gives rise to Social Representations, instead of myths, for instance, as an instrument used to make sense of reality as well as to shape reality.  If we agree with that then we must admit that all along the 20th century, and probably a good deal of the 19th century, Social Representations have been working inside our society even if nobody was aware of it, and it was not until the late fifties when Moscovici was ingenious and lucky enough to construct the adequate lens, that is, the adequate set of theoretical tools and methodological procedures, to scrutinize successfully his own society and to grasp Social Representations at work.  Before this "discovery", the object existed but not the theory which was required to make us aware of its existence.  Since then, a considerable amount of research has been devoted to draw with more and more precision the theoretical representation of Social Representations in such a way that we are currently very close to reaching a good correspondence between the two.

I am sure that not only Moscovici but any sensible person will consider ridiculous this first narrative.  In spite of that, this narrative seems to be assumed implicitly by some social psychologists and it is producing some unfortunate effects in European Social Psychology.  Moscovici has explained very clearly how abstract entities, or theoretical entities, are reified through a process of objectivation, and this is precisely what is happening in ongoing researches about Social Representations. It is as if the growing success of the theory had launched out many social psychologists on the track of Social Representations, rivalising between them in the art of setting up the most ingenious traps to capture some fine exemplars of Social Representations.  Some of the dialogues I have heard need no more than a slight distortion to sound like that:

- "Look, what a nice exemplar of social representation of AIDS I have got"

- "Well, It is quite interesting but it is not such a beautiful piece as my social representation of Europe"

Leaving the jokes aside, the point I want to make is simply that in the process of becoming more and more descriptive instead of explicative the research about social representations is loosing value at the same rate that it is increasing its holdings in the market of empirical research.  Of course Moscovici can not be blamed for that even if one can wonder if he could not have pushed more firmly the foot-brake to slow down this descriptive slope.

Anyway, as the first narrative is clearly unacceptable let me sketch the second narrative which is much more close to Moscovici's own claims about the ontological status of Social Representations.  In the same way as the first narrative was grounded on Moscovici's assertion that Social Representations are actually constitutive of modern societies' social thinking, this second narrative starts from Moscovici's affirmation that social representations are constructions, but constructions in which the represented object is also a construction.  In other words, here there is an interesting constructive loop in so far as in the process whereby it has been constructed a social representation constructs that which it is a construction of; it is a construction which constructs the thing of which it is a construction.  I think we have to take very seriously this constructive loop and we must congratulate Moscovici for being so clearly a constructionist when he explains how ordinary people makes sense of reality. But does Moscovici, as a social scientist, behave in a very different way from ordinary people? Why should we limit constructionism to just one kind of social thinking, the one manifested by ordinary people? Why not be constructionist for other types of social thinking, like the one which is developed by social scientists when they build their theories? If so we do, in other words if we do not stop halfway in adopting a constructionist standpoint, then we must conclude that social representations have been constructed in the process itself of constructing their theory. The problem is that if the constructionist assumptions of the Theory of Social Representations are accepted as valid and extended to the whole realm of social thinking, including scientific thinking, then the conclusions of the theory are necessarily false: there are no social representations in our societies. If I referred previously to the Theory of Social Representations as a "self validating theory" it seems that I should better speak now of a "self refuting theory".

But let me illustrate this point with the second narrative entitled: "The invention of Social Representations". This story could run something like this:

Through an "incessant babble" in the market places, cafes, scientific meetings, labs, and, above all, by conversing through their writings with germinal thinkers, like Durhkeim, Moscovici constructed, in the same movement, both an object which he called "Social Representations" and a theory about this object.  If we take this for granted then there is no reason to endow social representations with any ontological status which could be independent of the theory which created them.  There is no reason to think that the theory gave theoretical existence to objects which enjoyed actual existence, out-there, in society.  Both the object and the theory about it were literally created by Moscovici in the late fifties with a clear objective: to make sense of the way social thinking operates in contemporary societies. Since Moscovici's invention of both social representations and the Theory of Social Representations a considerable amount of work has been devoted to enlarging the elucidatory power of the theory in such a way that this theory is far more convincing today than it was thirty years ago.

If we take seriously this second narrative, which follows closely the constructive loop outlined by the Theory of Social Representations, it appears that in the process of making his theory Moscovici did not try to "represent" anything "out-there", and that the result of his theoretical work is by no means something which depicts some other thing.  What Moscovici did is no less and no more than to suggest a quite credible account about social thinking in our societies.  His theory is a well articulated discursive production which argues quite convincingly for a particular way of explaining social reality.

To this point, I think an important question can be asked: do ordinary people behave in a very different way than Moscovici did when they try to make sense of reality in their everyday life?  I am afraid they do not!  They, no more than Moscovici, do not construct "mental representations of things".  What they do is simply to construct things which are of a discursive nature in so far as they are objects of talk, thinking and communication. We cannot grasp whatever we call "a thing" and communicate about it without constructing it through the categories of our language, and what results from this construction is a discursive production; never a mental representation. In short, people do not live in a world of "representations"; they live in a world of "discursive productions" and that makes a great deal of difference.  One of the most stricking differences is that when the concept of "social representations" is substituted by the concept of "discursive productions" then the constructed nature of social world and of social beings themselves appears as an unavoidable feature. If people are seen as "discourse makers", or as "talking animals", it follows necesarily that they are at the same time "reality constructors" because "discourse" is precisely a practice which constructs sistematically the object of which it speaks (Foucault, 1972). The only way to deny this constructive character of people would be defining them as mere tape recorders which repeat mechanically what they have recorded. Moreover, the fact that people are engaged in a constant production of discourses about themselves and that they are inmersed in a host of discourses which speaks of them (directly, as the particular individual they are, or indirectly, as the type of being they are) suggests that they are themselves of a discursive nature and that they are literally constructed and self-constructed.

Notwithstanding, the drift towards discoursivity is far from being unproblematic. The discourse analysis trends which are developing currently in social sciences, and specially in social psychology, are full of controversies about the adequate procedures to conduct inquires, about what has to count properly as "discourse", and about the links between discourse and reality (Parker, 1990; Potter, Wetherell, Gill and Edwards, 1990; Edwards and Potter, 1992; Parker, 1992; Wetherell and Potter, 1992). This is not the place to deal with these controversies, or to develop in detail the conception of "discursivity" which stand back my claim about the discursive nature of our world, but a few general remarks will help to sketch an outlook of what I have in mind.

Many different traditions and theoretical works have contributed through history to emplasize the importance of language, but from mid century until now three major events have turned absolutely inscapable the concern with discursive practices. The first is Wittgenstein's reflection on language games and the linguistic boundaries of our world (Wittgenstein, 1953). The second is Gadamer's insistence on the hermeneutic nature of our way of "being in the world" (Gadamer, 1979). The third is Foucault's analysis of the role of discourses in the construction of such objects as madness or sexuality (Foucault, 1971, 1981). No doubt that other contributions have also been important bu these three have been decisive and they do not allow any other way out than to acknowledge that language (via discourses) is formative of reality.

There is now a long time ago since social sciences began to show a clear concern for language and symbolic productions. Ethnomethodologists renewed this interest in the last sixties, but it is not until the post-crisis developments that social psychology began to be really receptive to these questions. A first step in this direction was taken by Kenneth Gergen when he underlined the generative effects of knowledge (Gergen, 1982), a further step was done when social constructionists emphasized the role played by the vocabulary of psychology in the shaping of psychological attributes and stressed the role of language in the shaping of self and identity (Gergen and Davis, 1985; Harré, 1986; Shotter, 1984).

Currently great efforts are being devoted to analyse everyday explanations (Antaki, 1988), conversation processes, discourse and narratives (Gergen, 1988). All these efforts are contributing to place the "discursive dimension" as a fundamental feature of our reality and so doing they meet similar concerns in many other fields of knowledge (Post-Analitic  Philosophy, Post-structuralist trends in Anthropology, Sociology of Scientific Knowledge, and so on...)

All these developments make wiseful Moscovici's invitation to pay greater attention to language instead of cognitive processes. No doubt that we, as social psychologists, should give up the search for mental entities and mental formations in terms of images, categories, schematas, scripts, prototypes, etc; and we should concentrate upon discursivity.  This shift would enhance, among other things, the dynamic, constructive and active dimensions of social thinking.  This enhancing effect was sought by Moscovici himself when he opposed the nimble fluidity of Social Representations to the more static heaviness of ideologies.  We have now to accentuate this trend by opposing discursive productions to social representations.  It is because everyday discursive productions are both context-dependent and context-formative that the same person can make contradictory claims about the same issue depending upon with whom s/he is discussing it, and that this person can construct differently his/her position depending on the effects he/she is producing upon the discursive context (Billig, 1987).  This does not mean, of course, that a person is free to elaborate any possible standpoint, or that a person just reacts to a given context; there are constrictions which have prescriptive effects upon what can be said by somebody.  But it is clear that in comparison with the model of Social Representations, the model of Discursive Productions allows a wider range of possible standpoints for the person and, at the time, it emphasizes more strongly her role as an active agent.  In these aspects Discursive Productions stands to Social Representations in the same relation as the latter stand to ideologies: they are far more nimble and dynamic.

One of the reasons why the Theory of Social Representations restricts excessively the active role of agents, is because this theory engages in a constructionist direction but it stops halfway.  Effectively, this theory is concerned more directly with processes of re-construction, or transformation, than by processes of genuine construction.  Social Representations originate from pre-given objects which are already elaborated in particular discourses  - like scientific discourses or mediatic discourses -, and which have to be re-constructed in order to be assimilated.  A Social Representation is a re-representation of something which is already represented in other discourses.  It is an operation of transformation more than an operation of construction.  I am not suggesting that discursive productions are made "ex-nihil" or that they are not fed by pre-given information and on-going discourses, what I am saying is just that the underlying process is not one of "transformative appropriation" of that information and those discourses, it is a genuine process of construction which takes these elements as resources among other resources, like for instance human actions, social practices, colective memory and material productions. If the Theory of Social Representations fails to give a straight constructionist account of social life it may be because this theory is so deeply embeded in the prevailing "ideology of representation".

3. The Ideology of Representation

Some longstanding human practices, like drawing, painting, tracing maps, are prototipical of "figurative representation"; some other old human practices, like naming and counting, are prototipical of abstract "non-figurative" forms of representations.  But it was not until the invention of one of the most socially relevant technology, the technology of printing, that the concept of representation pervaded fully all our culture.  Effectively, the possibility to copy an original without distortions, the capacity to make countless reproductions of the same thing, and the activity of spreading representations of objects, put the idea of representation at the center of the stage and contributed to shape the concept of objectivity which was equated with the idea of an accurate representation (Rorty, 1979).  Later on, some other technological devices, like photography, films, or even tape recorders, intensified even more the extent to which the concept of representation shaped contemporary ideology.  This concept had a strong impact in science, including of course the social sciences.  Even if the ideology of representation is probably in a process of being weakened by the emergence of information technologies, like computers and so on1, it is hardly questionable that it is still forceful in contemporary society and that it displays strong power effects through making people being highly dependents.  Effectively, if reality is taken as an object which can be adequately represented, then we must pay allegiance to the established methods to ensure an accurate representation and we must rely on the experts who are able to draw these representations.  I am afraid that the Theory of Social Representations contributes inadvertently to reinforce this ideology and the "regime of truth" which it conveys (Ibañez, 1991).  But before getting to such a conclusion let me dig a bit deeper into the concept of representation.

When we use the concept of "representation" we necessarily constitute, at the same time, the concept of a "represented object" which is, by definition, something different from its representation.  In other words, by the mere fact of using the term "representation", we necessarily postulate the existence of an independent "pre-represented" reality which will serve as the referent to the representation, no matter if this "pre-represented reality" is of a natural kind or if it is a socially constructed object.  In short if we have "representation" at one hand we have also unavoidably what it is a "representation of" at the other hand.  The model which stands back to this conception is, obviously, the model of visual perception, but a model of perception which is quite naive and which pertains to the times when perception was defined as a kind of image which reflected reality like a mirror. The fact is that the concept of representation creates a straight duality between things and their image, between things and what stands for them.  As soon as this duality is created a set of problems emerges immediately which have afflicted philosophy for many centuries.  That is quite understandable because once duality is created you have to articulate immediately the means to transcend this duality, you have to construct a bridge between the two separate realms which have been established and you have to account for the strange road which leads from the object to its representation.

One form of resolving the problems created by duality is to state that all which counts for us are representations; that they constitute our actual world; and that we can bracket the question of the represented objects as long as what produces real effects upon us are their representation.  This solution, which has a clear Kantian and phenomenological flavour, leads nowhere because the proper use of the term "representation" forces upon to save the referent of representation even if it has been put into brackets.   In the case of the Theory of Social Representations this operation is realized in two ways.  First a "reified reality" is postulated and it is constituted as an object assigned to Science.  Second, the represented reality is put in the place of the reality which is represented, but this is done in a straight realistic way.  That means that, once constituted, the represented reality gets reified and constrains us in a prescriptive way as would do pre-represented reality itself.

It could be argued that there is no sense in making this type of argument insofar as thinking is a process which manipulates symbols and that symbols are precisely things which stands for other things; so to put representations in question is to put in question the process of thinking itself.  This claim would be convincing if Wittgenstein had not made us sensible to the fact that human thinking does not use symbols as elements which represents other things, and that words are not labels hanged by conventions to the referents they represent (Wittgenstein, 1953).  A word is not a label whose sense could be specified by pointing to its referent.  The sense of words is not any other thing than its use, and we have to search for uses instead of searching for referents if we want to grasp the process of thinking.  Effectively, all we have are not sets of things which can be represented by words, but "language games" which constructs, literally, what we objectify afterwards as pre-given things.  If we follow the assumption of Wittgenstein it results that we are far more active than the theory of Social Representations allow us to be because we are not active only in constructing the representation of things; we are also active to the extent that we construct things themselves through, and inside, our language games.

This is precisely what the ideology of representation prevent us to grasp.  This ideology masks the fact that our thinking activity does not work so has to represent anything at all; and that we neither construct representations nor represent constructions but construct the objects which constitutes our reality.  Of course, that is not to say that we do not construct images or that imagination is not one of our defining features as human beings, it is just that the relationship between these images and reality is not of a "representational" nature but of a constructive one, in both directions.  The only way to avoid the problems created by the duality between representation and reality is just not to create this duality, and that implies to be strongly critical of the idea of representation itself. Obviously, this process of construction does not go on "inside our heads", it is not just an operation of thought, it is embeded in our actions and our productions.

4. The split between "person as a thinker» and "person as a doer"

The second narrative I sketched a moment ago, in point 2, is questionable in many aspects, but one of its main failures is that it forgets absolutely that Moscovici did not construct the concept of Social Representations and the Theory of Social Representations just by thinking about it and conversing with others.  His thought was shaped by a whole set of practical activities he displayed alongside thinking: he wrote, he quoted books and articles, he constructed questionnaires, he quoted newspapers, he did content analysis of interviews, maybe he draw figurative schemes or diagrams and so on.  Without all these practical activities he could not have developed the kind of thinking which led him to the formulation of his theory.  The idea that thinking is closely linked to the practical activities and to the material productions issued by human beings is not a new one. Marx spoke of human beings as "tools makers", the anthropological theories about pre-historical humans stressed the role played by tools in the constitution of thought, and piagetian developmental psychology has focused on the role of concrete operations at a sensori-motor level on the development of abstract and formal operations. More recently, the accent has been put on the idea that technical devices are to be considered fully as  social operators (Roqueplo, 1983) and Bruno Latour is drawing upon Sociology of Scientific Knowledge to elaborate a provocative theory which ties toguether human beings and technological objects in a new kind of "hybrid objects" (Latour, 1991).

No doubt that Moscovici is fully aware of the impossibility to display thinking in a "practical vacuum", but the systematical emphasis put on "social thinking", as such, accentuates the widespread tendency to hipostasiate "thinking" as an activity sharply separated from "doing", to split the so called "mental"  realm from the "practical" realm and to divorce the symbolic dimension from praxis.  In effect, it seems difficult not to be trapped in these pitfalls when Social Representations are seen as "mental formations" born in the "incessant babble" of conversations.

From my point of view there are two senses in which we can not separate "thinking"  from "doing".  The first relies on what humanity has done all along its history and the second points to what people do actually in their everyday life.

In the first sense it looks reasonable to say that "social thinking", or "thinking"tout court is rooted in the cultural magma constituted by thousands of years of cultural productions.  The incessant babble of society is grounded on the accumulation of symbolic productions which have been generated through history and which conform collective memories. These collective memories are nourished by discursive materials like proverbs, myths, songs, all kind of writings, and by the proper language we use; but they are also constituted by tools, weapons, paintings, craftsmanship, buildings, domestic animals, vegetables, and all the "savoir-faire" which are linked to these objects. It will be a great mistake to separate clearly our material culture and our discursive culture, to split the thought from the things.  There is a strong tendency to acknowledge that material productions done by societies are concretions of processes of thought, and to ignore at the same time that thinking itself is shaped by material productions. There is today a whole array of evidences which shows clearly that our ways of thinking are strongly dependent of what some have called "the technologies of intelligence" (Lévy, 1990).  The invention of writing, the invention of print, the invention of micro-computers are the most outstanding inventions in the field of intelligence technologies but there are also many minor inventions, like pencils, rubbers, blackboards, abacus and so on, which have been shaping the way we actually think.  In a word, we do not think with independence of a whole set of material artifacts which do not just "help" to think but which are constitutive or our thinking.  To present Social Representations as "mental formations" originated in communication, and to define them as "a kind of thinking" foster the dualist conception of "thinking" as an autonomous activity clearly separated from "doing". On the contrary, in so far as discursive productions are not a mental type of stuff, it seems that those who work with this concept instead of the concept of social representations have it easier to avoid the trap of this dualism.

The second sense in which we can handle the problem of dualism between thinking and doing is tied to what people do actually in their everyday life.  It is obvious that in their ordinary life people are engaged in a diversity of practical activities which influence their discursive productions as much as their conversations with others may do. If people who belong to different social statues, or people who are involved in different professions, make different discourses about significant aspects of social reality, it is not so much because an abstract world of ideas surrounds each social position and each profession, it is mainly because the practical activities in which they are engaged as an upper class person versus a lower class one, or as an university professor versus an industrial worker, lays deep traces in their discourses.  Discursive productions are not generated only by mental processes and conceptual activity, they come also from non mental practices, from the objects which are used in daily life and from the operations allowed by these objects.

Do not ask me how this practical world impinges on our discursive productions.  This is precisely the kind of question which has not been studied yet by a social psychology which has assumed too eagerly the split between the mind and the world.

I am aware that a possible objection to my comments could take the following form: Well, we can accept that the type of "thinking" related to Social Representations is shaped by technologies, material surrounding, praxis and so on, but at the end what we get is the construction of mental formations that we have labelled as "Social Representations".  This is what is of concern for us, not the nature of thinking per se.  But, the fact is that this kind of statement does not free the Theory of Social Representations from the set of problems it faces, all the contrary, it leads to new difficulties. One of them is that this theory would restrict its scope to the mere description of these mental formations called "Social Representations" abandoning all attempt to elucidate their genesis. This would be hardly acceptable for all those who are convinced that social objects can not be understood independently of the processes which generate them.

5. Conclusions

Two important but hardly new ideas stand in the kernel of the Theory of Social Representations.

First, taking a strong position against the most shortminded behaviourist and positivist trends, this theory asserts that we do not respond, or react, to "reality as it is" but to our definition, or vision, of reality. This is indeed a very old idea splendidly coined by I.W.Thomas and D.S.Thomas at the beginning of this century when they wrote that a situation is effectively real if we think that it is (Thomas and Thomas, 1928), and further developed by symbolic interactionism, ethnomethodology and interpretative sociology. In accordance with this idea the Theory of Social Representations stresses the importance of knowing how people "see" their world if we are to elucidate (and eventually to predict) how they behave in it.

Second, taking a similar strong position against the most shortminded individualistic reductionism, this theory asserts that our definition of reality far from being an individual affair results from social commerce (the famous "inceasant babble" of everyday existence). We are again in front of a quite old idea which can be traced to the same traditions than the former one. In accordance with this idea the Theory looks for the formation of Social Representations in the conversational processes which mark social life.

There is no doubt that the positions against which the Theory of Social Representations constitutes itself arouse a first feeling of congeniality towards this theory. But this feeling is swept away when we take a closer look into the theory and when we begin to suspect that the positions explicitly opposed by the theory are covertly reintroduced in its basic assumptions.

The theory states that we react to our representation of the object, not to the object itself, but the theory states also that once a Social Representation is constituted it tends to constrain  the responses of all the members of the group who share this representation, just as if it was another brand of "pre-given stimuli". This tendency to reify representations should not be surprising if we take back the concept of Social Representation to its durkheimian origine and to the positivistic flavour which surrounds Durkheim's formulations.

The theory says that social Representations originate in social exchange, not in the head of individuals, but it is the heads of individuals where these representations are instantiated and where we have to look at if we want to find them (mainly via this kind of personal reports which are called "questionnaires"). This too is far from being surprising if we take in count the cognitivist climate which was flourishing when the Theory of Social Representations developed (Festinger's theories in the sixties and a more hard social cognitivism in the seventies).

It seems reasonable to conjecture that the surreptitious introduction of individualistic reductionism into the Theory of Social Representations is closely linked to its cognitivist assumptions. But cognitivism is, after all, a respectable position and if the conjecture was right then the theory would only deserve the usual criticisms that non-cognitivists (...belonging to non-behaviourist tradition...) devote to cognitivism. Nevertheless it is my contention that the case of Social Representations is more complex and that it deserves another type of criticism. To the extent that this theory plays the language game set up by the ideology of representation (Woolgar, 1989) and that it takes sense from inside this ideology as much as it contributes to foster its influence.

The concern with "how people see their world" leads to concentrate on the mechanisms activated by individuals to construct their image of the world. The contents of the image are enterely social since they are picked up from social discourses, the mechanisms which generate the image are partly social and partly cognitive, but the end product is enterely cognitive and the whole process rests heavly upon the metaphor of "vision" with the same distorting effects that this metaphor displays on dominant conceptions of scientific knowledge (Rorty, 1979). A sharp split is introduced between what "counts for reality" and what "reality effectively is", objects are seen as being relatively independents from subjects insofar as the only role attributed to subjects is restricted to the representational realm. Of course, subjects are fairly active in constructing their image of the world and this construction is grounded on social processes, but subjects are enterely passive with respect to the proper nature of objects. For sure, Moscovici has stated that objects themselves are constructed through representation but this assertion remains largely devoid of content because the Theory of Social Representations is unwillingly designed so as to bracket efficiently this contention, and it is the proper focus on "representation" which leads to this result. The notion of "representation" can not do any other thing than to drive the theory towards a more phenomenological type of theory than it pretends to be. The split between what "counts for reality" and what "reality effectively is" is internal to the notion of representation itself, and this split opens largely the doors to the intrusion of the concept of "objectivity". But if there is something like an "objective reality" independant from the subject (not in the individualistic sense but considered as a social entity), then twos consequencies follow. First, there must be means to accede to this reality (Science is a good candidate for that), and, second, there must be some representations of the same object which are more accurate than others. If there are means to accede to objective reality then those who master these means (specialists like scientists for instance) are endowed with the power of telling how this reality is, and those who do not master these means are compelled to submitt to these discourses. If representations differ in their accuracy, then we must promote a change towards the most accurate ones. This implies that those social scientists who are prone to emancipatory aims must work at the representational level and push people to get a new "vision" of things.

All these implications of the theory can be illustrated by a reference to racism. If we follow the Theory of Social Representations then we have to confront a reality composed of the existence of different races, and of relationships between different ethnic comunities. Racism appears as a consequence of some social representations linked to this reality, and the fight against racism can be best served by changing these representations. In other words, racial or ethnic differences are a feature of our world, and racism is a consequence of an insane representation of this feature. If we take some distance from the ideology of representation and adopt a more constructionist stance then we are to consider that we construct reality in terms of races, or ethnic differences, as a sensical feature of our world, and that it is this very construction which conveys the possibility of racism itself. In other words, there are not ethnic differences, by one hand, and our representation of them, by the other, but there are discursive practices which construct the world in terms of ethnic differences. Racism is not a question of how we "represent reality" but of how we construct reality. Change is not brought by acting at the phenomenological level of representations but at the ontological level (the type of reality we construct), and this points towards social practices more than towards representational outcomes. This move implies above all that we sweep away the split between what "counts for reality" and what "reality efectively is". The concern with "representation" is, I am afraid, to be seriously weaken by the denial of this split.

These concluding comments may appear as being too heavily pervaded by political considerations instead of academic ones. I acknowledge that this may be true, but this is internal to the proper position I have tried to develop, for I do not see how the excellence of a theory can be constructed from the standpoint of "value-free criteria" which would enable us to grasp things "as they are" or, in this case, theories "as they are".

Akcnowledgments
A shorter version of this paper was presented at the European Association of Experimental Social Psychology (EASP) and Society for Experimental Social Psychology (SESP) Joint Conference (Leuven; Belgium 14-18 July, 1992)
I would like to thank Charles Antaki for his helpful comments on an earlier draft, and Henderikus Stam for helping me through my difficulties with English.

1 The new tecnologies of intelligence, like computers for instance, are contributing decisively to change our assumptions about the links between knowledge and reality. In many domains the question is no longer to depict accurately reality or to check reality to see if it is in correspondance with our claims about it. The computer allow us to generate "virtual" realities out from a few parameters and it allows us to explore quite freely as many paths as possible in these "generated relities". We have not to wonder if our model is true, or if it accounts for reality, what is relevant is just if we can do useful things with the model, it does not "stands for" reality, it furnishes practical results. The best starting parameters are not those which lead to the most accurate image of reality, but those which generate more paths to be explored at the lower cost in time and complexity and which produce more practical results. Some of the values which have sustained during centuries the ideology of modernity lose gradually their relevance: objectivity, truth, correspondence with reality, the clear distinction between reality and its representation... One can wonder if information technologies are not precisely the kind of technological innovation which requires the end of the ideology of modernity and the construction of a post-modern ideology to be fully legitimized and to secure the changes it is promoting in our society.

Antaki,C. (Ed.) (1981). The Psychology of Ordinary Explanations of Social Behaviour. London: Academic Press.

Antaki,C. (Ed.) (1988). Analysing Everyday Explanation: A Casebook of Methods. London: Sage.

Bartlett,F. (1932). Remembering. Cambridge: Cambridge University Press.

Billig,M. (1987). Arguing and thinking. A rhetorical approach to Social Psychology. Cambridge: Cambridge University Press.

Billig,M. (1988). Social Representations, Objectivation and Anchoring:A Rhetoretical Analysis. Social Behaviour, 3: 1-16.

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

Foucault,M. (1971). Madness and Civilization. London: Tavistock.

Foucault,M. (1972). The Archeology of Knowledge. London: Tavistock.

Foucault,M. (1981). The History of Sexuality. Vol I. Harmondsworf: Pinguin.

Gadamer,H. (1979) Truth and Method. London: Sheed and Ward.

Gergen,K. (1982). Toward transformation in social knowledge. New York: Springer Verlag.

Gergen,K. (1985). The Social Constructionist Mouvement in Modern Psychology. American Psychologist 40(3): 266-275.

Gergen,K. & Davis,K.E (Eds.) (1985). The Social Construction of the person. New York: Springer Verlag.

Gergen,M. (1988). Narrative Structures in Social Explanation. In C. Antaki (Ed.). Analysing Everyday Explanation: A Casebook of Methods. London: Sage.94-112.

Harre,R. (1986). Some Reflections on the Concept of "Social Representations". Social Research. 51: 927-938.

Harre,R. (1986). Selves in talk. British Journal of Social Psychology. 25: 271-273.

Ibáñez,T. (1990). Henry, Serge...  and the next generation. The British Psychological Society-Social Psychology Section Newsletter. 24: 5-14.

Ibáñez,T. (1991). Social Psychology and the Rhetoric of Truth. Theory and Psychology, 1(2): 187-201.

Israel,J. & Tajfel,H. (Eds.) (1972). The context of Social Psychology: a critical assessment. London: Academic Press.

Jahoda,G. (1988) Critical Notes and Reflections on "Social Representations". European Journal of Social Psychology. 18 (3): 195-209.

Jodelet,D. (Ed.)(1989). Les Representations Sociales. Paris: P.U.F.

Lakatos,I. (1978). The methodology of scientific research programmes. Cambridge: Cambridge University Press.

Latour,B. (1991) Nous n'avons jamais eté modernes. Essai d'Antrophologie Symetrique. Paris: Seuil.

Lévy,P. (1990). Les technologies de l'intelligence. Paris: Éditions La Découverte.

McKinley,A; Potter,J. (1987). Social Representations: A Conceptual Critique. Journal for the Theory of Social Behaviour. 17: 471-487.

Moscovici,S. (1961). La Psychanalyse, son image et son public. Paris. P.U.F.

Moscovici,S. (1980). Toward a Theory of Conversion Behaviour. In L.Berkowitz (Ed.) Advances in Experimental Social Psychology. New York: Academic Press (Vol.13).

Moscovici,S & Mugny,G. (Eds.) (1987). Psychologie de la conversion. Cousset: Del Val.

Parker,I. (1987). "Social Representations": Social Psychology's (mis) use of Sociology. Journal for the Theory of Social Behaviour. 17: 447-469.

Parker,I. (1990). Discourse: Definitions and Contradictions. Philosophical Psychology. 3(2): 189-204.

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

Potter,J. & Litton,I. (1985). Some Problems Underlying the Theory of Social Representations. British Journal of Social Psychology. 24: 81-90.

Potter,J. & Wetherell,M. (1987). Discourse and Social Psychology. Beverly Hills: Sage.

Potter,J.; Wetherell,M.; Gill,R.; Edwards,D. (1990). Discourse: Noun, Verb or Social Practice?. Philosophical Psychology. 3(2): 205-.

Roqueplo,Ph. (1983). Penser la téchnique. Paris: Seuil.

Rorty,R. (1979). Philosophy and the mirror of nature. Princeton:Princeton University Press.

Shotter,J. (1984). Social Accountability and Selfhood. Oxford: Blackwell.

Tajfel,H. (1981). Human groups and social categories. Cambridge: Cambridge University Press.

Thomas,I.W.; Thomas,D.S. (1928). The Child in America: Behaviour, Problems and Programs. New York: Knopf.

Turner,J. (Ed.)(1987). Rediscovering the social group. A self categorization theory. Oxford: Blackwell.

Vygotsky,L. (1931). Mind in society, the development of higher psychological processes. Cambridge: Harvard University Press, 1978.

Wetherell,M.; Potter,J. (1992). Mapping the Language of Racism. Brighton: Harvester Wheatsheaf.

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

Woolgar,S. (1989). The ideology of Representation and the role of the agent. In H.Lawson & L.Appignanesi (Eds.) Dismantling Truth. London: Weidenfeld and Nicholson.