The Postmodern Condition (1979) publ. Manchester University Press, 1984. The First 5 Chapters of main body of work are reproduced here.
Our working hypothesis is that the status of knowledge is altered as societies enter what is known as the postindustrial age and cultures enter what is known as the postmodern age. This transition has been under way since at least the end of the 1950s, which for Europe marks the completion of reconstruction. The pace is faster or slower depending on the country, and within countries it varies according to the sector of activity: the general situation is one of temporal disjunction which makes sketching an overview difficult. A portion of the description would necessarily be conjectural. At any rate, we know that it is unwise to put too much faith in futurology.
Rather than painting a picture that would inevitably remain incomplete, I will take as my point of departure a single feature, one that immediately defines our object of study. Scientific knowledge is a kind of discourse. And it is fair to say that for the last forty years the “leading” sciences and technologies have had to do with language: phonology and theories of linguistics, problems of communication and cybernetics, modern theories of algebra and informatics, computers and their languages, problems of translation and the search for areas of compatibility among computer languages, problems of information storage and data banks, telematics and the perfection of intelligent terminals, to paradoxology. The facts speak for themselves (and this list is not exhaustive).
These technological transformations can be expected to have a considerable impact on knowledge. Its two principal functions — research and the transmission of acquired learning-are already feeling the effect, or will in the future. With respect to the first function, genetics provides an example that is accessible to the layman: it owes its theoretical paradigm to cybernetics. Many other examples could be cited. As for the second function, it is common knowledge that the miniaturisation and commercialisation of machines is already changing the way in which learning is acquired, classified, made available, and exploited. It is reasonable to suppose that the proliferation of information-processing machines is having, and will continue to have, as much of an effect on the circulation of learning as did advancements in human circulation (transportation systems) and later, in the circulation of sounds and visual images (the media).
The nature of knowledge cannot survive unchanged within this context of general transformation. It can fit into the new channels, and become operational, only if learning is translated into quantities of information.” We can predict that anything in the constituted body of knowledge that is not translatable in this way will be abandoned and that the direction of new research will be dictated by the possibility of its eventual results being translatable into computer language. The “producers” and users of knowledge must now, and will have to, possess the means of translating into these languages whatever they want to invent or learn. Research on translating machines is already well advanced.” Along with the hegemony of computers comes a certain logic, and therefore a certain set of prescriptions determining which statements are accepted as “knowledge” statements.
We may thus expect a thorough exteriorisation of knowledge with respect to the “knower,” at whatever point he or she may occupy in the knowledge process. The old principle that the acquisition of knowledge is indissociable from the training (Bildung) of minds, or even of individuals, is becoming obsolete and will become ever more so. The relationships of the suppliers and users of knowledge to the knowledge they supply and use is now tending, and will increasingly tend, to assume the form already taken by the relationship of commodity producers and consumers to the commodities they produce and consume — that is, the form of value. Knowledge is and will be produced in order to be sold, it is and will be consumed in order to be valorised in a new production: in both cases, the goal is exchange.
Knowledge ceases to be an end in itself, it loses its “use-value.”
It is widely accepted that knowledge has become the principle force of production over the last few decades, this has already had a noticeable effect on the composition of the work force of the most highly developed countries and constitutes the major bottleneck for the developing countries. In the postindustrial and postmodern age, science will maintain and no doubt strengthen its preeminence in the arsenal of productive capacities of the nation-states. Indeed, this situation is one of the reasons leading to the conclusion that the gap between developed and developing countries will grow ever wider in the future.
But this aspect of the problem should not be allowed to overshadow the other, which is complementary to it. Knowledge in the form of an informational commodity indispensable to productive power is already, and will continue to be, a major — perhaps the major — stake in the worldwide competition for power. It is conceivable that the nation-states will one day fight for control of information, just as they battled in the past for control over territory, and afterwards for control of access to and exploitation of raw materials and cheap labor. A new field is opened for industrial and commercial strategies on the one hand, and political and military strategies on the other.
However, the perspective I have outlined above is not as simple as I have made it appear. For the merchantilisation of knowledge is bound to affect the privilege the nation-states have enjoyed, and still enjoy, with respect to the production and distribution of learning. The notion that learning falls within the purview of the State, as the brain or mind of society, will become more and more outdated with the increasing strength of the opposing principle, according to which society exists and progresses only if the messages circulating within it are rich in information and easy to decode. The ideology of communicational “transparency,” which goes hand in hand with the commercialisation of knowledge, will begin to perceive the State as a factor of opacity and “noise.” It is from this point of view that the problem of the relationship between economic and State powers threatens to arise with a new urgency.
Already in the last few decades, economic powers have reached the point of imperilling the stability of the state through new forms of the circulation of capital that go by the generic name of multi-national corporations. These new forms of circulation imply that investment decisions have, at least in part, passed beyond the control of the nation-states.” The question threatens to become even more thorny with the development of computer technology and telematics. Suppose, for example, that a firm such as IBM is authorised to occupy a belt in the earth’s orbital field and launch communications satellites or satellites housing data banks. Who will have access to them? Who will determine which channels or data are forbidden? The State? Or will the State simply be one user among others? New legal issues will be raised, and with them the question: “who will know?”
Transformation in the nature of knowledge, then, could well have repercussions on the existing public powers, forcing them to reconsider their relations (both de jure and de facto) with the large corporations and, more generally, with civil society. The reopening of the world market, a return to vigorous economic competition, the breakdown of the hegemony of American capitalism, the decline of the socialist alternative, a probable opening of the Chinese market these and many other factors are already, at the end of the 1970s, preparing States for a serious reappraisal of the role they have been accustomed to playing since the 1930s: that of, guiding, or even directing investments. In this light, the new technologies can only increase the urgency of such a re-examination, since they make the information used in decision making (and therefore the means of control) even more mobile and subject to piracy.
It is not hard to visualise learning circulating along the same lines as money, instead of for its “educational” value or political (administrative, diplomatic, military) importance; the pertinent distinction would no longer be between knowledge and ignorance, but rather, as is the case with money, between “payment knowledge” and “investment knowledge” — in other words, between units of knowledge exchanged in a daily maintenance framework (the reconstitution of the work force, “survival”) versus funds of knowledge dedicated to optimising the performance of a project.
If this were the case, communicational transparency would be similar to liberalism. Liberalism does not preclude an organisation of the flow of money in which some channels are used in decision making while others are only good for the payment of debts. One could similarly imagine flows of knowledge travelling along identical channels of identical nature, some of which would be reserved for the “decision makers,” while the others would be used to repay each person’s perpetual debt with respect to the social bond.
That is the working hypothesis defining the field within which I intend to consider the question of the status of knowledge. This scenario, akin to the one that goes by the name “the computerisation of society” (although ours is advanced in an entirely different spirit), makes no claims of being original, or even true. What is required of a working hypothesis is a fine capacity for discrimination. The scenario of the computerisation of the most highly developed societies allows us to spotlight (though with the risk of excessive magnification) certain aspects of the transformation of knowledge and its effects on public power and civil institutions — effects it would be difficult to perceive from other points of view. Our hypotheses, therefore, should not be accorded predictive value in relation to reality, but strategic value in relation to the question raised.
Nevertheless, it has strong credibility, and in that sense our choice of this hypothesis is not arbitrary. It has been described extensively by the experts and is already guiding certain decisions by the governmental agencies and private firms most directly concerned, such as those managing the telecommunications industry. To some extent, then, it is already a part of observable reality. Finally, barring economic stagnation or a general recession (resulting, for example, from a continued failure to solve the world’s energy problems), there is a good chance that this scenario will come to pass: it is hard to see what other direction contemporary technology could take as an alternative to the computerisation of society.
This is as much as to say that the hypothesis is banal. But only to the extent that it fails to challenge the general paradigm of progress in science and technology, to which economic growth and the expansion of sociopolitical power seem to be natural complements. That scientific and technical knowledge is cumulative is never questioned. At most, what is debated is the form that accumulation takes — some picture it as regular, continuous, and unanimous, others as periodic, discontinuous, and conflictual.
But these truisms are fallacious. In the first place, scientific knowledge does not represent the totality of knowledge; it has always existed in addition to, and in competition and conflict with, another kind of knowledge, which I will call narrative in the interests of simplicity (its characteristics will be described later). I do not mean to say that narrative knowledge can prevail over science, but its model is related to ideas of internal equilibrium and conviviality next to which contemporary scientific knowledge cuts a poor figure, especially if it is to undergo an exteriorisation with respect to the “knower” and an alienation from its user even greater than has previously been the case. The resulting demoralisation of researchers and teachers is far from negligible; it is well known that during the 1960s, in all of the most highly developed societies, it reached such explosive dimensions among those preparing to practice these professions — the students — that there was noticeable decrease in productivity at laboratories and universities unable to protect themselves from its contamination. Expecting this, with hope or fear, to lead to a revolution (as was then often the case) is out of the question: it will not change the order of things in postindustrial society overnight. But this doubt on the part of scientists must be taken into account as a major factor in evaluating the present and future status of scientific knowledge.
It is all the more necessary to take it into consideration since — and this is the second point — the scientists’ demoralisation has an impact on the central problem of legitimation. I use the word in a broader sense than do contemporary German theorists in their discussions of the question of authority. Take any civil law as an example: it states that a given category of citizens must perform a specific kind of action. Legitimation is the process by which a legislator is authorised to promulgate such a law as a norm. Now take the example of a scientific statement: it is subject to the rule that a statement must fulfil a given set of conditions in order to be accepted as scientific. In this case, legitimation is the process by which a “legislator” dealing with scientific discourse is authorised to prescribe the stated conditions (in general, conditions of internal consistency and experimental verification) determining whether a statement is to be included in that discourse for consideration by the scientific community.
The parallel may appear forced. But as we will see, it is not. The question of the legitimacy of science has been indissociably linked to that of the legitimation of the legislator since the time of Plato. From this point of view, the right to decide what is true is not independent of the right to decide what is just, even if the statements consigned to these two authorities differ in nature. The point is that there is a strict interlinkage between the kind of language called science and the kind called ethics and politics: they both stem from the same perspective, the same “choice” if you will — the choice called the Occident.
When we examine the current status of scientific knowledge at a time when science seems more completely subordinated to the prevailing powers than ever before and, along with the new technologies, is in danger of becoming a major stake in their conflicts — the question of double legitimation, far from receding into the background, necessarily comes to the fore. For it appears in its most complete form, that of reversion, revealing that knowledge and power are simply two sides of the same question: who decides what knowledge is, and who knows what needs to be decided? In the computer age, the question of knowledge is now more than ever a question of government.
The reader will already have noticed that in analysing this problem within the framework set forth I have favoured a certain procedure: emphasising facts of language and in particular their pragmatic aspect. To help clarify what follows it would be useful to summarise, however briefly, what is meant here by the term pragmatic.
A denotative utterance such as “The university is sick,” made in the context of a conversation or an interview, positions its sender (the person who utters the statement), its addressee (the person who receives it), and its referent (what the statement deals with) in a specific way: the utterance places (and exposes) the sender in the position of “knower” (he knows what the situation is with the university), the addressee is put in the position of having to give or refuse his assent, and the referent itself is handled in a way unique to denotatives, as something that demands to be correctly identified and expressed by the statement that refers to it.
If we consider a declaration such as “The university is open,” pronounced by a dean or rector at convocation, it is clear that the previous specifications no longer apply. Of course, the meaning of the utterance has to be understood, but that is a general condition of communication and does not aid us in distinguishing the different kinds of utterances or their specific effects. The distinctive feature of this second, “performative,” utterance is that its effect upon the referent coincides with its enunciation. The university is open because it has been declared open in the above-mentioned circumstances. That this is so is not subject to discussion or verification on the part of the addressee, who is immediately placed within the new context created by the utterance. As for the sender, he must be invested “with the” authority to make such a statement. Actually, we could say it the other way around: the sender is dean or rector that is, he is invested with the authority to make this kind of statement — only insofar as he can directly affect both the referent, (the university) and the addressee (the university staff) in the manner I have indicated.
A different case involves utterances of the type, “Give money to the university”; these are prescriptions. They can be modulated as orders, commands, instructions, recommendations, requests, prayers, pleas, etc. Here, the sender is clearly placed in a position of authority, using the term broadly (including the authority of a sinner over a god who claims to be merciful): that is, he expects the addressee to perform the action referred to. The pragmatics of prescription entail concomitant changes in the posts of addressee and referent.
Of a different order again is the efficiency of a question, a promise, a literary description, a narration, etc. I am summarising. Wittgenstein, taking up the study of language again from scratch, focuses his attention on the effects of different modes of discourse; he calls the various types of utterances he identifies along the way (a few of which I have listed) language games. What he means by this term is that each of the various categories of utterance can be defined in terms of rules specifying their properties and the uses to which they can be put — in exactly the same way as the game of chess is defined by a set of rules determining the properties of each of the pieces, in other words, the proper way to move them.
It is useful to make the following three observations about language games. The first is that their rules do not carry within themselves their own legitimation, but are the object of a contract, explicit or not, between players (which is not to say that the players invent the rules). The second is that if there are no rules, there is no game, that even an infinitesimal modification of one rule alters the nature of the game, that a “move” or utterance that does not satisfy the rules does not belong to the game they define. The third remark is suggested by what has just been said: every utterance should be thought of as a “move” in a game.
This last observation brings us to the first principle underlying our method as a whole: to speak is to fight, in the sense of playing, and speech acts fall within the domain of a general agonistics. This does not necessarily mean that one plays in order to win. A move can be made for the sheer pleasure of its invention: what else is involved in that labor of language harassment undertaken by popular speech and by literature? Great joy is had in the endless invention of turns of phrase, of words and meanings, the process behind the evolution of language on the level of parole. But undoubtedly even this pleasure depends on a feeling of success won at the expense of an adversary — at least one adversary, and a formidable one: the accepted language, or connotation.
This idea of an agonistics of language should not make us lose sight of the second principle, which stands as a complement to it and governs our analysis: that the observable social bond is composed of language “moves.” An elucidation of this proposition will take us to the heart of the matter at hand.
If we wish to discuss knowledge in the most highly developed contemporary society, we must answer the preliminary question of what methodological representation to apply to that society. Simplifying to the extreme, it is fair to say that in principle there have been, at least over the last half-century, two basic representational models for society: either society forms a functional whole, or it is divided in two. An illustration of the first model is suggested by Talcott Parsons (at least the postwar Parsons) and his school, and of the second, by the Marxist current (all of its component schools, whatever differences they may have, accept both the principle of class struggle and dialectics as a duality operating within society).”
This methodological split, which defines two major kinds of discourse on society, has been handed down from the nineteenth century. The idea that society forms an organic whole, in the absence of which it ceases to be a society (and sociology ceases to have an object of study), dominated the minds of the founders of the French school. Added detail was supplied by functionalism; it took yet another turn in the 1950s with Parsons’s conception of society as a self-regulating system. The theoretical and even material model is no longer the living organism; it is provided by cybernetics, which, during and after the Second World War, expanded the model’s applications.
In Parsons’s work, the principle behind the system is still, if I may say so, optimistic: it corresponds to the stabilisation of the growth economies and societies of abundance under the aegis of a moderate welfare state. In the work of contemporary German theorists, systemtheorie is technocratic, even cynical, not to mention despairing: the harmony between the needs and hopes of individuals or groups and the functions guaranteed by the system is now only a secondary component of its functioning. The true goal of the system, the reason it programs itself like a computer, is the optimisation of the global relationship between input and output, in other words, performativity. Even when its rules are in the process of changing and innovations are occurring, even when its dysfunctions (such as strikes, crises, unemployment, or political revolutions) inspire hope and lead to belief in an alternative, even then what is actually taking place is only an internal readjustment, and its result can be no more than an increase in the system’s “viability.” The only alternative to this kind of performance improvement is entropy, or decline.
Here again, while avoiding the simplifications inherent in a sociology of social theory, it is difficult to deny at least a parallel between this “hard” technocratic version of society and the ascetic effort that was demanded (the fact that it was done in name of “advanced liberalism” is beside the point) of the most highly developed industrial societies in order to make them competitive — and thus optimise their “irrationality” — within the framework of the resumption of economic world war in the 1960s.
Even taking into account the massive displacement intervening between the thought of a man like Comte and the thought of Luhmann, we can discern a common conception of the social: society is a unified totality, a “unicity.” Parsons formulates this clearly: “The most essential condition of successful dynamic analysis is a continual and systematic reference of every problem to the state of the system as a whole … A process or set of conditions either “contributes” to the maintenance (or development) of the system or it is “dysfunctional” in that it detracts from the integration, effectiveness, etc., of the “system.” The “technocrats” also subscribe to this idea. Whence its credibility: it has the means to become a reality, and that is all the proof it needs. This is what Horkheimer called the “paranoia” of reason.
But this realism of systemic self-regulation, and this perfectly sealed circle of facts and interpretations, can be judged paranoid only if one has, or claims to have, at one’s disposal a viewpoint that is in principle immune from their allure. This is the function of the principle of class struggle in theories of society based on the work of Marx.
“Traditional” theory is always in danger of being incorporated into the programming of the social whole as a simple tool for the optimisation of its performance; this is because its desire for a unitary and totalising truth lends itself to the unitary and totalising practice of the system’s managers. “Critical” theory, based on a principle of dualism and wary of syntheses and reconciliations, should be in a position to avoid this fate. What guides Marxism, then, is a different model of society, and a different conception of the function of the knowledge that can be produced by society and acquired from it. This model was born of the struggles accompanying the process of capitalism’s encroachment upon traditional civil societies. There is insufficient space here to chart the vicissitudes of these struggles, which fill more than a century of social, political, and ideological history. We will have to content ourselves with a glance at the balance sheet, which is possible for us to tally today now that their fate is known: in countries with liberal or advanced liberal management, the struggles and their instruments have been transformed into regulators of the system; in communist countries, the totalising model and its totalitarian effect have made a comeback in the name of Marxism itself, and the struggles in question have simply been deprived of the right to exist. Everywhere, the Critique of political economy (the subtitle of Marx’s Capital) and its correlate, the critique of alienated society, are used in one way or another as aids in programming the system.
Of course, certain minorities, such as the Frankfurt School or the group Socialisme ou barbarie, preserved and refined the critical model in opposition to this process. But the social foundation of the principle of division, or class struggle, was blurred to the point of losing all of its radicality; we cannot conceal the fact that the critical model in the end lost its theoretical standing and was reduced to the status of a “utopia” or “hope,” a token protest raised in the name of man or reason or creativity, or again of some social category such as the Third World or the students — on which is conferred in extremes the henceforth improbable function of critical subject.
The sole purpose of this schematic (or skeletal) reminder has been to specify the problematic in which I intend to frame the question of knowledge in advanced industrial societies. For it is impossible to know what the state of knowledge is — in other words, the problems its development and distribution are facing today — without knowing something of the society within which it is situated. And today more than ever, knowing about that society involves first of all choosing what approach the inquiry will take, and that necessarily means choosing how society can answer. One can decide that the principal role of knowledge is as an indispensable element in the functioning of society, and act in accordance with that decision, only if one has already decided that society is a giant machine.
Conversely, one can count on its critical function, and orient its development and distribution in that direction, only after it has been decided that society does not form an integrated whole, but remains haunted by a principle of oppositions The alternative seems clear: it is a choice between the homogeneity and the intrinsic duality of the social, between functional and critical knowledge. But the decision seems difficult, or arbitrary.
It is tempting to avoid the decision altogether by distinguishing two kinds of knowledge. one, the positivist kind, would be directly applicable to technologies bearing on men and materials, and would lend itself to operating as an indispensable productive force within the system. The other the critical, reflexive, or hermeneutic kind by reflecting directly or indirectly on values or alms, would resist any such “recuperation.”
I find this partition solution unacceptable. I suggest that the alternative it attempts to resolve, but only reproduces, is no longer relevant for the societies with which we are concerned and that the solution itself is stilt caught within a type of oppositional thinking that is out of step with the most vital modes of postmodern knowledge. As I have already said, economic “redeployment” in the current phase of capitalism, aided by a shift in techniques and technology, goes hand in hand with a change in the function of the State: the image of society this syndrome suggests necessitates a serious revision of the alternate approaches considered. For brevity’s sake, suffice it to say that functions of regulation, and therefore of reproduction, are being and will be further withdrawn from administrators and entrusted to machines. Increasingly, the central question is becoming who will have access to the information these machines must have in storage to guarantee that the right decisions are made. Access to data is, and will continue to be, the prerogative of experts of all stripes. The ruling class is and will continue to be the class of decision makers. Even now it is no longer composed of the traditional political class, but of a composite layer of corporate leaders, high-level administrators, and the heads of the major professional, labor, political, and religious organisations.
What is new in all of this is that the old poles of attraction represented by nation-states, parties, professions, institutions, and historical traditions are losing their attraction. And it does not look as though they wilt be replaced, at least not on their former scale, The Trilateral Commission is not a popular pole of attraction. “Identifying” with the great names, the heroes of contemporary history, is becoming more and more difficult. Dedicating oneself to “catching up with Germany,” the life goal the French president [Giscard d’Estaing at the time this book was published in France] seems to be offering his countrymen, is not exactly exciting. But then again, it is not exactly a life goal. It depends on each individual’s industriousness. Each individual is referred to himself. And each of us knows that our self does not amount to much.
This breaking up of the grand Narratives (discussed below, sections 9 and 10) leads to what some authors analyse in terms of the dissolution of the social bond and the disintegration of social aggregates into a mass of individual atoms thrown into the absurdity of Brownian motion. Nothing of the kind is happening: this point of view, it seems to me, is haunted by the paradisaic representation of a lost organic” society.
A self does not amount to much, but no self is an island; each exists in a fabric of relations that is now more complex and mobile than ever before. Young or old, man or woman, rich or poor, a person is always located at “nodal points” of specific communication circuits, however tiny these may be. Or better: one is always located at a post through which various kinds of messages pass. No one, not even the least privileged among us, is ever entirely powerless over the messages that traverse and position him at the post of sender, addressee, or referent. One’s mobility in relation to these language game effects (language games, of course, are what this is all about) is tolerable, at least within certain limits (and the limits are vague); it is even solicited by regulatory mechanisms, and in particular by the self-adjustments the system undertakes in order to improve its performance. It may even be said that the system can and must encourage such movement to the extent that it combats its own entropy, the novelty of an unexpected “move,” with its correlative displacement of a partner or group of partners, can supply the system with that increased performativity it forever demands and consumes.
It should now be clear from which perspective I chose language games as my general methodological approach. I am not claiming that the entirety of social relations is of this nature — that will remain an open question. But there is no need to resort to some fiction of social origins to establish that language games are the minimum relation required for society to exist: even before he is born, if only by virtue of the name he is given, the human child is already positioned as the referent in the story recounted by those around him, in relation to which he will inevitably chart his course. Or more simply still, the question of the social bond, insofar as it is a question, is itself a language game, the game of inquiry. It immediately positions the person who asks, as well as the addressee and the referent asked about: it is already the social bond.
On the other hand, in a society whose communication component is becoming more prominent day by day, both as a reality and as an issue, it is clear that language assumes a new importance. It would be superficial to reduce its significance to the traditional alternative between manipulatory speech and the unilateral transmission of messages on the one hand, and free expression and dialogue on the other.
A word on this last point. If the problem is described simply in terms of communication theory, two things are overlooked: first, messages have quite different forms and effects depending on whether they are, for example, denotatives, prescriptives, evaluatives, performatives, etc. It is clear that what is important is not simply the fact that they communicate information. Reducing them to this function is to adopt an outlook which unduly privileges the system’s own interests and point of view. A cybernetic machine does indeed run on information, but the goals programmed into it, for example, originate in prescriptive and evaluative statements it has no way to correct in the course of its functioning — for example, maximising its own performance, how can one guarantee that performance maximisation is the best goal for the social system in every case. In any case the “atoms” forming its matter are competent to handle statements such as these — and this question in particular.
Second, the trivial cybernetic version of information theory misses something of decisive importance, to which I have already called attention: the agonistic aspect of society. The atoms are placed at the crossroads of pragmatic relationships, but they are also displaced by the messages that traverse them, in perpetual motion. Each language partner, when a “move” pertaining to him is made, undergoes a “displacement,” an alteration of some kind that not only affects him in his capacity as addressee and referent, but also as sender. These moves necessarily provoke “countermoves” and everyone knows that a countermove that is merely reactional is not a “good” move. Reactional countermoves arc no more than programmed effects in the opponent’s strategy; they play into his hands and thus have no effect on the balance of power. That is why it is important to increase displacement in the games, and even to disorient it, in such a way as to make an unexpected “move” (a new statement).
What is needed if we are to understand social relations in this manner, on whatever scale we choose, is not only a theory of communication, but a theory of games which accepts agonistics as a founding principle. In this context, it is easy to see that the essential element of newness is not simply “innovation.” Support for this approach can be found in the work of a number of contemporary sociologists, in addition to linguists and philosophers of language. This “atomisation” of the social into flexible networks of language games may seem far removed from the modern reality, which is depicted, on the contrary, as afflicted with bureaucratic paralysis. The objection will be made, at least, that the weight of certain institutions imposes limits on the games, and thus restricts the inventiveness of the players in making their moves. But I think this can be taken into account without causing any particular difficulty.
In the ordinary use of discourse — for example, in a discussion between two friends — the interlocutors use any available ammunition, changing games from one utterance to the next: questions, requests, assertions, and narratives are launched pell-mell into battle. The war is not without rules, but the rules allow and encourage the greatest possible flexibility of utterance.
From this point of view, an institution differs from a conversation in that it always requires supplementary constraints for statements to be declared admissible within its bounds. The constraints function to filter discursive potentials, interrupting possible connections in the communication networks: there are things that should not be said. They also privilege certain classes of statements (sometimes only one) whose predominance characterises the discourse of the particular institution: there arc things that should be said, and there are ways of saving them. Thus: orders in the army, prayer in church, denotation in the schools, narration in families, questions in philosophy, performativity in businesses. Bureaucratisation is the outer limit of this tendency.
However, this hypothesis about the institution is still too “unwieldy”: its point of departure is an overly “reifying” view of what is institutionalised. We know today that the limits the institution imposes on potential language “moves” are never established once and for all (even if they have been formally defined), Rather, the limits are themselves the stakes and provisional results of language strategies, within the institution and without. Examples: Does the university have a place for language experiments (poetics)? Can you tell stories in a cabinet meeting? Advocate a cause in the barracks? The answers are clear: yes, if the university opens creative workshops; yes, if the cabinet works with prospective scenarios; yes, if the limits of the old institution are displaced. Reciprocally, it can be said that the boundaries only stabilise when they cease to be stakes in the game.
This, I think, is the appropriate approach to contemporary institutions of knowledge.