For many years now historians have preferred to turn their attention to long periods, as if, beneath the shifts and changes of political events, they were trying to reveal the stable, almost indestructible system of checks and balances, the irreversible processes, the constant readjustments, the underlying tendencies that gather force, and are then suddenly reversed after centuries of continuity, the movements of accumulation and slow saturation, the great silent, motionless bases that traditional history has covered with a thick layer of events. The tools that enable historians to carry out this work of analysis are partly inherited and partly of their own making: models of economic growth, quantitative analysis of market movements, accounts of demographic expansion and contraction, the study of climate and its long-term changes, the fixing of sociological constants, the description of technological adjustments and of their spread and continuity. These tools have enabled workers in the historical field to distinguish various sedimentary strata; linear successions, which for so long had been the object of research, have given way to discoveries in depth. From the political mobility at the surface down to the slow movements of 'material civilisation', ever more levels of analysis have been established: each has its own peculiar discontinuities and patterns; and as one descends to the deepest levels, the rhythms become broader. Beneath the rapidly changing history of governments, wars, and famines, there emerge other, apparently unmoving histories: the history of sea routes, the history of corn or of gold-mining, the history of drought and of irrigation, the history of crop rotation, the history of the balance achieved by the human species between hunger and abundance. The old questions of the traditional analysis (What link should be made between disparate events? How can a causal succession be established between them? What continuity or overall significance do they possess? Is it possible to define a totality, or must one be content with reconstituting connections?) are now being replaced by questions of another type: which strata should be isolated from others? What types of series should be established? What criteria of periodisation should be adopted for each of them? What system of relations (hierarchy, dominance, stratification, univocal determination, circular causality) may be established between them? What series of series may be established? And in what large-scale chronological table may distinct series of events be determined?

At about the same time, in the disciplines that we call the history of ideas, the history of science, the history of philosophy, the history of thought, and the history of literature (we can ignore their specificity for the moment), in those disciplines which, despite their names, evade very largely the work and methods of the historian, attention has been turned, on the contrary, away from vast unities like 'periods' or 'centuries' to the phenomena of rupture, of discontinuity. Beneath the great continuities of thought, beneath the solid, homogeneous manifestations of a single mind or of a collective mentality, beneath the stubborn development of a science striving to exist and to reach completion at the very outset, beneath the persistence of a particular genre, form, discipline, or theoretical activity, one is now trying to detect the incidence of interruptions. Interruptions whose status and nature vary considerably. There are the epistemological acts and thresholds described by Bachelard: they suspend the continuous accumulation of knowledge, interrupt its slow development, and force it to enter a new time, cut it off from its empirical origin and its original motivations, cleanse it of its imaginary complicities; they direct historical analysis away from the search for silent beginnings, and the never-ending tracing-back to the original precursors, towards the search for a new type of rationality and its various effects. There are the displacements and transformations of concepts: the analyses of G. Canguilhem may serve as models; they show that the history of a concept is not wholly and entirely that of its progressive refinement, its continuously increasing rationality, its abstraction gradient, but that of its various fields of constitution and validity, that of its successive rules of use, that of the many theoretical contexts in which it developed and matured. There is the distinction, which we also owe to Canguilhem, between the microscopic and macroscopic scales of the history of the sciences, in which events and their consequences are not arranged in the same way: thus a discovery, the development of a method, the achievements, and the failures, of a particular scientist, do not have the same incidence, and cannot be described in the same way at both levels; on each of the two levels, a different history is being written. Recurrent redistributions reveal several pasts, several forms of connection, several hierarchies of importance, several networks of determination, several teleologies, for one and the same science, as its present undergoes change: thus historical descriptions are necessarily ordered by the present state of knowledge, they increase with every transformation and never cease, in turn, to break with themselves (in the field of mathematics, M. Serres has provided the theory of this phenomenon). There are the architectonic unities of systems of the kind analysed by M. Guéroult, which are concerned not with the description of cultural influences, traditions, and continuities, but with internal coherences, axioms, deductive connections, compatibilities. Lastly, the most radical discontinuities are the breaks effected by a work of theoretical transformation which establishes a science by detaching it from the ideology of its past and by revealing this past as ideological'. To this should be added, of course, literary analysis, which now takes as its unity, not the spirit or sensibility of a period, nor 'groups', 'schools', 'generations', or 'movements', nor even the personality of the author, in the interplay of his life and his 'creation', but the particular structure of a given œuvre, book, or text.

And the great problem presented by such historical analyses is not how continuities are established, how a single pattern is formed and preserved, how for so many different, successive minds there is a single horizon, what mode of action and what substructure is implied by the interplay of transmissions, resumptions, disappearances, and repetitions, how the origin may extend its sway well beyond itself to that conclusion that is never given - the problem is no longer one of tradition, of tracing a line, but one of division, of limits; it is no longer one of lasting foundations, but one of transformations that serve as new foundations, the rebuilding of foundations. What one is seeing, then, is the emergence of a whole field of questions, some of which are already familiar, by which this new form of history is trying to develop its own theory: how is one to specify the different concepts that enable us to conceive of discontinuity (threshold, rupture, break, mutation, transformation)? By what criteria is one to isolate the unities with which one is dealing; what is a science? What is an œuvre? What is a theory? What is a concept? What is a text? How is one to diversify the levels at which one may place oneself, each of which possesses its own divisions and form of analysis? What is the legitimate level of formalisation? What is that of interpretation? Of structural analysis? Of attributions of causality?

In short, the history of thought, of knowledge, of philosophy, of literature seems to be seeking, and discovering, more and more discontinuities, whereas history itself appears to be abandoning the irruption of events in favour of stable structures.

But we must not be taken in by this apparent interchange. Despite appearances, we must not imagine that certain of the historical disciplines have moved from the continuous to the discontinuous, while others have moved from the tangled mass of discontinuities to the great, uninterrupted unities; we must not imagine that in the analysis of politics, institutions, or economics, we have become more and more sensitive to overall determinations, while in the analysis of ideas and of knowledge, we are paying more and more attention to the play of difference; we must not imagine that these two great forms of description have crossed without recognising one another.

In fact, the same problems are being posed in either case, but they have provoked opposite effects on the surface. These problems may be summed up in a word: the questioning of the document. Of course, it is obvious enough that ever since a discipline such as history has existed, documents have been used, questioned, and have given rise to questions; scholars have asked not only what these documents meant, but also whether they were telling the truth, and by what right they could claim to be doing so, whether they were sincere or deliberately misleading, well informed or ignorant, authentic or tampered with. But each of these questions, and all this critical concern, pointed to one and the same end: the reconstitution, on the basis of what the documents say, and sometimes merely hint at, of the past from which they emanate and which has now disappeared far behind them; the document was always treated as the language of a voice since reduced to silence, its fragile, but possibly decipherable trace. Now, through a mutation that is not of very recent origin, but which has still not come to an end, history has altered its position in relation to the document: it has taken as its primary task, not the interpretation of the document, nor the attempt to decide whether it is telling the truth or what is its expressive value, but to work on it from within and to develop it: history now organises the document, divides it up, distributes it, orders it, arranges it in levels, establishes series, distinguishes between what is relevant and what is not, discovers elements, defines unities, describes relations. The document, then, is no longer for history an inert material through which it tries to reconstitute what men have done or said, the events of which only the trace remains; history is now trying to define within the documentary material itself unities, totalities, series, relations. History must be detached from the image that satisfied it for so long, and through which it found its anthropological justification: that of an age-old collective consciousness that made use of material documents to refresh its memory; history is the work expended on material documentation (books, texts, accounts, registers, acts, buildings, institutions, laws, techniques, objects, customs, etc.) that exists, in every time and place, in every society, either in a spontaneous or in a consciously organised form. The document is not the fortunate tool of a history that is primarily and fundamentally memory; history is one way in which a society recognises and develops a mass of documentation with which it is inextricably linked.

To be brief, then, let us say that history, in its traditional form, undertook to 'memorise' the monuments of the past, transform them into documents, and lend speech to those traces which, in themselves, are often not verbal, or which say in silence something other than what they actually say; in our time, history is that which transforms documents into monuments. In that area where, in the past, history deciphered the traces left by men, it now deploys a mass of elements that have to be grouped, made relevant, placed in relation to one another to form totalities. There was a time when archaeology, as a discipline devoted to silent monuments, inert traces, objects without context, and things left by the past, aspired to the condition of history, and attained meaning only through the restitution of a historical discourse; it might be said, to play on words a little, that in our time history aspires to the condition of archaeology, to the intrinsic description of the monument.

This has several consequences. First of all, there is the surface effect already mentioned: the proliferation of discontinuities in the history of ideas, and the emergence of long periods in history proper. In fact, in its traditional form, history proper was concerned to define relations (of simple causality, of circular determination, of antagonism, of expression) between facts or dated events: the series being known, it was simply a question of defining the position of each element in relation to the other elements in the series. The problem now is to constitute series: to define the elements proper to each series, to fix its boundaries, to reveal its own specific type of relations, to formulate its laws, and, beyond this, to describe the relations between different series, thus constituting series of series, or 'tables': hence the ever-increasing number of strata, and the need to distinguish them, the specificity of their time and chronologies; hence the need to distinguish not only important events (with a long chain of consequences) and less important ones, but types of events at quite different levels (some very brief, others of average duration, like the development of a particular technique, or a scarcity of money, and others of a long-term nature, like a demographic equilibrium or the gradual adjustment of an economy to climatic change); hence the possibility of revealing series with widely spaced intervals formed by rare or repetitive events. The appearance of long periods in the history of today is not a return to the philosophers of history, to the great ages of the world, or to the periodisation dictated by the rise and fall of civilisations; it is the effect of the methodologically concerted development of series. In the history of ideas, of thought and of the sciences, the same mutation has brought about the opposite effect; it has broken up the long series formed by the progress of consciousness, or the teleology of reason, or the evolution of human thought; it has questioned the themes of convergence and culmination; it has doubted the possibility of creating totalities. It has led to the individualisation of different series, which are juxtaposed to one another, follow one another, overlap and intersect, without one being able to reduce them to a linear schema. Thus, in place of the continuous chronology of reason, which was invariably traced back to some inaccessible origin, there have appeared scales that are sometimes very brief, distinct from one another, irreducible to a single law, scales that bear a type of history peculiar to each one, and which cannot be reduced to the general model of a consciousness that acquires, progresses, and remembers.

Second consequence: the notion of discontinuity assumes a major role in the historical disciplines. For history in its classical form, the discontinuous was both the given and the unthinkable: the raw material of history, which presented itself in the form of dispersed events - decisions, accidents, initiatives, discoveries; the material, which, through analysis, had to be rearranged, reduced, effaced in order to reveal the continuity of events. Discontinuity was the stigma of temporal dislocation that it was the historian's task to remove from history. It has now become one of the basic elements of historical analysis. its role is threefold. First, it constitutes a deliberate operation on the part of the historian (and not a quality of the material with which he has to deal): for he must, at least as a systematic hypothesis, distinguish the possible levels of analysis, the methods proper to each, and the periodisation that best suits them. Secondly, it is the result of his description (and not something that must be eliminated by means of his analysis): for he is trying to discover the limits of a process, the point of inflection of a curve, the inversion of a regulatory movement, the boundaries of an oscillation, the threshold of a function, the instant at which a circular causality breaks down. Thirdly, it is the concept that the historian's work never ceases to specify (instead of neglecting it as a uniform, indifferent blank between two positive figures); it assumes a specific form and function according to the field and the level to which it is assigned: one does not speak of the same discontinuity when describing an epistemological threshold, the point of reflexion in a population curve, or the replacement of one technique by another. The notion of discontinuity is a paradoxical one: because it is both an instrument and an object of research; because it divides up the field of which it is the effect; because it enables the historian to individualise different domains but can be established only by comparing those domains. And because, in the final analysis, perhaps, it is not simply a concept present in the discourse of the historian, but something that the historian secretly supposes to be present: on what basis, in fact, could he speak without this discontinuity that offers him history - and his own history - as an object? One of the most essential features of the new history is probably this displacement of the discontinuous: its transference from the obstacle to the work itself; its integration into the discourse of the historian, where it no longer plays the role of an external condition that must be reduced, but that of a working concept; and therefore the inversion of signs by which it is no longer the negative of the historical reading (its underside, its failure, the limit of its power), but the positive element that determines its object and validates its analysis.

Third consequence: the theme and the possibility of a total history begin to disappear, and we see the emergence of something very different that might be called a general history. The project of a total history is one that seeks to reconstitute the overall form of a civilisation, the principle material or spiritual - of a society, the significance common to all the phenomena of a period, the law that accounts for their cohesion - what is called metaphorically the 'face' of a period. Such a project is linked to two or three hypotheses; - it is supposed that between all the events of a well-defined spatio-temporal area, between all the phenomena of which traces have been found, it must be possible to establish a system of homogeneous relations: a network of causality that makes it possible to derive each o them, relations of analogy that show how they symbolise one another, or how they all express one and the same central core; it is also supposed that one and the same form of historicity operates upon economic structures, social institutions and customs, the inertia of mental attitudes, technological practice, political behaviour, and subjects them all to the same type of transformation; lastly, it is supposed that history itself may be articulated into great units - stages or phases - which contain within themselves their own principle of cohesion. These are the postulates that are challenged by the new history when it speaks of series, divisions, limits, differences of level, shifts, chronological specificities, particular forms of rehandling, possible types of relation. This is not because it is trying to obtain a plurality of histories juxtaposed and independent of one another: that of the economy beside that of institutions, and beside these two those of science, religion, or literature; nor is it because it is merely trying to discover between these different histories coincidences of dates, or analogies of form and meaning. The problem that now presents itself- and which defines the task of a general history - is to determine what form of relation may be legitimately described between these different series; what vertical system they are capable of forming; what interplay of correlation and dominance exists between them; what may be the effect of shifts, different temporalities, and various rehandlings; in what distinct totalities certain elements may figure simultaneously; in short, not only what series, but also what 'series of series' - or, in other words, what 'tables' it is possible to draw up. A total description draws all phenomena around a single centre - a principle, a meaning, a spirit, a world-view, an overall shape; a general history, on the contrary, would deploy the space of a dispersion.

Fourth and last consequence: the new history is confronted by a number of methodological problems, several of which, no doubt, existed long before the emergence of the new history, but which, taken together, characterise it. These include: the building-up of coherent and homogeneous corpora of documents (open or closed, exhausted or inexhaustible corpora), the establishment of a principle of choice (according to whether one wishes to treat the documentation exhaustively, or adopt a sampling method as in statistics, or try to determine in advance which are the most representative elements); the definition of the level of analysis and of the relevant elements (in the material studied, one may extract numerical indications; references - explicit or not - to events, institutions, practices; the words used, with their grammatical rules and the semantic fields that they indicate, or again the formal structure of the propositions and the types of connection that unite them); the specification of a method of analysis (the quantitative treatment of data, the breaking-down of the material according to a number of assignable features whose correlations are then studied, interpretative decipherment, analysis of frequency and distribution); the delimitation of groups and sub-groups that articulate the material (regions, periods, unitary processes); the determination of relations that make it possible to characterise a group (these may be numerical or logical relations; functional, causal, or analogical relations; or it may be the relation of the 'signifier' (signs) to the 'signified' (signifé).

All these problems are now part of the methodological field of history. This field deserves attention, and for two reasons. First, because one can see to what extent it has freed itself from what constituted, not so long ago, the philosophy of history, and from the questions that it posed (on the rationality or teleology of historical development (devenir), on the relativity of historical knowledge, and on the possibility of discovering or constituting a meaning in the inertia of the past and in the unfinished totality of the present). Secondly, because it intersects at certain points problems that are met with in other fields - in linguistics, ethnology, economics, literary analysis, and mythology, for example. These problems may, if one so wishes, be labelled structuralism. But only under certain conditions: they do not, of themselves, cover the entire methodological field of history, they occupy only one part of that field - a part that varies in importance with the area and level of analysis; apart from a number of relatively limited cases, they have not been imported from linguistics or ethnology (as is often the case today), but they originated in the field of history itself - more particularly, in that of economic history and as a result of the questions posed by that discipline; lastly, in no way do they authorise us to speak of a structuralism of history, or at least of an attempt to overcome a 'conflict' or 'opposition' between structure and historical development: it is a long time now since historians uncovered, described, and analysed structures, without ever having occasion to wonder whether they were not allowing the living, fragile, pulsating 'history' to slip through their fingers. The structure/development opposition is relevant neither to the definition of the historical field, nor, in all probability, to the definition of a structural method.

This epistemological mutation of history is not yet complete. But it is not of recent origin either, since its first phase can no doubt be traced back to Marx. But it took a long time to have much effect. Even now - and this is especially true in the case of the history of thought - it has been neither registered nor reflected upon, while other, more recent transformations - those of linguistics, for example - have been. It is as if it was particularly difficult, in the history in which men retrace their own ideas and their own knowledge, to formulate a general theory of discontinuity, of series, of limits, unities, specific orders, and differentiated autonomies and dependences. As if, in that field where we had become used to seeking origins, to pushing back further and further the line of antecedents, to reconstituting traditions, to following evolutive curves, to projecting teleologies, and to having constant recourse to metaphors of life, we felt a particular repugnance to conceiving of difference, to describing separations and dispersions, to dissociating the reassuring form of the identical. Or, to be more precise, as if we found it difficult to construct a theory, to draw general conclusions, and even to derive all the possible implications of these concepts of thresholds, mutations, independent systems, and limited series - in the way in which they had been used in fact by historians. As if we were afraid to conceive of the Other in the time of our own thought.

There is a reason for this. If the history of thought could remain the locus of uninterrupted continuities, if it could endlessly forge connections that no analysis could undo without abstraction, if it could weave, around everything that men say and do, obscure synthesis that anticipate for him, prepare him, and lead him endlessly towards his future, it would provide a privileged shelter for the sovereignty of consciousness. Continuous history is the indispensable correlative of the founding function of the subject: the guarantee that everything that has eluded him may be restored to him; the certainty that time will disperse nothing without restoring it in a reconstituted unity; the promise that one day the subject - in the form of historical consciousness - will once again be able to appropriate, to bring back under his sway, all those things that are kept at a distance by difference, and find in them what might be called his abode. Making historical analysis the discourse of the continuous and making human consciousness the original subject of all historical development and all action are the two sides of the same system of thought. In this system, time is conceived in terms of totalisation and revolutions are never more than moments of consciousness.

In various forms, this theme has played a constant role since the nineteenth century: to preserve, against all decentrings, the sovereignty of the subject, and the twin figures of anthropology and humanism. Against the decentring operated by Marx - by the historical analysis of the relations of reduction, economic determinations, and the class struggle - it gave place towards the end of the nineteenth century, to the search for a total history, in which all the differences of a society might be reduced to a single form, to the organisation of a world-view, to the establishment of a system of values, to a coherent type of civilisation. To the decentring operated by the Nietzschean genealogy, it opposed the search for an original foundation that would make rationality the telos of mankind, and link the whole history of thought to the preservation of this rationality, to the maintenance of this teleology, and to the ever necessary return to this foundation. Lastly, more recently, when the researches of psychoanalysis, linguistics, and ethnology have decentred the subject in relation to the laws of his desire, the forms of his language, the rules of his action, or the games of his mythical or fabulous discourse, when it became clear that man himself, questioned as to what he was, could not account for his sexuality and his unconscious, the systematic forms of his language, or the regularities of his fictions, the theme of a continuity of history has been reactivated once again; a history that would be not division, but development (devenir); not an interplay of relations, but an internal dynamic; not, a system, but the hard work of freedom; not form, but the unceasing effort of a consciousness turned upon itself, trying to grasp itself in its deepest conditions: a history that would be both an act of long, uninterrupted patience and the vivacity of a movement, which, in the end, breaks all bounds. If one is to assert this theme, which, to the 'immobility' of structures, to their 'closed' system, to their necessary 'synchrony', opposes the living openness of history, one must obviously deny in the historical analyses themselves the use of discontinuity, the definition of levels and limits, the description of specific series, the uncovering of the whole interplay of differences. One is led therefore to anthropologise Marx, to make of him a historian of totalities, and to rediscover in him the message of humanism; one is led therefore to interpret Nietzsche in the terms of transcendental philosophy, and to reduce his genealogy to the level of a search for origins; lastly, one is led to leave to one side, as if it had never arisen, that whole field of methodological problems that the new history is now presenting.- For, if it is asserted that the question of discontinuities, systems and transformations, series and thresholds, arises in all the historical disciplines (and in those concerned with ideas or the sciences no less than those concerned with economics and society), how could one oppose with any semblance of legitimacy 'development' and 'system', movement and circular regulations, or, as it is sometimes put crudely and unthinkingly, 'history' and 'structure'?

The same conservative function is at work in the theme of cultural totalities (for which Marx has been criticised, then travestied), in the theme of a search for origins (which was opposed to Nietzsche, before an attempt was made to transpose him into it), and in the theme of a living, continuous, open history. The cry goes up that one is murdering history whenever, in a historical analysis - and especially if it is concerned with thought, ideas, or knowledge - one is seen to be using in too obvious a way the categories of discontinuity and difference, the notions of threshold, rupture and transformation, the description of series and limits. One will be denounced for attacking the inalienable rights of history and the very foundations of any possible historicity. But one must not be deceived: what is being bewailed with such vehemence is not the disappearance of history, but the eclipse of that form of history that was secretly, but entirely related to the synthetic activity of the subject-, what is being bewailed ' is the 'development' (devenir) that was to provide the sovereignty of the consciousness with a safer, less exposed shelter than myths kinship systems, languages, sexuality, or desire; what is being bewailed is the possibility of reanimating through the project, the work of meaning, or the movement of totalisation, the interplay of material determinations, rules of practice, unconscious systems, rigorous but unreflected relations, correlations that elude all lived experience; what is being bewailed, is that ideological use of history by which one tries to restore to man everything that has unceasingly eluded him for over a hundred years. All the treasure of bygone days was crammed into the old citadel of this history; it was thought to be secure; it was secularised; it was made the last resting-place of anthropological thought; it was even thought that its most inveterate enemies could be captured and turned into vigilant guardians. But the historians had long ago deserted the old fortress and gone to work elsewhere; it was realised that neither Marx nor Nietzsche were carrying out the guard duties that had been entrusted to them. They could not be depended on to preserve privilege; nor to affirm once and for all - and God knows it is needed in the distress of today - that history, at least, is living and continuous, that it is, for the subject in question, a place of rest, certainty, reconciliation, a place of tranquillised sleep.

At this point there emerges an enterprise of which my earlier books Histoire de la role (Madness and Civilisation), Naissance de la clinique, and Les Mots et les choses (The Order of Things) were a very imperfect sketch. An enterprise by which one tries to measure the mutations that operate in general in the field of history; an enterprise in which the methods, limits, and themes proper to the history of ideas are questioned; an enterprise by which one tries to throw off the last anthropological constraints; an enterprise that wishes, in return, to reveal how these constraints could come about. These tasks were outlined in a rather disordered way, and their general articulation was never clearly defined. It was time that they were given greater coherence - or, at least, that an attempt was made to do so. This book is the result.

In order to avoid misunderstanding, I should like to begin with a few observations.

- My aim is not to transfer to the field of history, and more particularly to the history of knowledge (connaissances), a structuralist method that has proved valuable in other fields of analysis. My aim is to uncover the principles and consequences of an autochthonous transformation that is taking place in the field of historical knowledge. It may well be that this, transformation, the problems that it raises, the tools that it uses, the concepts that emerge from it, and the results that it obtains are not entirely foreign to what is called structural analysis. But this kind of analysis is not specifically used;

- my aim is most decidedly not to use the categories of cultural totalities (whether world-views, ideal types, the particular spirit of an age) in order to impose on history, despite itself, the forms of structural analysis. The series described, the limits fixed, the comparisons and correlations made are based not on the old philosophies of history, but are intended to question teleologies and totalisations;

- in so far as my aim is to define a method of historical analysis freed from the anthropological theme, it is clear that the theory that I am about to outline has a dual relation with the previous studies. It is an attempt to formulate, in general terms (and not without a great deal of rectification and elaboration), the tools that these studies have used or forged for themselves in the course of their work. But, on the other hand, it uses the results already obtained to define a method of analysis purged of all anthropologism. The ground on which it rests is the one that it has itself discovered. The studies of madness and the beginnings of psychology, of illness and the beginnings of a clinical medicine, of the sciences of life, language, and economics were attempts that were carried out, to some extent, in the dark: but they gradually became clear, not only because little by little their method became more precise, but also because they discovered - in this debate on humanism and anthropology - the point of its historical possibility.

In short, this book, like those that preceded it, does not belong - at least directly, or in the first instance - to the debate on structure (as opposed to genesis, history, development); it belongs to that field in which the questions of the human being, consciousness, origin, and the subject emerge, intersect, mingle, and separate off. But it would probably not be incorrect to say that the problem of structure arose there too.

This work is not an exact description of what can be read in Madness and Civilisation, Naissance de la clinique, or The Order of Things. It is different on a great many points. It also includes a number of corrections and internal criticisms. Generally speaking, Madness and Civilisation accorded far too great a place, and a very enigmatic one too, to what I called an 'experiment', thus showing to what extent one was still close to admitting an anonymous and general subject of history; in Naissance de la clinique, the frequent recourse to structural analysis threatened to bypass the specificity of the problem presented, and the level proper to archaeology; lastly, in The Order of Things, the absence of methodological signposting may have given the impression that my analyses were being conducted in terms of cultural totality. It is mortifying that I was unable to avoid these dangers: I console myself with the thought that they were intrinsic to the enterprise itself, since, in order to carry out its task, it had first to free itself from these various methods and forms of history; moreover, without the questions that I was asked,' without the difficulties that arose, without the objections that were made, I may never have gained so clear a view of the enterprise to which I am now inextricably linked. Hence the cautious, stumbling manner of this text: at every turn, it stands back, measures up what is before it, gropes towards its limits, stumbles against what it does not mean, and digs pits to mark out its own path. At every turn, it denounces any possible confusion. It rejects its identity, without previously stating: I am neither this nor that. It is not critical, most of the time; it is not a way of saying that everyone else ' is wrong. It is an attempt to define a particular site by the exteriority of its vicinity; rather than trying to reduce others to silence, by claiming that what they say is worthless, I have tried to define this blank space from which I speak, and which is slowly taking shape in a discourse that I still feel to be so precarious and so unsure.

'Aren't you sure of what you're saying? Are you going to change yet again, shift your position according to the questions that are put to you, and say that the objections are not really directed at the place from which you, are speaking? Are you going to declare yet again that you have never been what you have been reproached with being? Are you already preparing the way out that will enable you in your next book to spring up somewhere else and declare as you're now doing: no, no, I'm not where you are lying in wait for me, but over here, laughing at you?'

'What, do you imagine that I would take so much trouble and so much pleasure in writing, do you think that I would keep so persistently to my task, if I were not preparing - with a rather shaky hand - a labyrinth into which I can venture, in which I can move my discourse, opening up underground passages, forcing it to go far from itself, finding overhangs that reduce and deform its itinerary, in which I can lose myself and appear at last to eyes that I will never have to meet again. I am no doubt not the only one who writes in order to have no face. Do not ask who I am and do not ask me to remain the same: leave it to our bureaucrats and our police to see that our papers are in order. At least spare us their morality when we write.'

Source: The Archaeology of Knowledge (1969), publ. Routledge, 1972. Introduction, by Foucault.