The Archeology of Knowledge Michel Foucault contents part I introduction 3

5. CHANGE And TRANSFORMATIONS

Let us now turn to the archaeological description of change. Whatever theoretical criticisms one can make of the traditional history of ideas, it does at least take as its essential theme the phenomena of temporal succession and sequence, analyses them in accordance with schemata of evolution, and thus describes the historical deployment of dis-courses. Archaeology, however, seems to treat history only to freeze it. On the one hand, by describing discursive formations, it ignores the temporal relations that may he manifested in them; it seeks general rules that will he uniformly valid, in the same way, and at every point in time: does it not, therefore, impose the constricting figure of a synchrony on a development that may he slow and imperceptible? In this ^'world of ideas^', which is in itself so untrustworthy, in which apparently the most stable figures disappear so quickly, but in which so many irregularities occur that are later accorded definitive status, in which the future always anticipates itself, whereas the past is constantly shifting, is not archaeology valid as a sort of motionless thought? And, on the other hand, when it does have recourse to chronology, it is only, it seems, in order to fix, at the limits of the positivities, two pinpoints: the moment at which they are born and the moment at which they

((184))

disappear, as if duration was used only to fix this crude calendar, and was omitted throughout the analysis itself; as if time existed only in the vacant moment of rupture, in that white, paradoxically atemporal crack in which one sudden formulation replaces another. Whether as a syn­chrony of positivities, or as an instantaneity of substitutions, time is avoided, and with it the possibility of a historical description disap­pears. Discourse is snatched from the law of development and estab­lished in a discontinuous atemporality. It is immobilized in fragments: precarious splinters of eternity. But there is nothing one can do about it: several eternities succeeding one another, a play of fixed images disappearing in turn, do not constitute either movement, time, or history.

But the problem must be examined in greater detail.

Let us take first the apparent synchrony of discursive formations. One thing is true: it is no use establishing the rules in every statement, and they cannot therefore he put into operation with every statement, they do not change each time; they can he found at work in statements or groups of statements in widely separated periods. We have seen, for example, that for nearly a century — from Tournefort to Jussieu — the various objects of Natural History obeyed the same rules of formation; we have seen that the theory of attribution is the same and plays the same role in the work of Lancelot, Condillac, and Destutt de Tracy. Moreover, we have seen that the order of statements based on archaeo­logical derivation did not necessarily reproduce the order of succes­sions: one can find in Beauzee statements that are archeologically anterior to those to be found in. the Grammaire of Port-Royal. In such an analysis, therefore, there is a suspension of temporal successions — or, to be more precise, of the calendar of formulations. But this suspension is intended precisely to reveal the relations that characterize the temporal­ity of discursive formations and articulate them in series whose intersection in no way precludes analysis.
(a) Archaeology defines the rules of formation of a group of state 

ments. In this way it shows how a succession of events may, in the same

((185))

order in which it is presented, become an object of discourse, be recorded, described, explained, elaborated into concepts, and provide the opportunity for a theoretical choice. Archaeology analyses the degree and form of permeability of a discourse: it provides the prin­ciple of its articulation over a chain of successive events; it defines the operators by which the events are transcribed into statements. It does not challenge, for example, the relation between the Analysis of Wealth and the great monetary fluctuations of the seventeenth and early eight­eenth centuries; it tries to show what, in these crises, could be given as an object of discourse, how those crises could be conceptualized in such an object, how the interests that were in conflict throughout these processes could deploy their strategy in them. Or again, it does not claim that the cholera epidemic of 1832 was not an event that con­cerned medicine: it shows how clinical discourse put into operation such a body of rules that a whole domain of medical objects could then be reorganized, that a whole group of methods of recording and nota-tion could be used, that the concept of inflammation could be aban­doned and the old theoretical problem of fevers could be resolved definitively. Archaeology does not deny the possibility of new state­ments in correlation with ^'external^' events. Its task is to show on what condition a correlation can exist between them, and what precisely it consists of (what are its limits, its form, its code, its law of possibility). It does not try to avoid that mobility of discourses that makes them move to the rhythm of events; it tries to free the level at which it is set in motion — what might he called the level of ^'evential^' engage-ment. (An engagement that is specific for every discursive formation, and which does not have the same rules, the same operators, or the same sensibility in, for example, the Analysis of Wealth and in Political Economy, in the old medicine of the ^'constitutions^' and in modern epidemiology.)

(h) Moreover, all the rules of formation assigned by archaeology to a positivity do not have the same generality: some are more specific and derive from others. This subordination may be merely hierarchical but it may also involve a temporal vector. Thus in General Grammar, the theory of the verb-attribution and that of the noun-articulation are linked to one another: and the second derives from the first, but

((186))

without it being possible to determine an order of succession between them (other than the deductive or rhetorical order that has been chosen for the expose). On the other hand, the analysis of the com­plement or the search for roots could appear (or reappear) only when the analysis of the attributive sentence or the notion of the noun as an analytic sign of representation had been developed. Another example: in the Classical period, the principle of the continuity of beings is implied in the classification of species according to structural char­acters; and in this sense they are simultaneous; on the other hand, it is only when this classification is undertaken that the lacunae and gaps may be interpreted in the categories of a history of nature, of the earth, and of the species. In other words, the archaeological ramification of the rules of formation is not a uniformly simultaneous network: there exist relations, branches, derivations that are temporally neutral; there exist others that imply a particular temporal direction. Archaeology, then, takes as its model neither a purely logical schema of simultanei­ties; nor a linear succession of events; but it tries to show the intersec­tion between necessarily successive relations and others that are not so. It does not believe, therefore, that a system of positivity is a synchronic figure that one can perceive only by suspending the whole of the diachronic process. Far from being indifferent to succession, archae­ology maps the temporal vectors of derivation.
Archaeology does not set out to treat as simultaneous what is given as successive; it does not try to freeze time and to substitute for its flux of events correlations that outline a motionless figure. What it suspends is the theme that succession is an absolute: a primary, indissociable sequence to which discourse is subjected by the law of its finitude; it is also the theme that there is in discourse only one form and only one level of succession. For these themes, it substitutes analyses that reveal both the various forms of succession that are superposed in discourse (and by forms I do not simply mean the rhythms or causes, but the series themselves), and the way in which the successions thus specified are articulated. Instead of following the thread of an original calendar, in relation to which one would establish the chronology of successive or simultaneous events, that of short or lasting processes, that of momentary or permanent phenomena, one tries to show how it is

((187))

possible for there to be succession, and at what different levels distinct successions are to be found. To constitute an archaeological history of discourse, then, one must free oneself of two models that have for so long imposed their image: the linear model of speech (and partly at least of writing), in which all events succeed one another, without any effect of coincidence and superposition; and the model of the stream of consciousness whose presence always eludes itself in its openness to the future and its retention of the past. Paradoxical as it may be, dis-cursive formations do not have the same model of historicity as the flow of consciousness or the linearity of language. Discourse, at least as analysed by archaeology, that is, at the level of its positivity, is not a consciousness that embodies its project in the external form of language (langage); it is not a language (longue), plus a subject to speak it. It is a practice that has its own forms of sequence and succession.

Archaeology is much more willing than the history of ideas to speak of discontinuities, ruptures, gaps, entirely new forms of positivity, and of sudden redistributions. The practice of political economy was, trad­itionally, to seek everything that led up to Ricardo, everything that could foreshadow his analyses, methods, and principal notions, every-thing that tended to make his discoveries more probable; the practice of the history of comparative grammar was to rediscover — beyond Bopp and Rask — earlier research into the filiation and kinship of lan­guages; it was to determine how much Anquetil-Duperron contributed towards the constitution of an Indo-European domain; it was to uncover the first comparison (made in 1769) of Sanskrit and Latin conjugations; it may even lead one back to Harris or Ramus. Archae­ology proceeds in the opposite direction: it seeks rather to untie all those knots that historians have patiently tied; it increases differences, blurs the lines of communication, and tries to make it more difficult to pass from one thing to another; it does not try to show that the Physio­cratic analysis of production foreshadowed that of Ricardo; it does not regard it as relevant to its own analyses to say that Cceurdoux foreshadowed Bopp.

What does this insistence on discontinuities correspond to? In fact, it

((188))

is paradoxical only in relation to the practice of the historians of ideas. It is rather the history of ideas — with its concern for continuities, transitions, anticipations, and foreshadowings — that plays with para­dox. From Daubenton to Cuvier, from Anquetil to Bopp, from Graslin, Turgot, or Forbonnais to Ricardo — even such a chronologically small gap — the differences are innumerable: some are localized, others are more general; some concern methods, others concepts; sometimes they concern the domain of objects, at others the whole linguistic instrument. More striking still is the example of medicine: in a quarter of a century, from 1790 to 1815, medical discourse changed more profoundly than since the seventeenth century, probably than since the Middle Ages, and perhaps even since Greek medicine: a change that revealed new objects (organic lesions, deep sites, tissular alterations, ways and forms of inter-organic diffusion, anatomoclinical signs and correlations), techniques of observation, of detection of the patho­logical site, recording; a new perceptual grid, and on almost entirely new descriptive vocabulary; new sets of concepts and nosographical distributions (century-old, sometimes age-old categories such as fever or constitution disappeared, and diseases that are perhaps as old as the world — like tuberculosis — were at last isolated and named). Those who say that archaeology invents differences in an arbitrary way can never have opened La Nosographic philosophique and the Traite des membranes. Archaeology is simply trying to take such differences ser­iously: to throw some light on the matter, to determine how they are divided up, how they are entangled with one another, how they govern or are governed by one another, to which distinct categories they belong; in short, to describe these differences, not to establish a system of differences between them. If there is a paradox in archae­ology, it is not that it increases differences, but that it refuses to reduce them — thus inverting the usual values. For the history of ideas, the appearance of difference indicates an error, or a trap; instead of examining it, the clever historian must try to reduce it: to find beneath it a smaller difference, and beneath that an even smaller one, and so on until he reaches the ideal limit, the non-difference of perfect continuity. Archaeology, on the other hand, takes as the object of its description what is usually regarded as an obstacle: its aim is not to overcome differences, but to analyse them, to say what

((189))

exactly they consist of, to differentiate them. How does this differen-tiation operate?

1. Instead of considering that discourse is made up of a series of homogeneous events (individual formulations), archaeology dis-tinguishes several possible levels of events within the very density of discourse: the level of the statements themselves in their unique emergence; the level of the appearance of objects, types of enunci­ation, concepts, strategic choices (or transformations that affect those that already exist); the level of the derivation of new rules of forma-tion on the basis of rules that are already in operation – but always in the element of a single positivity; lastly, a fourth level, at which the substitution of one discursive formation for another takes place (or the mere appearance and disappearance of a positivity). These events, which are by far the most rare, are, for archaeology, the most im­portant: only archaeology, in any case, can reveal them. But they are not the exclusive object of its description; it would be a mistake to think that they have an absolute control over all the others, and that they lead to similar, simultaneous ruptures at the different levels distinguished above. All the events that occur within the density of discourse are not immediately below one another. Of course, the appearance of a discursive formation is often correlative with a vast renewal of objects, forms of enunciation, concepts, and strategies (a principle that is not universal however: General Grammar was estab­lished in the seventeenth century without much apparent alteration in grammatical tradition); but it is not possible to determine the particu­lar concept or object that suddenly manifests its presence. One should not describe such an event, therefore, in accordance with categories that may be suitable for the emergence of a formulation, or the appear­ance of a new word. It is useless to ask of such an event questions like: ^`Who is its author? Who is speaking? In what circumstances and in what context? With what intentions, what project in mind?^' The appearance of a new positivity is not indicated by a new sentence – unexpected, surprising, logically unpredictable, stylistically deviant – that is inserted into a text, and announces either the opening of a new chapter, or the entry of a new speaker. It is an event of a quite different type.

i

((190))

2. In order to analyse such events, it is not enough simply to indi­cate changes, and to relate them immediately to the theological, aes­thetic model of creation (with its transcendence, with all its originali­ties and inventions), or to the psychological model of the act of con­sciousness (with its previous obscurity, its anticipations, its favourable circumstances, its powers of restoration), or to the biological model of evolution. We must define precisely what these changes consist of: that is, substitute for an undifferentiated reference to change — which is both a general container for all events and the abstract principle of their succession — the analysis of transformations. The disappearance of one positivity and the emergence of another implies several types of trans-formation. By going from the more particular to the more general, one can and must describe: how the different elements of a system of formation were transformed (what, for example, were the variations in the rate of unemployment and labour needs, what were the political decisions concerning the guilds and the universities, what were the new needs and new possibilities of public assistance at the end of the eighteenth century — all these were elements in the system of forma-tion of clinical medicine); how the characteristic relations of a system of formation were transformed (how, in the middle of the seventeenth century, for example, the relation between the perceptual field, the linguistic code, the use of instruments, and information that was put into operation by the discourse on living beings was modified, thus making possible the definition of the objects proper to Natural His-tory); how the relations between different rules of formation were transformed (how, for example, biology modified the order and the dependence that Natural History had established between the theory of characterization and the analysis of temporal derivations); lastly, how the relations between various positivities were transformed (how the relations between philology, biology, and economics transform the relations between General Grammar, Natural History, and the Analysis of Wealth; how the interdiscursive configuration outlined by the priv­ileged relations of these three disciplines is decomposed; how their respective relations with mathematics and philosophy are modified; how a place emerges for other discursive formations and, in particular, for that interpositivity that was later to assume the name of the human sciences). Rather than refer to the living force of change (as if it were

((191))

its own principle), rather than seek its causes (as if it were no more than a mere effect), archaeology tries to establish the system of trans-formations that constitute ^'change^'; it tries to develop this empty, abstract notion, with a view to according it the analysable status of transformation. It is understandable that some minds are so attached to all those old metaphors by which, for a century and a half, history (movement, flux, evolution) has been imagined, that they see archae­ology simply as the negation of history and the crude affirmation of discontinuity; the truth is that they cannot accept that change should be cleansed of all these adventitious models, that it should be deprived of both its primacy as a universal law and its status as a general effect, and that it should be replaced by the analysis of various transformations.
3. To say that one discursive formation is substituted for another is not to say that a whole world of absolutely new objects, enunciations, concepts, and theoretical choices emerges fully armed and fully organ­ized in a text that will place that world once and for all; it is to say that a general transformation of relations has occurred, but that it does not necessarily alter all the elements; it is to say that statements are gov­erned by new rules of formation, it is not to say that all objects or concepts, all enunciations or all theoretical choices disappear. On the contrary, one can, on the basis of these new rules, describe and analyse phenomena of continuity, return, and repetition: we must not forget that a rule of formation is neither the determination of an object, nor the characterization of a type of enunciation, nor the form or content of a concept, but the principle of their multiplicity and dispersion. One of these elements — or several of them — may remain identical (preserve the same division, the same characteristics, the same structures), yet belong to different systems of dispersion, and be governed by distinct laws of formation. One can find in such phenomena therefore: elem­ents that remain throughout several distinct positivities, their form and content remaining the same, but their formations being heterogeneous (such as monetary circulation as an object first in the Analysis of Wealth, and then in political economy; the concept of character first in Natural History, then in biology); elements that are constituted, modified, organized in one discursive formation, and which, stabilized at last, figure in another (such as the concept of reflex, which, as

((192))

G. Canguilhem has shown, was formed in Classical science from Willis to Prochaska, then entered modern physiology); elements that appear later, as an ultimate derivation in a discursive formation, and which occupy an important place in a later formation (such as the notion of organism, which appeared at the end of the eighteenth century in Natural History, and as the result of a whole taxonomic enterprise of characterization, and which became the major concept of biology at the time of Cuvier; or the notion of lesional site, which Morgagni discovered, and which became one of the principal concepts of clinical medicine); elements that reappear after a period of desuetude, obliv­ion, or even invalidation (such as the return to a Linnaean type of fixism in a biologist like Cuvier; or the reactivation in the eighteenth century of the old notion of an original language). The problem for archaeology is not to deny such phenomena, nor to try to diminish their importance; but, on the contrary, to try to describe and measure them: how can such permanences or repetitions, such long sequences or such curves projected through time exist? Archaeology does not hold the content for the primary and ultimate dormee that must account for all the rest; on the contrary, it considers that the same, the repeti­tive, and the uninterrupted are no less problematic than the ruptures; for archaeology, the identical and the continuous are not what must be found at the end of the analysis; they figure in the element of a dis-cursive practice; they too are governed by the rules of formation of positivities; far from manifesting that fundamental, reassuring inertia which we like to use as a criterion of change, they are themselves actively, regularly formed. And to those who might be tempted to criticize archaeology for concerning itself primarily with the analysis of the discontinuous, to all those agoraphobics of history and time, to all those who confuse rupture and irrationality, I will reply: It is you who devalue the continuous by the use that you make of it. You treat it as the support-element to which everything else must be related; you treat it as the primary law, the essential weight of any discursive prac­tice; you would like to analyse every modification in the field of this inertia, as one analyses every movement in the gravitational field. But in according this status to continuity, you are merely neutralizing it, driv-ing it out to the outer limit of time, towards an original passivity. Archaeology proposes to invert this arrangement, or rather (for our

((193))

aim is not to accord to the discontinuous the role formerly accorded to the continuous) to play one off against the other; to show how the continuous is formed in accordance with the same conditions and the same rules as dispersion; and how it enters — neither more nor less than differences, inventions, innovations or deviations — the field of dis-cursive practice.^'
4. The appearance and disappearance of positivities, the play of substitutions to which they give rise, do not constitute a homogeneous process that takes place everywhere in the same way. We must not imagine that rupture is a sort of great drift that carries with it all discursive formations at once: rupture is not an undifferentiated inter-val — even a momentary one — between two manifest phases; it is not a kind of lapsus without duration that separates two periods, and which deploys two heterogeneous stages on either side of a split; it is always a discontinuity specified by a number of distinct transformations, between two particular positivities. The analysis of archaeological breaks sets out, therefore, to establish, between so many different changes, analogies and differences, hierarchies, complementarities, coincidences, and shifts: in short, to describe the dispersion of the discontinuities themselves.

The idea of a single break suddenly, at a given moment, dividing all discursive formations, interrupting them in a single moment and reconstituting them in accordance with the same rules — such an idea cannot be sustained. The contemporaneity of several transformations does not mean their exact chronological coincidence: each transform­ation may have its own particular index of temporal ^`viscosity^'. Natural History, General Grammar, and the Analysis of Wealth were constituted in similar ways, and all three in the course of the seventeenth century; but the system of formation of the Analysis of Wealth was linked with a great many conditions and non-discursive practices (the circulation of goods, monetary manipulations and their effects, the system of protect-ing trade and manufactures, fluctuations in the quantity of metal coined): hence the slowness of a process that lasted for over a century (from Grammont to Cantillon), whereas the transformations that had taken place in General Grammar and Natural History had extended over scarcely more than twenty-five years. Inversely, contemporary,

((194))

similar, and linked transformations do not belong to a single model that is reproduced several times on the surface of discourses, and imposes on all a strictly identical form of rupture: when one describes the archaeological break that led to philology, biology, and economics, one is showing how these three positivities were linked (by the disap­pearance of the analysis of the sign, and of the theory of representa­tion), what symmetrical effects it could produce (the idea of a totality and of an organic adaptation among living beings; the idea of morpho­logical coherence, and of a regulated evolution in languages; the idea of a form of production that has its internal laws and its limits of devel­opment); but it also shows what were the specific differences of these transformations (how in particular historicity is introduced in a par­ticular way in these three positivities, how their relation to history cannot therefore be the same, even though they all have a particular relation with it).

Lastly, there are important shifts between different archaeological ruptures — and sometimes even between discursive formations that are very close and linked by a great many relations. Let us take the discip­lines of languages and historical analysis: the great transformation that gave rise at the beginning of the nineteenth century to a historical, comparative grammar preceded by a good half-century the mutation in historical discourse: as a result, the system of interpositivity in which philology was involved was profoundly affected in the second half of the nineteenth century, without the positivity of philology ever being put into question. Hence phenomena of ^'fragmented shift^', of which we can cite at least another famous example: concepts like those of surplus value or falling rate of profit, as found in Marx, may be described on the basis of the system of positivity that is already in operation in the work of Ricardo; but these concepts (which are new, but whose rules of formation are not) appear — in Marx himself — as belonging at the same time to a quite different discursive practice: they are formed in that discursive practice in accordance with specific laws, they occupy in it a different position, they do not figure in the same sequences: this new positivity is not a transformation of Ricardo^'s analyses; it is not a new political economy; it is a discourse that occurred around the derivation of certain economic concepts, but which, in turn, defines the conditions in which the discourse of

((195))

economists takes place, and may therefore be valid as a theory and a critique of political economy.

Archaeology disarticulates the synchrony of breaks, just as it des­troyed the abstract unity of change and event. The period is neither its basic unity, nor its horizon, nor its object: if it speaks of these things it is always in terms of particular discursive practices, and as a result of its analyses. The Classical age, which has often been mentioned in arch­aeological analyses, is not a temporal figure that imposes its unity and empty form on all discourses; it is the name that is given to a tangle of continuities and discontinuities, modifications within positivities, dis-cursive formations that appear and disappear. Similarly, rupture is not for archaeology the prop of its analyses, the limit that it indicates from afar, without being able either to determine it or to give it specificity; rup­ture is the name given to transformations that bear on the general rules of one or several discursive formations. Thus the French Revolution — since up to now all archaeological analyses have been centred on it — does not play the role of an event exterior to discourse, whose divisive effect one is under some kind of obligation to discover in all dis-courses; it functions as a complex, articulated, describable group of transformations that left a number of positivities intact, fixed for a number of others rules that are still with us, and also established posi­tivities that have recently disappeared or are still disappearing before our eyes.

((196))