The Archeology of Knowledge Michel Foucault contents part I introduction 3

2. THE ORIGINAL And THE REGULAR

In general, the history of ideas deals with the field of discourses as a domain with two values; any element located there may be character­ized as old or new; traditional or original; conforming to an average type or deviant. One can distinguish. therefore between two categories of formulation: those that are highly valued and relatively rare, which appear for the first time, which have no similar antecedents, which may serve as models for others, and which to this extent deserve to be regarded as creations; and those, ordinary, everyday, solid, that are not responsible for themselves, and which. derive, sometimes going so far as to repeat it word for word, from what has already been said. To each of these two groups the history of ideas gives a status; and it does not subject them to the same analysis: in describing the first, it recounts the history of inventions, changes, transformations, it shows how truth freed itself from error, how consciousness awoke from its successive slumbers, how new forms rose up i.n turn to produce the landscape that we know today; it is the task of the historian to rediscover on the basis of these isolated points, these successive ruptures, the continuous line of an evolution. The second group, on the other hand, reveals history as inertia and weight, as a slow accumulation of the past, a silent

((158))

sedimentation of things said; in this second group, statements must he treated by weight and in accordance with what they have in common; their unique occurrence may be neutralized; the importance of their author^'s identity, the time and place of their appearance are also dimin­ished; on the other hand, it is their extent that must be measured; the extent of their repetition in time and place, the channels by which they are diffused, the groups in which they circulate; the general horizon that they outline for men^'s thought, the limits that they impose on it; and how, in characterizing a period, they make it possible to distinguish it from others; one then describes a series of overall figures. In the first case, the history of ideas describes a succession of events in thought; in the second, there are uninterrupted expanses of effects; in the first, one reconstitutes the emergence of truths of forms; in the second, one re-establishes forgotten solidities, and refers discourses to their relativity.

It is true that, between these two authorities, the history of ideas is continuously determining relations; neither analysis is ever found in its pure state; it describes conflicts between the old and the new, the resistance of the acquired, the repression that it exercises over what has so far never been said, the coverings by which it masks it, the oblivion to which it sometimes succeeds in confining it; but it also describes the conditions, which, obscurely and at a distance, will facilitate the emer­gence of future discourses; it describes the repercussions of discover­ies, the speed and extent of their diffusion, the slow processes of replacement or the sudden upheavals that overthrow familiar language (langage); it describes the integration of the new in the already struc­tured field of the acquired, the progressive fall from the original into the traditional, or, again, the reappearances of the already-said, and the uncovering of the original. But this intersection does not prevent it from always maintaining a bipolar analysis of the old and the new. An analysis that reinvests in the empirical element of history, and in each of its stages, the problematic of the origin: in every æuvre, in every hook, in the smallest text, the problem is to rediscover the point of rupture, to establish, with the greatest possible precision, the division between the implicit density of the already-said, a perhaps involuntary fidelity to acquired opinion, the law of discursive fatalities, and the vivacity of creation, the leap into irreducible difference. Although this description

((159))

of originalities may seem obvious enough, it poses two very different methodological problems; that of resemblance and that of procession. It presupposes, in effect, that one can establish a sort of single, great series in which every formulation would assume a date in accordance with homogeneous chronological guide-lines. But, to examine the question more closely, does Grimm, with his law of vowel-gradations, precede Bopp (who quoted him, used him, applied and modified what he said) in the same way and on the same temporal line; and did Cceurdoux and Anquetil-Duperron (in observing analogies between Greek and Sanskrit) anticipate the definition of the Indo-European lan­guages, and precede the founders of comparative grammar? Was Saus-sure ^'preceded^' by Peirce and his semiotics, by Arnauld and Lancelot with the Classical analysis of the sign, and by the Stoics and the theory of the ^'signifier^', in the same series and in accordance with the same mode of anteriority? Precession is not an irreducible and primary donnee; it cannot play the role of an absolute measure that makes it possible to gauge all discourse and to distinguish the original from the repetitive. The mapping of antecedents is not enough, in itself, to determine a discursive order; on the contrary, it is subordinated to the discourse that one is analysing, at the level that one chooses, on the scale that one establishes. By deploying discourse throughout a calen­dar, and by giving a date to each of its elements, one does not obtain a definitive hierarchy of precessions and originalities; this hierarchy is never more than relative to the systems of discourse that it sets out to evaluate.

Similarly, the resemblance between two or several successive formu­lations also poses a whole series of problems. In what sense and in accordance with what criteria can one affirm: ^'this has been said^'; ^'the same thing can already he found in this or that text^', etc.? What is identity, partial or total, in the order of discourse? The fact that two enunciations are exactly identical, that they are made up of the same words used with the same meaning, does not, as we know, mean that they are absolutely identical. Even when one finds, in the work of Diderot and Lamarck, or of Benoit de Maillet and Darwin, the same formulation of the principle of evolution, one cannot consider that one is dealing in each case with the same discursive event, which has been subjected at different times to a series of repetitions. Identity is not a

((160))

criterion even when it is exhaustive; even less so when it is partial, when words are not used each time in the same sense, or when the same nucleus of meaning is apprehended through different words: to what extent can one affirm that it is the same organicist theme that emerges in the so very different discourses and vocabularies of Buffon, Jussieu, and Cuvier? And, inversely, can one say that the word ^'organ­ization^' has the same meaning in the work of Daubenton, Blumenbach, and Geoffroy Saint-Hilaire? Generally speaking, does one find the same type of resemblance between Cuvier and Darwin, and between Cuvier and Linnaeus (or Aristotle)? There is no immediately recognizable resemblance between the formulations: their analogy is an effect of the discursive field in which it is mapped.

It is not legitimate, then, to demand, point-blank, of the texts that one is studying their title to originality, and whether they really possess those degrees of nobility that are measured here by the absence of ancestors. The question can have meaning only in very precisely defined series, in groups whose limits and domain have been estab­lished, between guide-lines that delimit sufficiently homogeneous dis-cursive fields.^' But to seek in the great accumulation of the already-said the text that resembles ^'in advance^' a later text, to ransack history in order to rediscover the play of anticipations or echoes, to go right back to the first seeds or to go forward to the last traces, to reveal in a work its fidelity to tradition or its irreducible uniqueness, to raise or lower its stock of originality, to say that the Port-Royal grammarians invented nothing, or to discover that Cuvier had more predecessors than one thought, these are harmless enough amusements for historians who refuse to grow up.

Archaeological description is concerned with those discursive prac­tices to which the facts of succession must be referred if one is not to establish them in an unsystematic and naive way, that is in terms of merit. At the level in which they are, the originality/banality oppos­ition is therefore not relevant: between an initial formulation and the sentence, which, years, centuries later, repeats it more or less exactly, it establishes no hierarchy of value; it makes no radical difference. It tries

It was in this way that M. Canguilhem established the series of propositions which, from Willis to Prochaska, made possible the definition of the reflex.

((161))

only to establish the regularity of statements. In this sense, regularity is not in opposition to irregularity, which, in the margins of current opinion or the most frequent texts, characterizes the deviant statement (abnormal, prophetic, retarded, pathological, or the product of genius); it designates, for every verbal performance (extraordinary, or banal, unique in its own kind or endlessly repeated), the set of conditions in which the enunciative function operates, and which guarantees and defines its existence. In this sense, regularity does not characterize a certain central position between the ends of a statistical curve — it is not valid therefore as an index of frequency or probability; it specifies an effective field of appearance. Every statement bears a certain regularity and it cannot be dissociated from it. One must not therefore oppose the regularity of a statement with the irregularity of another (that may he less expected, more unique, richer in. innov­ation), but to other regularities that characterize other statements.

Archaeology is not in search of inventions; and it remains unmoved at the moment (a very moving one, I admit) when for the first time someone was sure of some truth; it does not try to restore the light of those joyful mornings. But neither is it concerned with the average phenomena of opinion, with the dull grey of what everyone at a par­ticular period might repeat. What it seeks in the texts of Linnaeus or Buffon, Petty or Ricardo, Pinel or Bichat, is not to draw up a list of founding saints; it is to uncover the regularity of a discursive practice. A practice that is in operation, in the same way, in the work of their predecessors; a practice that takes account in their work not only of the most original affirmations (those that no one else dreamt of before them), but also of those that they borrowed, even copied, from their predecessors. A discovery is no less regular, from the enunciative point of view, than the text that repeats an.d diffuses it; regularity is no less operant, no less effective and active, in a banal as in a unique forma-tion. In. such a description one cannot admit a difference in nature between creative statements (which reveal something new, which emit hitherto unknown information, and which are 'active^' in the same way) and imitative statements (which receive and repeat information, an.d remain, as it were, ^'passive^'). The field of statements is not a group of inert areas broken up by fecund moments; it is a domain that is active throughout.

((162))

This analysis of enunciative regularities opens up in several directions that one day perhaps will be explored with greater care.
1. A group of statements is characterized, then, by a certain form of regularity, without it being either necessary or possible to dis­tinguish between what is new and what is not. But these regularities — we shall come hack to them later — are not given once and for all; the same regularity is not to be found at work in Tournefort and Darwin, Lancelot and Saussure, Petty and Keynes. There are, then, homogeneous fields of enunciative regularities (they characterize a discursive forma-tion), but these fields are different from one another. The movement from one field of enunciative regularities to another need not be accompanied by corresponding changes at all other levels of discourse. There are verbal performances that are identical from the point of view of grammar (vocabulary, syntax, and the language (langue) in general); that are also identical. from the point of view of logic (from the point of view of propositional structure, or of the deductive system in which it is placed) ; but which. are enunciatively different. Thus the formation of the quantitative relation between prices and monetary mass in circula­tion may be expressed in the same words — or synonymous words — and be obtained by the same reasoning; but it is not enunciatively identical in Gresham or Locke and the nineteenth-century marginalists; it does not belong in each case to the same system of formation of objects and concepts. We must distinguish, then, between linguistic ana-logy (or translatability), logical identity (or equivalence), and enunciative homogeneity. It is with these homogeneities and these alone that archae­ology is concerned. It can see the appearance of a new discursive practice through verbal formulations that remain linguistically analo­gous or logically equivalent (by taking up again, sometimes word for word, the old theory of sentence-attribution and verb-copula the Port-Royal grammarians opened up an enunciative regularity whose speci­ficity it is the duty of archaeology to describe). Inversely, it may ignore differences of vocabulary, it may pass over semantic fields or different deductive organizations, if it is capable of recognizing in each case, despite their heterogeneity, a certain enunciative regularity (from this point of view, the theory of the language (langage) of action, the search for the origin of languages (longues), the establishment of primitive

((163))

roots, as they are found in the eighteenth century, are not ^'new^' in relation to Lancelot^'s ^'logical^' analyses)

One can see the emergence therefore of a number of disconnexions and articulations. One can no longer say that a discovery, the formula-tion of a general principle, or the definition of a project, inaugurates, in a massive way, a new phase in the history of discourse. One no longer has to seek that point of absolute origin or total revolution on the basis of which everything is organized, everything becomes possible and necessary, everything is effaced in order to begin again. One is deal-ing with. events of different types and levels, caught up in distinct historical webs; the establishment of an enunciative homogeneity in no way implies that, for decades or centuries to come, men will say and think the same thing; nor does it imply the definition, explicit or not, of a number of principles from which everything else would flow, as inevitable consequences. Enunciative homogeneities (and heterogeneities) intersect with linguistic continuities (and changes), with logical identities (and differences), without any of them pro­ceeding at the same pace or necessarily affecting one another. But there must exist between them a number of relations and inter-dependences whose no doubt highly complex domain must be described.

2. Another direction of research: the interior hierarchies within enunciative regularities. We have seen that every statement belongs to a certain regularity — that consequently none can be regarded as pure creation, as the marvellous disorder of genius. But we have also seen that no statement can be regarded as inactive, and be valid as the scarcely real shadow or transfer of the initial statement. The whole enunciative field is both regular and alerted: it never sleeps; the least statement — the most discreet or the most banal — puts into operation a whole set of rules in accordance with which its object, its modality, the concepts that it employs, and the strategy of which it is a part, are formed. These rules are never given in a formulation, they ^'traverse^' formulations, and set up for them a space of coexistence; one cannot therefore rediscover the unique statement that would articulate them for themselves. However, certain groups of statements put these rules into operation in their most general and most widely applicable form;
((164))

using them as a starting-point, one can see how other objects, other concepts, other enunciative modalities, or other strategic choices may be formed on the basis of rules that are less general and whose domain of application is more specified. One can thus describe a tree of enun­ciative derivation: at its base are the statements that put into operation rules of formation in their most extended form; at its summit, and after a number of branchings, are the statements that put into operation the same regularity, but one more delicately articulated, more clearly delimited and localized in its extension.

Archaeology — and this is one of its principal themes — may thus constitute the tree of derivation of a discourse. That of Natural History, for example. It will place at the root, as governing statements, those that concern the definition of observable structures and the field of possible objects, those that prescribe the forms of description and the per­ceptual codes that it can use, those that reveal the most general possibilities of characterization, and thus open up a whole domain of concepts to be constructed, and, lastly, those that, while constituting a strategic choice, leave room for the greatest number of subsequent options. And it will find, at the ends of the branches, or at various places in the whole, a burgeoning of ^`discoveries^' (like that of fossil series), conceptual transformations (like the new definition of the genus), the emergence of new notions (like that of mammals or organ-ism), technical improvements (principles for organizing collections, methods of classification and nomenclature). This derivation from governing statements must not be confused with a deduction that is made on the basis of axioms; nor must it be identified with the ger­mination of a general idea, or a philosophical nucleus whose signifi­cance emerges gradually in experience or precise conceptualizations; lastly, it must not he taken as a psychological genesis based on a dis­covery whose consequences and possibilities gradually develop and unfold. It is different from all these courses, and it must be described in its autonomy. One can thus describe the archaeological derivations of Natural History without beginning with its undemonstrable axioms or its fundamental themes (the continuity of nature, for example), and without taking as one^'s starting-point and guiding-thread the first dis­coveries or the first approaches (those of Tournefort before those of Linnaeus, those of Jonston before those of Tournefort). The archaeo-
((165))

logical order is neither that of systematici ties, nor that of chronological successions.

But one can see that a whole domain of possible questions is open-ing up here. For these different orders cannot be specific and autono­mous; there must be relations and dependences between them. For certain discursive formations, the archaeological order is perhaps not very different from the systematic order, as in other cases it may follow the thread of chronological successions. These parallelisms (contrary to the distortions met with elsewhere) are worthy of analysis. In any case, it is important not to confuse these different orders, not to seek in an ^'initial^' discovery or in the originality of a formulation the principle from which everything can be deduced and derived; not to seek in a general principle the law of enunciative regularities or individual inventions; not to demand of archaeological derivation that it reproduce the order of time or reveal a deductive schema.
Nothing would be more false than to see in the analysis of dis-cursive formations an attempt at totalitarian periodization, whereby from a certain moment and for a certain time, everyone would think in the same way, in spite of surface differences, say the same thing, through a polymorphous vocabulary, and produce a sort of great dis-course that one could travel over in any direction. On the contrary, archaeology describes a level of enunciative homogeneity that has its own temporal articulations, and which does not carry with it all the other forms of identity and difference that are to be found in language; and at this level, it establishes an order, hierarchies, a whole burgeon-ing that excludes a massive, amorphous synchrony, given totally once and for all. In those confused unities that we call ^'periods^', it reveals, with all their specificity, ^'enunciative periods^' that are articulated, but without being confused with them, upon the time of concepts, on theoretical phases, on stages of formalization and of linguistic development.

((166))