The Archeology of Knowledge Michel Foucault contents part I introduction 3

3. THE DESCRIPTION OF STATEMENTS

I now find that the analysis has shifted its ground to a quite consider-able extent; it was my intention to return to the definition of the statement, which, at the outset, I had left in suspense. It was as if I had regarded the statement as a unit that could be established without difficulty, and that all I had to do was describe its possibilities and laws of combination. I now realize that I could not define the statement as a unit of a linguistic type (superior to the phenomenon of the word, inferior to the text); but that I was dealing with an enunciative function that involved various units (these may sometimes be sentences, some-times propositions; but they are sometimes made up of fragments of sentences, series or tables of signs, a set of propositions or equivalent formulations); and, instead of giving a ^'meaning^' to these units, this function relates them to a field of objects; instead of providing them with a subject, it opens up for them a number of possible subjective positions; instead of fixing their limits, it places them in a domain of coordination and coexistence; instead of determining their identity, it places them in a space in which they are used and repeated. In short, what has been discovered is not the atomic statement — with its apparent meaning, its origin, its limits, and its individuality — but the

((120))

operational field of the enunciative function and the conditions accord-ing to which it reveals various units (which may be, but need not be, of a grammatical or logical order). But I now feel that I must answer two questions: what do I now understand by the task, which I originally set myself, of describing statements? How can this theory of the statement be adjusted to the analysis of discursive formations that I outlined previously?

I. First task: fix the vocabulary. If we agree to call verbal performance, or, better, linguistic performance, any group of signs produced on the basis of a natural (or artificial) language (langue), we could call formulation the individual (or possibly collective) act that reveals, on any material and according to a particular form, that group of signs: the formulation is an event that can always be located by its spatio-temporal coordinates, which can always be related to an author, and which may constitute in itself a specific act (a ^'performative^' act, as the British analysts call it); we can call sentence or proposition the units that grammar or logic may recognize in a group of signs: these units may always be characterized by the elements that figure in them, and by the rules of construction that unite them; in relation to the sentence and the proposition, the questions of origin, time and place, and context are merely subsidiary; the decisive question is that of their correctness (if only under the form of ^'acceptability^'). We will call statement the modality of existence proper to that group of signs: a modality that allows it to be something more than a series of traces, something more than a succession of marks on a substance, something more than a mere object made by a human being; a modality that allows it to be in relation with a domain of objects, to prescribe a definite position to any possible subject, to be situated among other verbal performances, and to be endowed with a repeatable materiality. We can now understand the reason for the equivocal meaning of the term discourse, which I have used and abused in many different senses: in the most general, and vaguest way, it denoted a group of verbal performances; and by discourse, then, I meant that which was produced (perhaps all that was produced) by the groups of signs. But I also meant a group of acts of formulation, a series

((121))

of sentences or propositions. Lastly — and it is this meaning that was finally used (together with the first, which served in a provisional capacity) — discourse is constituted by a group of sequences of signs, in so far as they are statements, that is, i.n so far as they can be assigned particular modalities of existence. And if I succeed in showing, as I shall try to do shortly, that the law of such a series is precisely what I have so far called a discursive formation, if I succeed in showing that this discursive formation really is the principle of dispersion and redistribu­tion, not of formulations, not of sentences, not of propositions, but of statements (in the sense in which I have used this word), the term discourse can he defined as the group of statements that belong to a single system of formation; thus I shall be able to speak of clinical discourse, economic discourse, the discourse of natural history, psychiatric discourse.

I am well aware that most of these definitions do not conform with current usage: linguists usually give the word discourse a quite different meaning; logicians and analysts use the term statement in a different way. But my intention here is not to transfer to some hitherto benighted domain. a set of concepts, a form of analysis, and a theory that have been formed elsewhere; and I do not intend to use a model by applying it, with its own efficacy, to new contents. Not, of course, that I wish to question the value of such a model; not that I wish, even before trying it, to limit its application, or to lay down the threshold that it must not cross. But I would like to reveal a descriptive possibility, outline the domain of which. it is capable, define its limits and its autonomy. This descriptive possibility is articulated upon others; i.t does not derive from them.

In particular, then, the analysis of statements does not claim to he a total, exhaustive description of ^'language^' (langage), or of ^'what was said^'. In the whole density implied by verbal performances, it is situ­ated at a particular level that must be distinguished from the others, characterized in relation to them, and abstract. In particular, it does not replace a logical analysis of propositions, a grammatical analysis of sentences, a psychological or contextual analysis of formulations: it is another way of attacking verbal performances, of dissociating their complexity, of isolating the terms that are entangled in its web, and of locating the various regularities that they obey. By confronting the

((122))

statement with the sentence or the proposition, I am not trying to rediscover a lost totality, or to resuscitate, as many would nostalgically like to do, the plenitude of living speech, the richness of the Word, the profound unity of the Logos. The analysis of statements corresponds to a specific level of description.

2. The statement, then, is not an elementary unity that can be added to the unities described by grammar or logic. It cannot be isolated like a sentence, a proposition, or an act of formulation. To describe a state-ment is not a matter of isolating and characterizing a horizontal seg-ment; but of defining the conditions in which the function that gave a series of signs (a series that is not necessarily grammatical or logically structured) an existence, and a specific existence, can operate. An exist­ence that reveals such a series as more than a mere trace, but rather a relation to a domain of objects; as more than the result of an action or an individual operation, but rather a set of possible positions for a subject; as more than an organic, autonomous whole, closed in upon itself and capable of forming meaning of its own accord, but rather an element in a field of coexistence; as more than a passing event or an inert object, but rather a repeatable materiality. The description of statements is concerned, in a sort of vertical dimension, with the con­ditions of existence of different groups of signifiers (signifiants). Hence a paradox: the description of statements does not attempt to evade verbal performances in order to discover behind them or below their appar­ent surface a hidden element, a secret meaning that lies buried within them, or which emerges through them without saying so; and yet the statement is not immediately visible; it is not given in such a manifest way as a grammatical or logical structure (even if such a structure is not entirely clear, or is very difficult to elucidate). The statement is neither visible nor hidden.

Not hidden, by definition, since it characterizes the modalities of existence proper to a group of effectively produced signs. The analysis of statements can never confine its attention to the things said, to the sentences that were actually spoken or written, to the ^'signifying^' elements that were traced or pronounced — and, more particularly, to that very uniqueness that gives them existence, offers them to the view of the reader, to a possible reactivation, to innumerable uses or possible

((123))

transformations, among other things, but not like other things. It can-not concern only realized verbal performances since it analyses them at the level of their existence: it is a description of things said, precisely as they were said. The analysis of statements, then, is a historical analysis, but one that avoids all interpretation: it does not question things said as to what they are hiding, what they were ^'really^' saying, in spite of themselves, the unspoken element that they contain, the proliferation of thoughts, images, or fantasies that inhabit them; but, on the con­trary, it questions them as to their mode of existence, what it means to them to have come into existence, to have left traces, and perhaps to remain there, awaiting the moment when they might be of use once more; what it means to them to have appeared when and where they did — they and no others. From this point of view, there is no such thing as a latent statement: for what one is concerned with is the fact of language (langage).

A difficult thesis to sustain. We know — and this has probably been the case ever since men began to speak — that one thing is often said in place of another; that one sentence may have two meanings at once; that an obvious meaning, understood without difficulty by everyone, may conceal a second esoteric or prophetic meaning that a more subtle deciphering, or perhaps only the erosion of time, will finally reveal; that beneath a visible formulation, there may reign another that con­trols it, disturbs it, and imposes on it an articulation of its own; in short, that in one way or another, things said say more than themselves. But, in fact, these apparent duplications, this unsaid that is nevertheless said, do not affect the statement, at least as it has been defined here. Polysemia — which justifies hermeneutics and the discovery of another meaning — concerns the sentence, and the semantic fields that it employs: the same group of words may give rise to several meanings, and to several possible constructions; there may be, therefore, inter-woven or alternating, different meanings operating on the same enun­ciative base. Similarly, the suppression of one verbal performance by another, their substitution or interference, are phenomena that belong to the level of the formulation (even if they have incidences on the linguistic or logical structures); but the statement itself is not con­cerned with this duplication or this suppression: since it is the modal­ity of existence of the verbal performance as it has taken place. The

((124))

statement cannot be regarded as the cumulative result or the crystalliza­tion of several fluctuating, scarcely articulated, and mutually opposed statements. The statement is not haunted by the secret presence of the unsaid, of hidden meanings, of suppressions; on the contrary, the way in which these hidden elements function, and in which they can be restored, depends on the enunciative modality itself: we know that the ^'unsaid^', the ^'suppressed^', is not the same — either in its structure or in its effect — in the case of a mathematical statement, a statement in economics, an autobiography, or the account of a dream.

However, to all these various modalities of the unsaid that may be located against the background of the enunciative field, should no doubt be added a lack, which, instead of being inside seems to be correlative with this field and to play a role in the determination of its very existence. There may in fact be — and probably always are — in the conditions of emergence of statements, exclusions, limits, or gaps that divide up their referential, validate only one series of modalities, enclose groups of co-existence, and prevent certain forms of use. But one should not confuse, either in its status or in its effect, the lack that is characteristic of an enunciative regularity and the meanings concealed in what is formulated in it.

3. Although the statement cannot be hidden, it is not visible either; it is not presented to the perception as the manifest bearer of its limits and characteristics. It requires a certain change of viewpoint and attitude to be recognized and examined in itself. Perhaps it is like the over-familiar that constantly eludes one; those familiar transparencies, which, although they conceal nothing in their density, are nevertheless not entirely clear. The enunciative level emerges in. its very proximity.

There are several reasons for this. The first has already been given: the statement is not just another unity — above or below — sentences and propositions; it is always invested in unities of this kind, or even in sequences of signs that do not obey their laws (and which may be lists, chance series, tables); it characterizes not what is given in them, but the very fact that they are given, and the way in which they are given. It has this quasi-invisibility of the ^'there is^', which is effaced in the very thing of which one can say: ^'there is this or that thing^'.

Another reason: the ^'signifying^' structure of language (langage)

((125))

always refers back to something else; objects are designated by it; meaning is intended by it; the subject is referred back to it by a number of signs even if he is not himself present in them. Language always seems to be inhabited by the other, the elsewhere, the distant; it is hollowed by absence. Is it not the locus in which something other than itself appears, does not its own existence seem to be dissipated in this function? But if one wishes to describe the enunciative level, one must consider that existence itself; question language, not in the direction to which it refers, but in the dimension that gives it; ignore its power to designate, to name, to show, to reveal, to be the place of meaning or truth, and, instead, turn one^'s attention to the moment — which is at once solidified, caught up in the play of the 'signifier' and the ^'signi­fied^' — that determines its unique and limited existence. In the examin­ation of language, one must suspend, not only the point of view of the ^'signified^' (we are used to this by now), but also that of the ^'signifier^', and so reveal the fact that, here and there, in relation to possible domains of objects and subjects, in relation to other possible formulations and re-uses, there is language.

The last reason for this quasi-invisibility of the statement: it is implied, but never made explicit, in all other analyses of language. If language is to he taken as an object, decomposed into distinct levels, described and analysed, an ^'enunciative datum^' must exist that will always be determined and not infinite: the analysis of a language (longue) always operates on a corpus of words and texts; the uncovering and interpretation of implicit meanings always rests on a limited group of sentences; the logical analysis of a system implies a given group of propositions in the rewriting, in a formal language (langage). The enun­ciative level is neutralized each time: either it is defined only as a representative sample that enables one to free endlessly applicable structures; or it disappears into a pure appearance behind which the truth of words is revealed; or it acts as a neutral substance that serves as a support for formal relations. The fact that, each time, it is indispens­able if an analysis is to take place deprives it of all relevance for the analysis itself. If one adds to this that all these descriptions can be made only when they themselves form finite groups of statements, it will be clear why they are surrounded on all sides by the enunciative field, why they cannot free themselves from it, and why they cannot take it

((126))

directly as its theme. In considering statements in themselves, we will not seek, beyond all these analyses and at a deeper level, some secret or some root of language (langage) that they have omitted. We shall try to render visible, and analysable, that immediate transparency that constitutes the element of their possibility.

Neither hidden, nor visible, the enunciative level is at the limit of language (langage): it is not, in itself, a group of characteristics that are presented, even in an unsystematic way, to immediate experience; but neither is it the enigmatic, silent remainder that it does not translate. It defines the modality of its appearance: its periphery rather than its internal organization, its surface rather than its content. But the fact that one can describe this enunciative surface proves that the ^'given^', the datum, of language is not the mere rending of a fundamental silence; that the words, sentences, meanings, affirmations, series of propositions do not hack directly on to a primeval night of silence; but that the sudden appearance of a sentence, the flash of meaning, the brusque gesture of the index finger of designation, always emerge in the operational domain of an enunciative function; that between lan­guage as one reads and hears it, and also as one speaks it, and the absence of any formulation, there is not a profusion of things half said, sentences left unfinished, thoughts half expressed, an endless mono­logue of which only a few fragments emerge; but, before all — or in any case before it (for it depends on them) — the conditions according to which the enunciative function operates. This also proves that it is vain to seek, beyond structural, formal, or interpretative analyses of lan­guage, a domain that is at last freed from all positivity, in which the freedom of the subject, the labour of the human being, or the opening up of a transcendental destiny could be fulfilled. One should not object to linguistic methods or logical analyses: ^'When you have said so much about the rules of its construction, what do you do with language itself, in the plenitude of its living body? What do you do with this freedom, or with this meaning that is prior to all signification, without which individuals could not understand one another in the never-ending work of language? Are you not aware that as soon as one has crossed the finite systems that make possible the infinity of discourse, but which are incapable of founding it and of accounting for it, what one finds is the mark of a transcendence, or the work of the human being?

((127))

Do you know that you have described only a few of the characteristics of a language (langage) whose emergence and mode of being are entirely irreducible to your analyses?^' Such objections must be set aside: for if it is true that there is a dimension there that belongs neither to logic nor to linguistics, it is not, for all that, a restored transcend­ence, nor a way that has been reopened in the direction of an inaccess­ible origin, nor a creation by the human being of his own meanings. Language, in its appearance and mode of being, is the statement; as such, it belongs to a description that is neither transcendental nor anthropological. The enunciative analysis does not lay down for lin­guistic or logical analyses the limit beyond which they must renounce their power and recognize their powerlessness; it does not mark the line that encloses their domain; it is deployed in another direction, which intersects them. The possibility of an enunciative analysis, if it is established, must make it possible to raise the transcendental obstacle that a certain form of philosophical discourse opposes to all analyses of language, in the name of the being of that language and of the ground from which it should derive its origin.

I must now turn to the second group of questions: how can the description of statements, thus defined, be adjusted to the analysis of discursive formations, the principles of which I outlined above? And inversely: to what extent can one say that the analysis of discursive formations really is a description of statements, in the sense in which I have used this word? It is important to answer these questions; for it is at this point that the enterprise to which I have devoted myself for so many years, which I have developed in a somewhat blind way, but of which I am now trying — even if I readjust it, even if I rectify a number of errors or imprudences — to recapture the general outline, must close its circle. As has already become clear, I am not trying to say here what I once tried to say in this or that concrete analysis, or to describe the project that I had in mind, the obstacles that I encountered, the attempts that I was forced to abandon, the more or less satisfactory results that I managed to obtain; I am not describing an effective trajec­tory in order to indicate what should have been and what will be from

((128))

now on: I am trying to elucidate in itself – in order to measure it and to determine its requirements – a possibility of description that I have used without being aware of its constraints and resources; rather than trying to discover what I said, and what I might have said, I shall try to reveal, in its own regularity – a regularity that I have not yet succeeded in mastering – what made it possible to say what I did. But one can also see that I am not developing here a theory, in the strict sense of the term: the deduction, on the basis of a number of axioms, of an abstract model applicable to an indefinite number of empirical descriptions. If such an edifice were ever possible, the time for it has certainly not yet arrived. I am not inferring the analysis of discursive formations from a definition of statements that would serve as a basis; nor am I inferring the nature of statements from what discursive formations are, as one was able to abstract them from this or that description; but I am trying to show how a domain can be organized, without flaw, without con­tradiction, without internal arbitrariness, in which statements, their principle of grouping, the great historical unities that they may form, and the methods that make it possible to describe them are all brought into question. I am not proceeding by linear deduction, but rather by concentric circles, moving sometimes towards the outer and some-times towards the inner ones: beginning with the problem of dis-continuity in discourse and of the uniqueness of the statement (the central theme), I have tried to analyse, on the periphery, certain forms of enigmatic groupings; but the principles of unification with which I was then presented, and which are neither grammatical, nor logical, nor psychological, and which consequently cannot refer either to sentences, propositions, or representations, forced me to return to the centre, to that problem of the statement; to try to elucidate what is meant by the term statement. And I will consider, not that I have constructed a rigorous theoretical model, but that I have freed a coher­ent domain of description, that I have, if not established the model, at least opened up and arranged the possibility of one, if I have been able to 'loop the loop^', and show that the analysis of discursive formations really is centred on a description of the statement in its specificity. In short, if I have been able to show that they really are the proper dimen­sions of the statement that are at work in the mapping of discursive formations. Rather than founding a theory – and perhaps before being

((129))

able to do so (I do not deny that I regret not yet having succeeded in doing so) — my present concern is to establish a possibility.

In examining the statement what we have discovered is a function that has a bearing on groups of signs, which is identified neither with grammatical ^'acceptability^' nor with logical correctness, and which requires if it is to operate: a referential (which. is not exactly a fact, a state of things, or even an object, but a principle of differentiation); a subject (not the speaking consciousness, not the author of the formula-tion, but a position that may be filled in certain conditions by various individuals) ; an associated field (which is not the real context of the formulation, the situation in which it was articulated, but a domain of coexistence for other statements); a materiality (which is not only the substance or support of the articulation, but a status, rules of transcrip­tion, possibilities of use and re-use). Now, what has been described as discursive formations are, strictly speaking, groups of statements. That is, groups of verbal performances that are not linked to one another at the sentence level by grammatical (syntactical or semantic) links; which are not linked to one another at the proposition level by logical links (links of formal coherence or conceptual connexion); and which are not linked either at the formulation level by psychological links (either the identity of the forms of consciousness, the constancy of the mental­ities, or the repetition of a project); but which are linked at the statement level. That which implies that one can define the general set of rules that govern their objects, the form of dispersion that regularly divides up what they say, the system of their referentials; that which implies that one defines the general set of rules that govern the different modes of enunciation, the possible distribution of the subjective positions, and the system that defines and prescribes them; that which implies that one defines the set of rules common to all their associated domains, the forms of succession, of simultaneity, of the repetition of which they are capable, and the system that links all these fields of co-existence together; lastly, that which implies that one can define the general set of rules that govern the status of these statements, the way in which they are institutionalized, received, used, re-used., combined together, the mode according to which they become objects of appropriation, instruments for desire or interest, elements for a strat­egy. To describe statements, to describe the enunciative function of

((130))

which they are the bearers, to analyse the conditions in which this function operates, to cover the different domains that this function presupposes and the way in which those domains are articulated, is to undertake to uncover what might be called the discursive formation. Or again, which amounts to the same thing, but in the opposite direc­tion; the discursive formation is the general enunciative system that governs a group of verbal performances — a system that is not alone in governing it, since it also obeys, and in accordance with its other dimensions, logical, linguistic, and psychological systems. What has been called ^'discursive formation^' divides up the general plane of things said at the specific level of statements. The four directions in which it is analysed (formation of objects, formation of the subjective positions, formation of concepts, formation of strategic choices) cor­respond to the four domains in which the enunciative function oper­ates. And if the discursive formations are free in relation to the great rhetorical unities of the text or the book, if they are not governed by the rigour of a deductive architecture, if they are not identified with the oeuvre of an author, it is because they bring into play the enunciative level, together with the regularities that characterize it, and not the grammatical level of sentences, or the logical level of propositions, or the psychological level of formulation.

On this basis, we can now advance a number of propositions that lie at the heart of these analyses:

1 . It can be said that the mapping of discursive formations, inde­pendently of other principles of possible unification, reveals the specific level of the statement; but it can also be said that the descrip­tion of statements and of the way in which the enunciative level is organized leads to the individualization of the discursive formations. The two approaches are equally justifiable and reversible. The analysis of the statement and that of the formation are established correlatively. When the time finally comes to found a theory, it will have to define a deductive order.
2. A statement belongs to a discursive formation as a sentence belongs to a text, and a proposition to a deductive whole. But whereas the regularity of a sentence is defined by the laws of a language (longue),

((131))

and that of a proposition by the laws of logic, the regularity of state­ments is defined by the discursive formation itself. The fact of its belonging to a discursive formation and the laws that govern it are one and the same thing; this is not paradoxical since the discursive forma-tion is characterized not by principles of construction but by a disper­sion of fact, since for statements it is not a condition of possibility but a law of coexistence, and since statements are not interchangeable elem­ents but groups characterized by their modality of existence.

3. 
   So we can now give a full meaning to the definition of ^'dis-course^' that we suggested above. We shall call discourse a group of statements in so far as they belong to the same discursive formation; it does not form a rhetorical or formal unity, endlessly repeatable, whose appearance or use in history might be indicated (and, if necessary, explained) ; it is made up of a limited number of statements for which a group of conditions of existence can be defined. Discourse in this sense is not an ideal, timeless form that also possesses a history; the problem is not therefore to ask one-self how and why it was able to emerge and become embodied at this point in time; it is, from beginning to end, historical — a fragment of history, a unity and discontinuity in history itself, posing the problem of its own limits, its divisions, its transform­ations, the specific modes of its temporality rather than its sudden irruption in the midst of the complicities of time.
   

3. 
   Lastly, what we have called ^'discursive practice^' can now be defined more precisely. It must not he confused. with the expressive operation by which can individual formulates an idea, a desire, an image; nor with the rational activity that may operate in a system of inference; nor with the ^'competence^' of a speaking subject when he constructs grammatical sentences; it is a body of anonymous, historical rules, always determined in the time and space that have defined a given period, and for a given social, economic, geographical, or linguistic area, the conditions of operation of the enunciative function.
   

It remains for me now to invert the analysis and, after referring dis-cursive formations to the statements that they describe, to seek in another direction, this time towards the exterior, the legitimate use of

((132))

these notions: what can be discovered through them, how they can take their place among other methods of description, to what extent they can modify and redistribute the domain of the history of ideas. But before operating this inversion, and in order to operate it more surely, I shall remain a little longer in the dimension that I have been exploring, and try to define what the analysis of the enunciative field and of the formations that divide it up require and exclude.

((133))