October 22, 1985
The archive refers to historical formations. The archive is always the archive of a formation. This does not tell us at all what an historical formation is, nor what an archive is. And then here is Foucault telling us, in The Use of Pleasure, so, in a very late book, here is Foucault telling us: my books have been historical studies, but not an historian’s work. Historical studies and not an historian’s work.
Seminar Introduction
After Michel Foucault’s death from AIDS on June 25, 1984, Deleuze decided to devote an entire year of his seminar to a study of Foucault’s writings. Deleuze analyses in detail what he took to be the three “axes” of Foucault’s thought: knowledge, power, and subjectivation. Parts of the seminar contributed to the publication of Deleuze’s book Foucault (Paris: Minuit, 1986), which subsequently appeared in an English translation by Seán Hand (Minneapolis: University of Minnesota Press, 1988).
For archival purposes, the English translations are based on the original transcripts from Paris 8, all of which have been revised with reference to the BNF recordings available thanks to Hidenobu Suzuki, and with the generous assistance of Marc Haas.
English Translation
To introduce the first axis of Foucault’s thought, historical formations, Deleuze draws from his articles on Foucault, on The Archeology of Knowledge and Discipline and Punishment, the first two chapters of the Foucault book corresponding, respectively, to “A New Archivist” [Critique 274 (1970)] and “A New Cartographer” [Critique 343 (1975)]. Deleuze here provides an overview of the initial segment to prepare the detailed analysis. Following a bibliographical overview of Foucault’s writing, Deleuze points to Foucault’s first axis on the discipline of the archives, i.e., of historical formations, that Foucault approaches from the perspectives of “seeing” (le voir) and “speaking” (le parler). Deleuze adds several other dyads to these: the lesson of things and the lesson of grammar (that “seeing” is not “saying”), also the drawing and the text, and a third theme, the visible and the statable, i.e., visibilities and statements. Moreover, since “seeing” and “speaking” are conditions specific to each period (époque) which make behaviors and mentalities possible, Deleuze reviews how Foucault’s perspective on “seeing” and “speaking” developed from unreason (déraison) in the seventeenth century to disciplinary techniques in the eighteenth century. Deleuze emphasizes that, for Foucault, a statement regime is called a “discursivity” linked to types of evidence, as well as to non-discursive formations, distinctions that connect in turn, through the works Raymond Roussel, to a kind of “making seen”, first, in machines or mechanical processes, then in the linguistic process or the statable regime. Foucault also insists on the irreducibility of the statement and visibility, proceeding along two tracks, descriptions and statements, with the primacy given to statements over visibilities. Deleuze emphasizes Foucault’s debt to Maurice Blanchot (cf. The Infinite Conversation) and his statement “seeing is not speaking” as well as his term for the disjunction as “non-relation”. The four theses from Foucault that Deleuze defines here provide the seminar’s agenda: to understand how historical formations are constituted through the combination of regimes of statements and fields of visibility. Moreover, these forms, statements and visibilities, also constitute knowledge as the procedure without which truth has no existence.

Download
Gilles Deleuze
Seminar on Foucault
Part I: Knowledge (Historical Formations)
Lecture 01, 22 October 1985
Transcription: Annabelle Dufourcq (avec l’aide du College of Liberal Arts, Purdue University); time stamp and additional revisions, Charles J. Stivale
Translated by Mary Beth Mader
Part 1
… A version of it was published in 10/18. I alert you to the fact that this edition is abridged. What does The History of Madness concern? It concerns two things. [Pause; noises of chairs being moved, interrupting Deleuze] It’d be better to wait. Yes. Yes! Ah. [Pause]
Lucien Gouty: That guy must have had a cup of coffee!
Deleuze: Yeh, I suppose. [Interruption of the recording]
… What’s the issue? For Foucault, the issue is knowing how a mode was formed; but a mode of what? For the moment, let’s say—at the cost of surprises down the road—a mode of confining mad people. In what? In what was called at the time a general hospital or houses of correction. And this confinement of mad people, or this constitution of a general hospital that includes mad people, among others, appeared in the 17th century, that is, in the Classical Age. And where, in this same period, is medicine? Which medicine? When I say medicine, can I say psychiatry? Obviously not—psychiatry does not exist, it does not exist as a discipline. And people talked about nervous ailment or illness of the humors or illness of the head. There is no reason to say: that’s the precursor of psychiatry. It’s a branch of medicine in the 17th century. And then what Foucault studies is how the general hospital and the asylum, and medicine also, developed such that at the end of the 18th century and the start of the 19th century there comes about what is often presented as a sort of liberation of mad people, namely: releasing the chains. And what is this supposed liberation about? So, roughly—but truly superficially—those are the broad subject areas of History of Madness.
1963: a book on a poet, a poet from the start of the 20th century, Raymond Roussel. Raymond Roussel. What’s it about? It’s about an apparently peculiar body of work and this peculiar body of work seems to ground or to enclose what Roussel himself calls a language process [un procédé de langage], a language process that Roussel tried to explain in a book entitled How I Wrote Certain of My Books, and where he gives the following example: Here are two propositions, ‘the white man’s letters on the hordes of the old plunderer’ [les lettres du blanc sur les bandes du vieux pillard] and ‘the white letters on the cushions of the old billiard table’ [les lettres du blanc sur les bandes du vieux billard].[1]
Between these two propositions a whole peculiar story will unfold. And one quickly sees that in the course of his analysis, Foucault lends an essential importance to a theme found very often in Roussel—the theme of the double or doubling. The double or doubling. So. That’s how I characterize, as superficially as it may be, the . . . these books by Foucault so that those who have not read them might select the book they will take on, since, once again, I strongly advise you to pick one book if you have not read any. For those of you who are interested in this aspect, the analysis of Roussel’s poetry, for example, and in the theme of the double, you could add a later preface of Foucault’s, a preface for another inventor of peculiar language. This is a preface he wrote for the re-publication of a bizarre book, Jean-Pierre Brisset, also an author from the start of the 20th century, an inventor of a kind of language and of a very, very bizarre interpretation of language, a book entitled Grammaire logique, the Grammaire logique by Brisset, from Editions Tchou.[2] Foucault writes the preface for it and returns to Roussel and tries to analyze what he presents as processes of language. He considers three processes of language at the boundary of a peculiar literature and of what Foucault calls an uncertain madness, that is: is he crazy, is he not crazy? The three processes are Roussel’s process, whose analysis he takes up again, Brisset’s process, which he analyzes, and the process of an American, in this case, entirely current, a contemporary American named [Louis] Wolfson, who had invented a special treatment of language for himself. There you have it. So, Raymond Roussel, that’s 1963.
The same year, in 1963, Foucault published Birth of the Clinic [Naissance de la clinique], with Presses Universitaires de France. And what is Birth of the Clinic? It is the examination of two things. How illnesses are grouped into symptoms, and at the same time, to what medical statements do these symptoms refer? This is examined across two periods, the 18th and 19th centuries, the 18th century and the start of the 19th century, the period of the clinic, the birth of the clinic. You see that the theme of places is present constantly in Foucault’s corpus. The asylum, the general hospital, the clinic, and thereafter, pathological anatomy.
1966, Gallimard, The Order of Things [Les Mots et les choses]. What is The Order of Things about? Is it about words and things? Perhaps. We’ll see about all that. But it is, rather, a highly elaborate analysis of representation in the Classical Age, that is, in the 17th and 18th centuries. And then of how, at the end of the 18th and in the 19th century, representation is subjected to a critique from which forces external to representation—life, labor, and language—will emerge. That’s ’66.
1969: Archaeology of Knowledge [L’Archéologie du savoir]. Archaeology of Knowledge is a theory of statements and it’s the great theory of statements in Foucault—once we specify that Foucault gives the word ‘statement’ a sense and a status that no one before him has given it.
Fast on the heels of Archaeology of Knowledge comes Discourse on Language [L’ordre du discours] in 1971, also from Gallimard.
But 1971 also saw the publication of an article on Nietzsche, which will receive a good deal of attention, entitled “Nietzsche, Genealogy, History,” [“Nietzsche, la généalogie, l’histoire”], published in an edited volume from PUF, Presses Universitaires de France, in honor of Jean Hippolyte. [Pause]
In 1973, a small publishing house, Fata Morgana, published a very curious text by Foucault that would fit well . . . that you could group with Roussel and Brisset, and which is a commentary on a painter, [René] Magritte. It was published under the title, This is Not a Pipe [Ceci n’est pas une pipe]. Why? Because This is Not a Pipe was the title of a painting by Magritte. And the odd thing is that Magritte’s painting simply represented a pipe, very well illustrated. And set out below it in writing there was: Ceci n’est pas une pipe—that was the title of the painting. How can a painting that purely and simply represents a pipe with utter clarity be called This is Not a Pipe? Very well. This very much interests Foucault. Why? Because you understand that . . . you surely can sense that there is the problem of the relation between a drawing and a statement here. What relation is there between a drawing and a statement? The relation need not be complex, since the statement that designates the illustrated pipe immediately becomes “This is not” and not “this is.” What is happening in this turn of phrase?
1975: Discipline and Punish [Surveiller et punir]. Here again, it’s a matter of a place: the prison. And no longer the asylum. In fact, it is quite striking that fourteen years later Discipline and Punish should be composed in a way comparable to History of Madness. In History of Madness, it’s a question of a place, the asylum or the general hospital, and of a set of medical statements. In Discipline and Punish, it’s a matter of a place: the prison. How is the prison born? How is the prison formed? How does a regime that is the regime of the prison become established? And, at the same time, it’s a study of a regime of statements, the statements of penal law. In what period? In the 18th century. So, that was ’75.
1976 saw the publication of the first volume of the project Foucault pursues from then on, The History of Sexuality [L’Histoire de la sexualité]. He has a specific conception of it in ’76, when he publishes the first volume of that history under the title: The Will to Knowledge [La Volonté de savoir]. So, that’s ’76. On this point, one must observe, if you’ve been following the dates, that whereas the pace of Foucault’s books is fairly regular, there is a great silence. A great and long silence. Because the second volume of The History of Sexuality will not come out until 1984. Foucault did not conceal, he explicitly said, that he had come to redo his first plan. He had found something that led him to do a complete reworking. What happened in those years of silence? What had he found? How did he rework his project? In any case, it is in ’84 that the two subsequent volumes come out: The Use of Pleasure [L’Usage des plaisirs], in which he explains the reworking of his The History of Sexuality project and why he had come to this reworking, and volume three. Volume 2: The Use of Pleasure, ’84 and, the same year, The Care of the Self [Le Souci de soi], also ‘84. Very well, we’ll have to manage amid all of this.
One must note, since we’ll have to think about this point, that Foucault destroyed some manuscripts that were already very well underway. Notably, there is a piece on Manet, or rather there was a piece on Manet, a book on Manet. That matters to us since This is Not a Pipe is a book on Magritte and it includes several extremely interesting pages, we will see, on Paul Klee. The Order of Things opens with a famous description that is among the best-known pages of Foucault’s, and which is the description of a Velásquez painting, Las Meninas, or The Ladies in Waiting. So, that there existed, or that he had written, an already very long manuscript on Manet should interest us since we will perhaps come to the point of wondering what in fact this manuscript could have contained. And the fact is that this manuscript, it does seem that this manuscript was destroyed, that Foucault destroyed this manuscript at the end of his life. On the other hand, I think that the will is formal, and that it prohibits any publication of work posthumously. This is consequential for us, since when he stopped publishing The History of Sexuality that did not mean that there was not a fourth volume with the title Confessions of the Flesh [Les Aveux de la chair]. It explored the Church Fathers and the period of the formation of Christianity with regard to the formation of sexuality, thus an essential moment, a fundamental moment. It seems thus far that this book . . . should not be published, if it is true that the will bears this definitive prohibition. There you have it.
So, in light of this, I said all of this only so that, once again, if you have not read any Foucault, you could spot a direction in which to go. For my part, what I would like is that some of you pick Discipline and Punish. But if you don’t know Foucault at all, or very little, it would be better to pick an entire book rather than skipping from one book to another. So. With that, well, we can start. But what time is it?
Lucien Gouty: Quarter to ten.
Deleuze: Quarter to ten. Then, if it wouldn’t bother you, we’ll wait until ten o’clock. Because I’m afraid that . . . [Interruption of the recording]
… who can get here only around quarter to ten or ten o’clock. Nevertheless, I will be here anyway, I’ll be here at nine o’clock. And I would like, for those who can, to be here as well, or at least those who have reason to see me. I mean that from nine o’clock to quarter to ten will be the time when you would work the most, that is, when with respect to what we will have done in the previous meeting, we could, for those who would have questions to ask, we could expand, we could return to a particular point, etcetera. And then at quarter to ten, ten o’clock, I . . . I’d proceed. You see? So, I’ll repeat so it’s quite clear. I’ll be here at nine o’clock. Very well, there could be five or ten or fifteen of you . . . And with those of you here, we’ll go back over the previous session. And we’ll expand upon what should be expanded upon or else those who would have questions to ask me . . . or something to say, any of that, or . . . or to say: ‘No, that’s not right, we have to go back over that.’ And then at quarter to ten, from quarter to ten, we’ll hold the new session until about one o’clock. There. And like this we’ll make progress, and each week, we’ll go back for a little bit to . . . so. Ahh! That seems to get you down! Oh, well, listen, we’re starting. We’re starting.
So, I would like to clearly indicate the sectioning [découpages]. Today, I would like to start with a kind of trial-and-error process. I am extending an invitation to you. The invitation is to trust the author you study. But what does “trusting an author” mean? It means, it means the same thing as trial and error, the same thing as proceeding by a kind of trial and error. Before really understanding the problem that someone sets out, or the problems someone sets out, you’ve got to . . . I don’t know what, you’ve got to ruminate a lot. You have to classify a good deal, to re-classify, the notions the author is in the process of inventing. One must at all costs silence the voices of objection in oneself. The voices of objection are those that would too hastily say: oh, but there’s something wrong here. And to trust the author is to say to oneself, let’s not speak too quickly, let . . . you have to let the author speak. You have to let the author speak. But all this, this all consists . . . you’ve almost got to, before knowing the meaning he gives to the words, you have to carry out a kind of . . . of frequency analysis. To be sensitive to the frequency of words. To be sensitive to his style itself. To be sensitive to his own obsessions. Today, I would like to distinguish [diviser] well, so that it’s clear, because this, this isn’t simple . . . in fact, Foucault’s thought is not simple. Why? Because I believe it is a thought that invents coordinates, it is a thought that develops along axes. [Pause]
And there is one of these axes, and in my view, it is the first, it is the first that Foucault develops, in his corpus, and this first axis, he will call archaeology. And archaeology is the discipline of the archives. But what does Foucault call an archive? He will try to say this in one particular book, Archaeology of Knowledge. But as for us, we don’t want to take for granted right away what is . . . we won’t . . . we’ll almost say: so, look, it revolves around what? Foucault’s whole first period, I’d say, roughly, from The History of Sexuality to Discipline and Punish, it revolves around what? What it revolves around will let us define the archive. And there is no doubt that the archive has something to do with history. The object of the archive is the historical formation. The archives refer to historical formations. At first sight, this does not move us forward, we are spinning . . . you see, this is what I wanted to do today, we are spinning in the words. Good. The archive refers to historical formations. The archive is always the archive of a formation. This does not tell us at all what an historical formation is, nor what an archive is.
And then here is Foucault telling us, in The Use of Pleasure, so, in a very late book, here is Foucault telling us: my books have been historical studies, but not an historian’s work. Historical studies and not an historian’s work. Everyone knows that Foucault has a very close relation with the adherents of what was called the New History, roughly, the students of [Fernand] Braudel, the Annales School. But a relation may be very complex. He tells us absolutely: I am not an historian. He tells us absolutely: I am and I remain a philosopher. And nevertheless, a whole part of this corpus treats historical formations. He repeats: granted, they are historical studies, [but] they are not an historian’s works. What does he mean? He specifies a bit. He specifies a bit when he says, still in The Use of Pleasure: don’t expect a history of behaviors or of mindsets from me. Here, the allusion is clear. It is true that the Annales School, at least in part, offers us a history of behaviors and of mindsets. For example: what is a history of behaviors? Here, too, we’re touching on very broad, very cursory, signs. I think, I’m thinking of a very, very interesting book of history: Comment on meurt en Anjou au XVII siècle, au XVIII siècle. [How One Dies in 17th and 18th Century Anjou] Comment on meurt en Anjou. [How One Dies in Anjou][3] You can’t put it any better: that’s a history of behavior. I can do the history of a behavior, the behavior of death. I could also do: how one is born. How one is born in Picardy at a given point in time. You can easily see that this marshals archives. But Foucault tells us: I do not do a history of behaviors. One can imagine a history of the maternal instinct; that has been done.
In short, the field of a history of behaviors is infinite. How do people eat? How do they die? How do they marry? How are they born? How are children raised? How does one give birth? Etc., etc., and this can be either a history of behaviors or a history of mindsets. And I think that many people took Foucault’s work at its beginnings to be of this kind. This is why people aligned him so closely with the New History. However, Foucault told us absolutely: No, I have nothing to do with that. It’s not at all that he says that it’s not interesting, he says that that is not his problem. And why is it not his problem? What interests him then?
All of a sudden, we’ve got a light. We’ve got a light if . . . if you have read a bit of Foucault, or even a lot—in that case, it’s all the more so. What interests him are not the behaviors, but what? Seeing. Seeing. Foucault’s histories always revolve around seeing. You’ll say to me: ‘But seeing—what is that? All you have to do is add that to the behaviors; there are visual behaviors.’ No, not for Foucault. On this point, we’ll have to wait—it’s going to be very complicated. But for him seeing is of another order than behavior. And what else interests him? Speaking. Speaking. Okay. And speaking, one can always say: well, speaking conveys a mindset. Well, no, for Foucault. It’s even the inverse. Seeing—and you’ve already got to get used to this idea, but it will not be easy—seeing is not a behavior among other behaviors, it is the condition for any behavior in a period. Speaking is not an expression of the mindset, it is a condition for the mindset in a period. Good. In other words, in speaking to us of seeing and speaking, Foucault claims to surpass a history of behaviors and mindsets to accede to the conditions for historical behaviors and historical mindsets. What can justify such an ambition? It will be up to us to try to find out. All we’ve got is the impression that this is so. Foucault does not treat seeing and speaking as variables of behavior or of mindsets. He treats them as conditions. There is a search for the conditions of historical formation. What are these conditions for an historical formation? What is said in a period. What is seen in a period. Each period will be defined, for the time being, . . . all this will change, as we make our way . . . but we’re using very imprecise words, for the time being. It’s as if each period defines itself above all by what it sees and makes seen and by what it says. Okay. This is to say that seeing and making seen and saying are not on the same level as behaving, and as having such and such an idea. A regime of saying is the condition for all the ideas of a period. A regime of seeing is a condition for all that a period does.
Good, okay, at this point, I’ll return to my theme. You see, before one has even understood, ten or twelve objections come to mind. So, it’s now or never to say to oneself: let’s calm down! Let’s wait. It’s already quite peculiar. If you followed me here, it’s very peculiar this sort of elevation of seeing and speaking as conditions. Let’s investigate, then. Let’s investigate—after all, perhaps I’m wrong. If I’m wrong, I won’t have any confirmation. Let’s look for confirmations. Consequently, I’ll try to make a chart. On one side of my chart, I’ll put “seeing,” and on the other side, “speaking.” I’ll make a line, and I think: what comes to mind? Okay, so, I’ll fill in my chart. Seeing, speaking. And I try to fill in my chart so as to be sure, before having started, that I’ve not betrayed Foucault. Good.
And right away I come upon—I’m not going in chronological order—right away I come upon the book entitled Les Mots et les choses [The Order of Things].[4] So, you’ll say to me, “But things, aren’t they merely what is visible? Let’s wait. Words and Things, what a strange dualism. It’s not merely what is visible, no. But anyway, things, that’s what’s visible; words, that’s what’s sayable. To see, to speak. Obviously, that’s not good enough. On this point, an objection: Foucault would be the first to denounce the title. He would say, ‘they didn’t understand, they didn’t at all understand what I meant by Les Mots et les choses, because this doesn’t mean words and that doesn’t mean things.’ The title has to be taken ironically. Okay, however, at first glance, the irony escapes one. Why is Les Mots et les choses [Words and Things] ironic? Let’s wait. Words, things.
One more step: the lesson of things.[5] You know what that is, in primary school, right? In primary school, in the old days, there were two basic disciplines: the lesson of things . . . uh, are there no more chairs? It’s those others who’ve stolen them from us. Are there any there? And… [Pause; one hears the noise from a chair]
In primary school, they had that, the lesson of things, which was distinct from what? Well, from the lesson of words, the grammar lesson. Lesson of words, lesson of grammar, and those were the two heads of the primary school. Good. There was the hour on the salt marsh, they showed us a salt marsh, that is—I’m taking one step, I’m taking one step forward—an image of a salt marsh, a figure of a salt marsh, the visible salt marsh or the visible umbrella, or let’s just say it all—the visible pipe. That was the lesson of things. The teacher said: this is a pipe. This is a salt marsh. And then what came after that, the following hour, was grammar. And now, it was the order of saying and not the order of the drawing. And the order of saying is something other than the order of drawing. And if saying is something other than seeing, at that point, saying “this is a pipe” is stated necessarily as “this is not a pipe.” That is: saying is not a seeing.
The lesson of things and the lesson of grammar refer in this case to the little book, about which I spoke, by Foucault, a commentary on Magritte, Magritte’s painting being a lesson of things, an illustration [dessin appliqué] of a pipe, the title of the painting: This Is Not A Pipe. And, necessarily, “this is a pipe” becomes “this is not a pipe” since saying is not seeing. Consequently, if what I see is a pipe, what I say, necessarily, is not a pipe. Very well. We’ll see what this means. For now, I would just like for you to be convinced of this: that before understanding anything at all, things have to churn about in your head. And if you don’t let them churn about in your head, then at that point you’ll have completely sensible objections, but at the same time, that you have completely sensible objections, you’ll understand nothing at all. So, refrain from any objections.
So. I’ve come to my second section. From things and words, we have slid over to: lesson of things, lesson of grammar. Or, if you prefer: the drawing, the text. As he will tell us in This Is Not a Pipe. Which gives me my third pair: drawing-text. [Deleuze writes on the blackboard] For now, then, I’ve mentioned two books by Foucault: The Order of Things and This Is Not A Pipe.
The third theme, which recurs really constantly in one particular book by Foucault: The Birth of the Clinic. The visible and the statable [l’énonçable]. The visible and the statable, this is a pair of notions that The Birth of the Clinic refers to constantly. In what form? How is a disease visible in a given period? What makes it be seen? [Qu’est-ce qui la fait voir?] The symptom, that’s what makes a disease be seen. How is the clinic defined when it is formed in the 18th century? Above all, the clinic is a new way of making disease be seen. But at the same time, disease is not simply a set of symptoms, that is, visible; it is also statable. It is a combination of signs. And the sign is readable just as much as the symptom is visible. The visible and the readable are not the same thing. The subject of The Birth of the Clinic is the visible and the statable in relation to disease, and with regard equally to the development of the clinic and to the development of pathological anatomy. One more step, then. Couldn’t we say: Here it is, what fundamentally interests Foucault, following this first axis that will allow us to define archaeology, what fundamentally interests Foucault is the pair—that gives me one more pair . . . [He writes on the board] the third one: visible and statable; so, the fourth one: visibilities and statements.
Oh, indeed? What does this imply? This advance, this move to a new pair: visibilities and statements? It implies that what is statable is the statement, what is visible is visibility. You’ll tell me: that’s pitiful! Not at all! Not at all. It complexifies. If this formulation has a meaning—‘that which is visible is visibility, that which is statable is the statement’—it simply means that statements will not be given ready-made… [Interruption of the recording] [46:40]
Part 2
. . . to speak. Visibilities are not to be confused with seen objects or qualities. So, this will become more complex. Okay, but that’s exactly what ‘statements are statable and visibilities are visible’ means.
So, I’ve made a bit of progress by moving to this—one, two, three, four—in moving to this fifth pair. But now I’m getting carried away. Ah, can’t do that! Here, again, I have to hold back. And why? Well, yes, visibility and statement: in what way is that constant throughout this whole first period of Foucault’s? Visibility, statement. Well, let’s look. History of Madness. I’m taking the two books that have already seemed to me to have a sort of parallelism. History of Madness, about the asylum, and Discipline and Punish, about the prison. What’s going on?
History of Madness tells us, well, in the 17th century the general hospital, the house of correction, the asylum, appears. But what is this? It is an architecture. An architecture. The prison, likewise, is an architecture. What is an architecture? It is an aggregate of stones, let’s say: it’s an aggregate of things, it’s a material aggregate. Okay. It that what it is? Yes, yes, of course it’s that. But if I define it, if I define the general hospital or the prison in this way, does that truly tell me something? Not much. Not much. I could always talk about a “prison” style. When I say, oh, this apartment building, it looks like a prison, I certainly do mean something. There is a prison style, there is a hospital style . . . What else could be it be, besides this? The hospital is a place where we see. Or if you prefer, it is a place that makes us see [un lieu qui fait voir]. The prison is a place where we see, it’s a place that makes us see.
What does that mean? It means a number of things. It means a number of things. It means that the asylum implies, or the general hospital implies, a certain way of seeing madness. Does it presuppose this way of seeing? No, the inverse is also true. It itself presupposes this way of seeing. In the asylum, the mad are seen. They are seen in the strict sense. They are seen in the strict sense because in the 17th century there are tours, the mad are displayed.
Note that still today the mad are displayed. Of course, it’s not done in the same way. There are tours where people, in the 17th century, go to see the mad behind bars. The general hospital is a place of visibility. Not simply because there are . . . tours. Obviously, for a deeper reason. The general hospital implies a new way of seeing madness. What does that mean? It means that the way is not the same as in the Renaissance, that it is not the same as in the Middle Ages. Did one see madness in the Renaissance and in the Middle Ages? Obviously, one saw it, but not in the same way. There wasn’t the same visibility of madness. Okay. Certainly, this visibility covered, included, the way in which the mad person saw himself. Okay. A certain way of seeing.
In fact, here is one of the essential themes of History of Madness: the general hospital united and assembled into a group, the mad, but also vagabonds, beggars, jobless people, and also libertines, reprobates . . . all in the same grouping, the grouping of the general hospital. Okay, understand. We just move along bit by bit. Fairly quickly, since the 18th century and the end of the 17th century, voices spoke out. Voices spoke out to denounce this mixture. Sometimes in order to say, which does not amount to the same thing, one must separate the vagabonds from the mad; the vagabonds, the jobless, did not deserve to be likened to the mad. At other times, to say: the mad deserve special care, they should not be mixed in with the vagabonds. So, given that these voices spoke out to denounce the mixing of vagabonds and mad people, the mixing of the jobless and the mad in the general hospital, given that these voices spoke out this way, it thus looks like the 17th century did not make the distinction. Through a sort of shortcoming or through a sort of incapacity. Through a sort of blindness.
What Foucault shows is that it’s not that at all. It’s that if the 17th century mixed the mad with the vagabonds and jobless, this was done in the name of its own perception of madness; it was not in the name of a certain incapacity, it was in the name of a perfectly articulated perception, which, simply, would be articulated entirely otherwise in another historical formation. In History of Madness, he tells us that it is in the name of a collective sensibility, here’s the idea of collective perception . . . There is a collective perception of madness in the 17th century that grounds its assimilation to “vagabonds, beggars, libertines, jobless.”
Thus, I could try to define a way of seeing madness in the 17th century which would not be the same as the prior definition—Middle Ages, Renaissance—and which would not be the same as the subsequent one. This will be visibility, and indeed, what is an architecture? Well, an architecture is of course an aggregate of stones. But it is more, and it is above all, a place of visibility. Before sculpting stones what one sculpts is light. That’s Foucault’s idea.
Moreover, it becomes, at this juncture, a big idea. I don’t know if it’s true, if it’s not true, but his starting point is that architecture is a place of visibility. Architecture arranges visibilities. Architecture is the establishment of a field of visibilities. You see: always this ascent towards visibility as an ascent towards the condition. What conditions architecture is the visibility that one intends to realize. A way of seeing. Seeing madness in such and such a way. So, it’s as a function of the way in which the 17th century sees madness and not as a function of a blindness, that the mad are mixed with the vagabonds, the . . . etc. You’ll say: why? We’ll see all that. For now, we’re only looking for points of orientation.
And the prison? Shouldn’t we say the same thing about it, even though the prison comes about in the 18th century? It is an architecture, it is stones harder than any stone. Well: no. Prior to being stones harder than any stone, the condition for the prison is a regime of luminosity. It is a sculpture in light and there is no other definition of the prison. Light, seeing—but seeing what? Obviously, not seeing just anything whatsoever. Seeing those who are in prison, that is, seeing crime. Seeing crime. Prison is the place of the visibility of crime, just as the asylum is the place of the visibility of madness. This by itself should make us reflect on one point.
We said that—and we must also come back to this point, as well, but for now we are grouping certain notions together—we said that Foucault was particularly and almost exclusively interested in places of confinement: the asylum, the prison. Some solid objections were even advanced to him. For example, there is a very interesting page by Paul Virilio where he says: but Foucault, it’s . . . uh, it’s irksome because confinement is very outdated, and today the forms under which we are living are no longer forms of confinement, but something even worse. And so much so that Virilio thought that Foucault left out something fundamental about our modern societies, which no longer operate on confinement. Ahh.
Virilio’s page is interesting, it is very interesting. But it is—obviously, like every objection, this doesn’t detract at all from his page—it absolutely does not meet its mark. Why not? Because if there is anyone, who, before Virilio, said that the problem was not that of confinement, it is Foucault. He already said this with respect to places of confinement. In what sense? In the sense that the general hospital and the prison are only secondarily places of confinement. They are primarily spaces of visibility. That is, spaces of visual gridding.
And, in Discipline and Punish, Foucault works this out very much for and about the prison. He will say: What is a prison? What is the function of the prison? The function of the prison. And he’ll select a really fascinating text by an author from the end of the 18th century, part of a reformist movement at the end of the 18th century, namely, [Jeremy] Bentham, B E N T H A M, who had written a book called The Panopticon. The Panopticon. The panoptic. And the panoptic, which was the model prison, what was it? The panoptic was a place where the inhabitants, called “the prisoners,” had to be seen at every moment without themselves seeing. And the prisoners would be seen at each moment by people called “overseers” [surveillants] over-seers [sur-veillants], who, themselves, would see them without themselves being seen. So, here we have a distribution of light and shadow. How was this distribution to be made? It’s not hard, the panoptic, was, roughly, [He draws on the blackboard] an encirclement. A thick encirclement with window slots along the outside periphery and along the inside periphery so that light penetrates. At the center, there is a tower, a tower with shutters. Here, on the periphery, with an outside window, and an inside window, are the cells. Light permeates the cells. The central tower, the control tower, has shutters so that the prisoner can see nothing of what happens in the tower. However, from the tower everything that goes on in the cells can be seen. But the prisoners cannot see the cell next to theirs. On the one hand, you have: being seen without seeing. On the other hand, you have: seeing without being seen. This is the panoptic.
In other words, the prison is a form of light, is a distribution of lights and shadows prior to being a pile of stones. Does this implicate a conception of painting, in Foucault? Perhaps . . . according to which light itself would also be a condition of painting, a condition of the act of painting. We have reason to wonder about this. A painting is a visibility. I can say of painting that it is the art of visibilities. Perhaps, in this we might have one reason for which painting and architecture have an essential relation, a close relation.
So, I’ve substantiated the idea of visibility as much with respect to History of Madness as with respect to Discipline and Punish. And I repeat: prison is the visibility of crime, crime brought to light, just as the general hospital is the visibility of madness in the 17th century. The way in which the 17th century sees madness, brings it to light. However, on the other hand, there is the statement. Why “on the other hand?” Why? The fact is . . . the fact is . . What does that mean? At the same time as the hist . . . At the same time as the asylum in the 17th century, there is a certain state of medicine that includes a category of illnesses—once again: illnesses of the humors, illnesses of the head, illnesses of the nerves. There is no question of mental illnesses. Foucault’s analyses are definitive ones: the 17th century knew nothing of the category of mental illness, for simple reasons we will see, which are that with respect to medicine, it never distinguished soul and body and there is no psychology; therefore, there is no mental illness. But there are illnesses of the head, there are illnesses of the humors, there are illnesses of the nerves, that is: neuroses — the word appears starting in the 18th century— neuroses, which are nervous illnesses. Very well.
So, there are a certain number, a certain body of statements, a certain set of statements about a group of illnesses. These illnesses are statables [des énonçables]. But here’s the raw fact, as an historical fact: medicine does not enter the general hospital. The general hospital does not have medicine as its origin. Even more: The general hospital, the asylum, the house of correction have nothing to do with medicine. One does not treat in the general hospital. Ah, one does not treat in the general hospital . . . So, then, where does the general hospital come from? Foucault’s answer—I’m really condensing this analysis and very simply—it comes from the police [la police], absolutely not from medicine. And the medicine that treats, it treats, but it treats outside of the general hospital. So. It’s as if there were a heterogeneity between the general hospital, the place of madness’s visibility, and medicine, the place of the statability of illnesses of the head. Will they connect at all? Yes, once the two of them are constituted, there will be contact. But the genealogy is independent. There will be contact, but it is not the same formation. And when I spoke about a parallelism with Discipline and Punish, you will find a theme, you will find the same theme treated more deeply in Discipline and Punish.
Let me return to History of Madness for a brief moment. So, statements, medical statements, what are they about, since they do not enter, they do not concern, what one sees at the general hospital? What do they concern? What is the object of these statements? It is—and this is undoubtedly what is of powerful interest in the 17th century or the Classical Age, to have formed this notion—medical statements concern a notion that is specific to the 17th century: unreason. Unreason. The general hospital contains mad people and makes them be seen, but medicine states unreason. Once again, the question is not to know whether there is a connection between the two, the question is first to say and to show that the formation of the two is altogether different. Good.
I return to Discipline and Punish. [Pause] Prison is a way to see crime, it is a place of the visibility of crime, of crime punished, of crime as punished crime. To make punished crime be seen. And, actually, it is a place of light, that distributes light and shadow. At the same time, in the same period, what is the regime of the statement? Well, in the same period, there is a penal law. Even more than that: there is a whole movement to reform the penal law, in the 18th century, a very interesting movement. And it’s all the more interesting for what reason? It’s all the more interesting because as soon as one studies this great movement of penal law and its reform, one notices that this movement is not about the prison—and that the prison is foreign to penal law. So, needless to say there will be a connection . . . That’s not the question. The point is that, from the point of view of penal law, the prison is but one particular sanction in certain very precise cases of crime, but that all kinds of other sanctions are planned and that, furthermore, penal law considers the prison only with a kind of fundamental unease. As if, when faced with the prison, penal law constantly formulated the statement: this is not a prison. This is not a pipe, this is not a prison. Penal law conceives of a whole regime of sanctions without referring to the prison or with minimal reference to prison. Prison is, in penal law, a genuine foreign body. So, then where does the prison come from if it does not come from the law? It comes from something entirely different, it comes from what Foucault will call disciplinary techniques. Disciplinary techniques of work, of the army, of school. And the prison will be born, not from a juridical grouping [ensemble], but from an extra-juridical disciplinary grouping.
And, in fact, what are the statements of the law about? Just as the 17th century had medical statements that concerned unreason, and invented this notion, so the statements of the law, in the 18th century, concerned delinquency and invented that notion. Between the statements of delinquency and the prison as a place of visibility, there is a heterogeneity. There will be contact, there will be all sorts of relations, but there is a heterogeneity; it’s not the same archaeological formation. Just as there is between the general hospital and the statements of medicine. Thus, here is a new pair: visibilities, statements. What is needed is a comparative analysis of: the prison as a place of crime’s visibility; and penal law as the statement of delinquency; likewise, a comparative analysis of the general hospital as the visibility of madness; and medical statements as statements of unreason. Each one has its line, its own line, independent of the other.
Let’s continue our research into pairs. Now, we’ll try to think out how the pair visibilities-statements is grounded in Foucault. You see that the visible and the statable in Birth of the Clinic is as if transformed into visibility-statement. Once again, architecture must be taken in this sense, architecture as a regime of light. A new pair. I can just as well say that in the prison . . . and on a number of occasions in Discipline and Punish Foucault says this . . . expresses himself this way: the prison is an evidence [une évidence]. As the general hospital is an evidence. This use of the word ‘evidence’ is interesting since evidence is a visibility. Foucault thus makes evidence an historical conception. Each historical formation has evidences. Its evidences. And, in the subsequent period, what was an evidence ceases to be one. That the mad could and even must be grouped together with the vagabonds, etcetera, etcetera, is an evidence for the 17th century—it is not a blindness. In other words, it seems to me that here Foucault’s great historical principle comes into view: every historical formation sees all that it is capable of seeing; every historical formation sees all it can see. And its correlate: every historical formation says all it can say.
An historical formation will be defined by its evidences, that is, by its regime of light. And what?—its discursivities, a statement regime will be called a discursivity by Foucault. Evidence and discursivity. In Archaeology of Knowledge, we will once again be perturbed because the pair changes, but it changes in such a way that we run a great risk. What is this great risk? It’s that we risk seeing only one term of the pair. And yet the other one is there, but it is no longer designated any way other than negatively. And we’ll find the expressions: non-discursive formation-discursive formation. So, here, seeing, visibilities, are now designated only negatively under the name of non-discursive formations. Why is this? What has happened? The first answer, but an insufficient one, is simple and consists of saying that Archaeology of Knowledge is a book entirely dedicated to the question: “what does ‘statement’ [‘énoncé’] mean?” Thus, since he does not consider the other pole, the other head, visibilities, insofar as he alludes to them there, he alludes to them there solely in a negative way. But the question would just pop up again, it pops up again—I leave it as it is, namely: why does Foucault feel the need to devote a book to statements, separated from visibilities?
I continue my research, restricted to terminology, and I see that, again, the pair will shift, but always under the classifications of seeing/speaking. In the book on Raymond Roussel, Foucault tells us that, according to him, there are two kinds of works by this poet, two kinds of works. . . In fact, the book will even spot—but I can’t yet take this into account—three kinds of works by Roussel. But we will stick with the two that are easiest. He tells us: on the one hand, there are books that describe machines, extraordinary machines, and these machines do what? They allow us to see a strange spectacle. [Pause] Machines that allow you to see a strange spectacle. Okay. So, what does this mean, machines . . . ? And, indeed, we will see . . . later, we will talk about Raymond Roussel as Foucault sees him and Raymond Roussel in his own terms. In fact, there are extraordinary machines that make one see spectacles, crazy spectacles. And the whole of Roussel’s book describes these spectacles produced from mysterious machines, where, for example, the dead recount the final moment of their life, try to recount it in glass cages. That’s interesting, that idea of the machine. You can sense that we’re close to a confirmation. What we are looking for are the confirmations of what we said earlier about architecture. Plainly, all machines are not optical. It would be stupid to say: every machine is optical. On the other hand, what is less stupid is to say: every machine whatsoever allows something to be seen.
Every machine makes something be seen, in addition to what it does. The steam engine makes something be seen. So . . . optical machines, all the more so, make something be seen. But to say: the machine necessarily makes something be seen is the same as saying: architecture is a sculpture in light. Thus, machines that make something be seen that could not be seen apart from the machine. In other words, visibility is inseparable from a kind of process [processus] that should in fact be called—well, it’s not . . . it’s not important—a kind of process that should be called a machinic process. And descriptions of machines or of machinic processes define a first kind of work by Raymond Roussel. What is the second kind of work? It is books, in this instance, that do not rely on descriptions of machines insofar as they make a spectacle be seen, but that rely on a linguistic operation [procédé]. No longer on a process [processus], but on an operation [procédé]. And what is a linguistic operation? It is an enunciative regime [un régime énonciatif]. The process is machinic, but the operation is enunciative. The second kind of Raymond Roussel’s works concerns enunciative operations [procédés]. Very good. Here we have a new pair: machinic process/enunciative operation.
If I were to try to summarize all this, I will thus appeal to terms that are not Foucault’s, to try to cover all of this . . . You can see that we followed, what, a kind of dualism, a very curious dualism that shifted across the range of Foucault’s books, that alters, that takes on varying aspects. So, I’m thinking of a terminology that a linguist uses, namely, the linguist [Louis] Hjelmslev. Hjelmslev speaks of . . . I’m just restricting this to the words alone, because they seem able to perhaps help us here. He talks about the form of content and the form of expression. He says: there are forms of content and forms of expression. And this is where my invocation of Hjelmslev stops. Why? Because in his work, he’s a linguist, a pure linguist, in his work, the form of expression is a way to re-baptize what linguists called the signifier—there are reasons to want this new baptism—and the form of content is a way to re-baptize what linguists called the signified. We know that this is the case in Hjelmslev’s work.
I’m borrowing Hjelmslev’s words, but I’m saying right away: so, let’s assume that form of content has nothing to do with the signified, let’s assume that form of expression has nothing to do with the signifier. And, in a certain way, this is just what Foucault will show. What is the form of expression in a given historical formation? It is the regime of statements. What is the form of content? I would say, for example, that the prison is a form of content. The general hospital is a form of content.
And, for me in fact, we find the expression, ‘the prison-form,’ with a hyphen between the two words, numerous times in Discipline and Punish. The prison-form is a form of content, since it is not a form of expression; the form of expression is penal law. Penal law is a form of expression, and the prison is a form of content. Medicine is a form of expression, with its statements. The general hospital or the asylum is a form of content. Only, in Foucault’s work, and we will see why this is, the form of expression no longer has nothing to do with a signifier, the form of content no longer has anything to do with a signified. Why? Because visibilities are irreducible to a signified, just as statements are irreducible to a signifier.
Foucault says constantly: discursivities nullify themselves by entering into the order of the signifier. Moreover, visibilities are not of the nature of a signified, either. Why? That’s something we’ll have to search for. But we’ve really made some progress here. You’ll have to grant me, you’ll have to grant me, at this point, this whole list, concerning seeing and speaking, but we’ve at least reached an expression, a formulation of a much more rigorous foundational pair: field of visibilities/regime of statements, or, if you prefer, visibilities and statements.
Foucault’s very curious dualism is founded on this: the irreducibility of one form to another. Seeing and speaking. Once we specify that to see is not simply the empirical exercise of the eye, it is to constitute visibilities. To see or to make see. To state is not the empirical exercise of language, it is to constitute statements. However, it is not easy, to constitute visibilities, to constitute statements, it is not at all easy. It does not exist ready-made since it varies across time periods. And how is such a regime of statements constituted? How is such a place of visibility formed?
Well, we just may have made a tiny bit of progress, since . . . First question: hasn’t the way that Foucault claims to go beyond a history of behaviors and mindsets become clearer? Aren’t –we’re not trying to justify everything, but aren’t the following formulations becoming a little bit clearer? — visibilities are not things among other things, and visions, evidences, are not actions among others actions. Seeing. But it’s a condition in virtue of which any action, passion, etc. arises. All that is done in a period can only be done if it arises in the light. What is done and undergone in a period presupposes its regime of light. And likewise, all that is thought in a period, all the ideas of a period, presuppose its statement regime. Statements are not ideas among other ideas, nor are they simple communications between ideas. They are conditions for the deployment [le déploiement] of all the networks of ideas that operate in a period. Visibilities are not only givens like other givens; they are conditions of light that make possible bringing to light, raising into light, what is done and what is undergone in a period. Foucault does not do a history of mindsets and a history of behaviors, he ascends to the conditions proper to each period, which make both behaviors and mindsets possible. In other words, he works as a philosopher and not as an historian. Seeing and speaking determine the conditions because seeing is superseded toward fields of visibilities and speaking is superseded toward regimes of statements.
And, so, this is what . . . [Interruption of the recording] [1:33:26]
Part 3
… So, I . . . Ahhh. I’ll repeat what I said and what those who were not here missed. Uh . . . I’m thinking of our sessions in the following way. This is so that you can manage your arrival times well. I’ll be here at 9am every time, whence my state already . . . Around 9am, I’ll be here and I’ll review with those who will be here the points deemed necessary from the preceding session, or else we’ll go a bit further in a one direction or another. That will put us at 9:45 or 10am, at which point I’ll start the new session. Because there are lots of you who are arriving between 9:30 and 10, which bothers me. You see, it’s in this way that I shift… I do this. And, so, each time, there will be a beginning that will consist re-examining points raised in the preceding session in greater depth. Do you follow me? Bad faith . . . Well, in any case, it doesn’t apply to you, you arrive at 10 o’clock.
Claire Parnet: But your course is at 10 o’clock, that’s . . .
Deleuze: No. This year it’s at 9 o’clock.
Claire Parnet: Oh, really!
Deleuze: Everything changed due to the reforms.
Claire Parnet: Six hours of class, if you review each time, you’ll always end at 1:30 [several inaudible words]!
Deleuze [laughing]: Ill will, bad faith. Well. Anyway, you’ll see, surely.
Okay, see, it amounts to saying what? It amounts to speaking in a way, uh . . . a philosophical way. Namely: seeing and speaking are pure elements. Seeing and speaking are part of an analysis into elements. Seeing is thus not a behavior, speaking is not an idea. It’s not a history of ideas, a history of behaviors. It is ascending truly to conditions, that is, to visibilities and statements. Only we’re not finished here. Because right away, at least what came before has strengthened us for something. For what has it made us strong? For avoiding, not a misinterpretation of Foucault, but a kind of mutilation. The mutilation would be this: in virtue of the force and originality of Foucault’s theory of statements, of the thoroughly novel notion of the statement he contrives, to make of Foucault someone for whom there are only statements, to the point that the visible would merely be the residue of statements, or illusions projected by the statement, and in which Foucault would be reduced to what? Let’s put it entirely simply: to a philosopher of language, or, worse, to an analytic philosopher. I mean: Anglo-American analytic philosophy. There is nonetheless good reason to compare Foucault to these authors, some linguists, some analytic philosophers, assuredly. But what I insist upon, is, once again, this: “so bizarre a dualism.” The visibility/statement dualism.
And no doubt, a certain number of you have seen Foucault during his lifetime. Those of you who saw him, all you had to do was see him, I believe, to know two things. To know that, of course, he was less passionate about what he himself said than about what he heard said. Uh . . . But he also had a passion for seeing. He had a sort of visual genius, no less than a genius for the statement. And why am I saying this, and why do I introduce a note of a personal nature, here? Because I recall a text in which a commentator on El Greco, the painter, spoke about the elongated forms of Greco’s, elongated bodies, and suggests that it is not without relation to El Greco’s astigmatism. He was astigmatic, but in any case, okay . . . that sets out a good problem. Is the elongation of the body as a pictorial technique an effect of the astigmatism? No, certainly not. But why should there not be a complex relation between the two? Indeed, I want to emphasize Foucault’s myopia. I mean: his passion for seeing, in a certain way, must we, it is legitimate to, relate this sort of passion for seeing with what he meant by ‘seeing?’ For, when he spoke of visibilities . . . I was telling you, visibilities, are not things and objects, and one can justify this philosophically—this cannot be things and objects, it’s something else.
But when Foucault speaks about it, what is it? The expressions that recur constantly are: glimmering, glittering, radiance. Radiance, glimmering, glittering . . . his very style, his style is a style of light. The materiality of his style is extraordinarily luminous, that is, it’s shimmerings, glitterings, radiance. The visible is not the thing or the object, nor even the quality. The visible is the glimmering, the glittering, the radiance.
That’s what a visibility is. We must investigate why that is so. Just as I said: statements are something altogether different than words and sentences. Okay. So, I’m saying: he has a passion for seeing no less than a passion for stating. And for this reason, he is a truly great describer of paintings. If you will, I could also have put, among my pairs, and I’ll add it as the last one: description/stating [énonciation]. Description . . . description is always description of a painting. A description is an entirely different thing than a statement. Moreover, on that point, he crafts descriptions throughout his corpus that are either descriptions of paintings, of the painting by Velasquez, description of the painting by Magritte . . . but also descriptions of the prison, descriptions of the asylum . . . This is how I answer the question ‘How can we speak of a visibility? It’s that speaking of a visibility is describing, and describing is not the same thing as stating. And Foucault advances on two tracks, along two non-symmetrical paths: descriptions and statements.
As a result, we must consider: what is a painting? What is Foucault’s theory of descriptions and not merely: what is the theory of statements? That said . . . That said, I seem to be saying the contrary, but it must be said. It does not seem contestable to me that there is a primacy of statements. Statements have primacy over visibilities. Only, what does “to have primacy” mean? What we have here are problems for the future. What does “to have primacy” mean? The primacy of statements over visibilities culminates in The Archaeology of Knowledge, but Foucault will never renounce this point. In fact, you find in The Archaeology of Knowledge a very curious expression, as Foucault tells us: “the discursive has discursive relations with the non-discursive.” There is no better way to express the primacy of the discursive. The discursive has discursive relations with the non-discursive, that is, the statement has enuciative relations [des relations énonciatives] with the non-statement [le non-énoncé], with the visible.
Primacy of statements? We must say: yes. There is a primacy of statements. Except that what I would like to point out here is this, and I ask you already to reflect here, this point will occupy us greatly later, it’s a very exacting point, very precise. It’s that primacy has never meant “reducibility,” and this appears evident to me. Any understanding of Foucault that would translate the primacy of the statement over the visible in terms of the reducibility of the visible to the statement deeply mutilates Foucault’s thought. Further, something only has primacy over something else to the extent that this other thing is of another nature. You cannot draw the conclusion of reduction from primacy, on the contrary. The statement has primacy over the visible—what does “primacy” mean? In any case, “primacy” does not mean that the visible reduces to the statement.
What is more, at the point where we now find ourselves, we should posit all sorts of theses at once. Four theses. If you’ll grant me these seeing/speaking pairs. First thesis . . . And I believe that they are all to be found in Foucault’s work. First thesis: there is a difference in nature between seeing and speaking, between the visible and the statable. In more scholarly terms: there is no isomorphism. What does “isomorphism” mean? There is no isomorphism, Foucault says this constantly, and he said it explicitly in This Is Not a Pipe and in Birth of the Clinic, really constantly, in all his books, isomorphism is rejected. There is no isomorphism between seeing and speaking, that is, between the visible and the statable. This means: there is no conformity. “What does “conformity” mean? It means: either a common form, or a correspondence between the two forms. A bijective function between the two forms. Foucault denies the seeing-speaking conformity, correspondence, isomorphism. In other words, we never see what we speak about, and we never speak about what we see.
Does he say this? Yes, he says this, he says it constantly. So, this is why I say: this is the first great thesis. Notably, at the opening of The Order of Things. We’ll look at these texts later, in detail, but I summarize one of Foucault’s great texts: what we see never resides in what we say. What we see . . uh, on page 21, as I recall it. “What we see never resides in what we say” and he adds, “and even if what we say proceeds by metaphor, etc.”—there’s a pretty long sentence here—“even if what we say proceeds by metaphor, displacement, etc., it never makes us see what we see.”[6] Which means what? To speak is not to see, to see is not to speak.
There is a disjunction between seeing and speaking. There is a disjunction, there is no conjunction, there is a disjunction between seeing and speaking. You’ll say to me: but that’s not true; I can speak about what I see and I can see what I speak about. If you said that to me, you wouldn’t have followed so far. Of course, you always can, but the question is: is this of any interest? When you speak about what you see, isn’t it . . . it’s of interest only if you communicate it to someone who himself does not see it. Because if it’s someone who sees the same thing you do, I don’t quite see, uh . . . the point of saying what you see. Look, a boat! Okay, that’s about it. Look, someone who . . . Uh. Okay. So. Do you understand? So: the seeing-speaking disjunction. What does that mean, again? And what does it stimulate? In this regard, Foucault, plainly, is a part, I don’t know, of a set of thoughts, since he is in a certain relation to other thinkers. We will have to scrutinize this theme of a fundamental disjunction between seeing and speaking rather carefully. Once again, this seems like a platitude or even something contestable, but do realize that it must be understood in a particular way, at the point where we are now.
What is interesting is not that speaking and seeing are not the same thing. What is important, once again, is not speaking and seeing. It is, rather, what speaking refers to and what seeing refers to. Perhaps seeing and speaking don’t go together, but what doesn’t go together is what seeing refers to and what speaking refers to—once we specify that, according to Foucault, what seeing refers to is visibilities, and what speaking refers to is statements. Thus, this amounts to saying: between statements and visibilities there is a difference in nature, there is no isomorphism. And once again what did he show in History of Madness? That there was no isomorphism between the hospital and medicine, each had its own formation. The hospital comes from the police, unreason comes from medicine. Likewise, the prison came from the disciplinary, delinquency came from the judicial. There is no isomorphism, there is no common form.
Moreover, I’m saying: this should make us think of other authors . . . Very well. Yes, above all of Blanchot, and Foucault frequently indicated his gratit . . . his debt to Blanchot. There is no question of considering him to be a disciple of Blanchot, I think that there is more of an encounter between the two, especially since, starting from the same thesis—seeing is not speaking—Foucault and Blanchot will work out the theme in two fairly different ways.
And, indeed, “seeing is not speaking” is a great text of Blanchot’s, it runs throughout his corpus, and it’s a particular chapter, one of the most beautiful, among the most beautiful pages from Maurice Blanchot, in The Infinite Conversation.[7] To speak is not to see. To speak is not to see. You will notice that Blanchot does not add: to see is not to speak. I believe that Foucault is the one who makes it reciprocal: to see is not to speak. On what condition, in what . . . and what is the meaning, what difference from Blanchot? Well, no matter, that’s what we’ll see later on. This will be a problem for us: the relations with Blanchot. But the very regime of a disjunction between seeing and speaking, those of you who were here last year, you will perhaps recall that we ran into it in another context, which was the context of cinema. And it had seemed to us, when we worked on speech in the cinema, that certain directors [auteurs] in modern cinema made very special use of speech and of speaking, in the sense that it was a use that was disjunctive in relation to the visual image—between speaking and the visual image, there were relations of disjunction. It has seemed to us that this was above all confirmed or represented by three great current directors: the Straubs, Marguerite Duras, and Syberberg. What is this seeing-speaking disjunction? As Marguerite Duras says, it’s as if there were two films, two films, that is, without isomorphism, the film of voices and the visual film. And the voices evoke an event that we will not see, whereas the visual image presents places without events—empty places or mute places. Think, for example, those of you who have seen this film, of India Song, where the visual image is shunted off in one direction and the film of voices in another. Okay.
What we see does not reside in what we say. What we say does not make us see. There is seeing and there is saying, but they are in a disjunctive relation, which is to say, a non-relation. “Non-relation,” this odd expression, is from Blanchot. Blanchot says: ‘between seeing and saying, between seeing and speaking, between speaking and seeing, there is a non-relation.’ And he adds: ‘this non-relation that is perhaps more absolute than every relation.’ That’s curious. This non-relation that is perhaps more absolute . . . You’ll see, if you happen to read This Is Not a Pipe, that Foucault takes up the expression “non-relation,” in This Is Not a Pipe, saying: between the drawing and its title, that is, between the visible and the statement, there is a non-relation. Which amounts to saying: there is a disjunction. Ah, is there a disjunction? Okay, but at the same time, we return to the following point: this non-relation must certainly, in a certain way, be a relation and even be more profound than any relation. The two forms are irreducible. The form of the visible and the form of the statable are irreducible. There is neither conformity, nor correspondence, nor isomorphism. There is non-relation, there is disjunction.
And this non-relation must even be, in a certain paradoxical original way, a relation and this will not be a relation between two forms, this will not be a conformity. Do you notice what we are circling around? We are really at the heart of the problem of truth as Foucault will set it out. Because the truth has always been defined by conformity, conformity between the thing and the representation, conformity of saying and seeing. Very well.
So, Foucault’s first thesis is, and he will never recant this, is the heterogeneity of the two forms, difference in nature: what we see is not contained in what we say and vice versa. Second thesis: Which one comes first? The question cannot be asked. There is a reciprocal presupposition here. Each one presupposes the other. There is a reciprocal presupposition here. Third thesis: The statement nonetheless has primacy over the visible. We come back to our problem. Where does this primacy come from? You see that the primacy does not imply any reduction. Granted, the two are irreducible. Granted, one has primacy over the other. But far from entailing a reduction, primacy presupposes irreducibility. I can exert a primacy or a power only over what resists me, only over that which has another form.
And, finally, the fourth thesis: there are, in one direction as in the other, mutual captures, namely—this is surely what will be the most complicated in Foucault—statements constantly grasp, capture, the visible. Visibilities constantly grasp, capture, statements. But how is this possible, since these are two heterogeneous, irreducible forms. Now this, this is going to be the most beautiful part. In my view, this is what is most beautiful in Foucault, this system of double capture. Visibilities take hold of statements, statements take hold of visibilities.
And it is in this little book, This Is Not a Pipe, that Foucault works it out most thoroughly. He speaks of the incision of statements into the visible, of the incursion of the visible into the statement. Incursion, incision, he says: each—here, I’m quoting from memory, but by heart, almost by heart—“each one sends its arrow into the target of the other.” Each one sends its arrow into the target of the other, and he adds—here I can guarantee the word—it’s a “battle.” It’s a battle. You see why he adds “it’s a battle” and really experiences the relation between visibilities and statements as a battle. He necessarily experiences it as a battle, and it must be a battle because they are not of the same form. There can be no agreement between statements and visibilities. There can be no agreement. Why? Because agreement is a common form or a correspondence of form to form. But there is neither a common form, nor a formal correspondence between the visible and the statable. If there are phenomena of capture, this will thus not be in the form of an agreement, of a conformity, it will be in the form of a capture, of a violent capture, a battle. So, you’ll say to me: those are just words. Yes, but these words allow us to circumscribe the problem.
As a result, I find myself with these four fundamental theses: difference of nature or heterogeneity of the two forms, the form of the visible, the form of the statable. Hence, non-relation. Reciprocal presupposition of the two, each one presupposing the other. Third: primacy of one over the other, primacy of the statement over visibility. Fourth: mutual capture, wrestlers’ holds between visibilities and statements, as in a battle.
What does this amount to saying? This already gives us an entire agenda. We’ll have to make do with these four theses. All that I can conclude for the moment is what? What is an archive? And, listen, I’ll answer, I’ll give a final answer which will be, in a sense, the most disappointing, and will stop being disappointing if you take into account everything it took for us to get to it. I’ll say: well, here it is: the archive is fundamentally audiovisual. And so: see you later, my work here is done! The archive is audiovisual—only this commonplace has been transformed by Foucault. Because, let’s go on: right away, what is archaeology? Archaeology is the study of historical formations. In what way is it something different than history? Because it concerns ascending right to the conditions, the visible and the statable, and because history will never be able to identify pure statements or visibilities. These are pure elements: they require a philosophical analysis.
And so: What is an historical formation? I now can say: an historical formation is an arrangement [agencement] of the visible and the statable, it’s a combination, it’s a way of combining visibilities and statements, once we specify that, as we’ve seen, any statement . . . The two are irreducible, but it is not the case that any statement can combine with any visibility. There are combinations or captures that rule out anything going together with anything else. The coherence of a period stems from the fact that its visibilities, in virtue of their own form, are combinable with its statements in virtue of their own form. That’s what will define this interlacing, this intertwining, of visible [things] and statable [things] that varies across historical formations. No historical formation sees from another one, that is, has the visibilities and the statabilities of another one. And we can even say that this is what defines . . . as soon as there is a variation in the statement regime and the fields of visibility, you can say: we are entering another historical formation. Hence, archaeology, this arrangement [agencement] of the visible and the statable as constitutive of the historical formation, is what Foucault will call, in his own terminology, a deployment [dispositif]. A deployment.
And, lastly—you’ve had enough, huh. . . And, lastly, I’ll close on the following point: the two, the two—seeing and speaking, that is, visibles and statables—make up what he will call a savoir, a form of knowledge. To know [savoir] is always to produce the non-relation between the visible and the statable, it is to combine the visible and the statable, to produce mutual captures of the visible and the statable. At stake is the problem of truth. You will perhaps notice that I have defined in the same way the archive, audiovisual, the historical formation, a combination of the visible and the statable, and knowledge [savoir]. Well, yes. The reason is that for Foucault there is nothing, there is nothing beneath knowledge. Everything is a knowledge. Everything is knowledge. There is no experience prior to knowledge. This is his break with phenomenology. There is not, as Merleau-Ponty said, a ‘wild experience,’ there is not lived experience or, rather, lived experience is already a knowledge.
Except that not all knowledge is a science. But there is nothing beneath knowledge. What is an historical formation? They are thresholds of knowledge, varying greatly amongst themselves, that knowledge has constituted in a time period, a stacking of thresholds, a stacking of thresholds set out in various directions. Whence ‘archaeology of knowledge.’ It is knowledge that is the object of an archaeology, because knowing is precisely to combine the visible and the statable. In what sense? In the sense, I’ll just start into, now, what I’d like to do next time and . . . in the sense that the visible refers to a process [processus], as we saw, the statable refers to an operation [procédé]. To put knowledge together, that is, to combine the process [processus] of visibility with the operation [procédé] of statability, the operation [procédé] for the creation of statements—what is this? The process [processus] plus the operation [procédé] is a procedure [procédure]. Knowledge is a procedure [procédure]. Truth does not exist independently of the procedure [procédure], and the procedure [procédure] is the combination of the visible process [processus] and the statable operation [procédé]. So, now we have an entire group of notions. We did a whole survey, just about, of what? Of what is, plainly, just the first axis of Foucault’s thought.
So, next time, starting at 9 o’clock, I’ll be here and . . . to examine with you, to examine with those here then, how . . . if there are things to revisit . . . all of that. But I would like for you to think about all this. It’s sort of the plan for our upcoming sessions. So, those of you who can tell that it won’t concern them, you of course don’t come back. Those who think that it does concern them, come back, and at the start, if you have questions, at the start, we’ll take care of those questions and then go on from there. [End of the recording] [2:09:31]
Notes
[1] See Michel Foucault, Death and the Labyrinth: The World of Raymond Roussel, trans. Charles Ruas (New York: Continuum, 2004), 15. Orig: Michel Foucault, Raymond Roussel (Paris: Gallimard, 1963), 21.
[2] Michel Foucault, “Sept propos sur le septième ange?” Dits et écrits, I, 1954-1975 (2001). Paris : Gallimard, Quarto, 13-25. Preface to La Grammaire logique de Jean-Pierre Brisset (Paris: Tchou, 1970).
[3] This is close to a title published in 1971: Les Hommes et la mort en Anjou aux XVIIe et XVIII siècles. Essai de démographie et de psychologie historiques (Paris-La Haye, Mouton, 1971). (English title would be: Men and Death in 17th and 18th Century Anjou. Essay in Historical Demography and Psychology).
[4] The English title would be “Words and Things”.
[5] Or “object lesson,” la leçon de choses.
[6] The Order of Things, 9; orig. 25. Translation adapted.
[7] Maurice Blanchot, The Infinite Conversation, trans. Susan Hanson (Minneapolis, MN, University of Minnesota Press, 1992); L’entretien infini (Paris, Éditions Gallimard, 1969).
For archival purposes, the transcription of this seminar by Annabelle Dufourcq took place starting in 2011 with the support of a Purdue University College of Liberal Arts grant. The translation was completed in 2018, and subsequent revision of the transcript, translation and time stamp were completed in December 2021 and April 2022. Additional revisions were added in February 2024.