December 17, 1985
Visibility is always a picture. Why? Because visibility is a being of light before it is a solid being. Well, light, just like the statement, is an integration of singularities, singular points. And you can only define a light…and the path of a light as going from one singularity to another; that is, there are luminous series just as there are verbal series.
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
Shifting now to the discussion of power, Deleuze examines a particular passage in The Archeology of Knowledge, Foucault’s choice of a statement example, AZERT (QUERTY equivalent on French keyboards). Deleuze emphasizes key points on the as a condition of designating and as neither banal nor original, but regular, a regularity implying “facultative rules” of enunciation and an emission of singularities that Deleuze demonstrates with several drawings in order to introduce the key term, a series of regular points, with all statements being serial. Elaboring the linkages between enunciation and singularities with reference to different mathematical perspectives (cf. Albert Lautman, Henri Poincaré), Deleuze shows that the statement is a function, and that for Foucault, a linkage exists between the emission of singularities and “rapports de force”, i.e., force relations, in short, knowledge as integrating force relations between things, people, letters, shadow and light, between everything. Deleuze then indicates that Foucault’s method of singularities, developed in The Archeology of Knowledge, was applicable to other domains: for the aesthetic domain, Deleuze takes Mallarmé’s poem, “A Throw of the Dice Will Never Abolish Chance”, to apply this method, concerning chance and singular points. For the social domain, the confession and related terms (sacrament, guilt) imply an envelope of singularities relating to the concept of “pastoral power” in contrast to royal power (cf. Confessions of the Flesh). Given Foucault’s shift of emphasis in his final books, Deleuze offers examples from Foucault, the constellation of singularities in the formation of psychiatry and of pathological anatomy. Linking these relations back to the knowledge series, to light as the integration of singular points, the verbal series and the luminous series, Deleuze draws from Velázquez’s Las Meninas and Roussel, arguing that the two series can be interlaced as an instances arising from another dimension, and he also insists that the Foucault-Blanchot intersection emerges within the three notions: the neutral or the “one”; the singular; and the multiple opposing the same, hence, the force relations and singularities constituting the bases for the knowledge power-system.

Download
Gilles Deleuze
Seminar on Foucault, 1985-1986
Part I: Knowledge (Historical Formations)
Lecture 08, 17 December 1985
Transcribed by Annabelle Dufourcq; additional revisions and time stamp, Charles J. Stivale
Translated by Samantha Bankston; additional revisions, Charles J. Stivale
Part 1
So, you see what the problem is. It coincides well with the end of the term. It’s… after painting a kind of picture of knowledge along Foucault’s lines, we were almost pushed… I mean, it wasn’t a matter of will, it wasn’t… we truly were pushed towards a second domain, that of power. And I mean: I have a feeling that this is also what happened to Foucault. Meaning, he actually started with an epistemology, or with an attempt to construct a doctrine of knowledge, and it was this doctrine of knowledge that literally pushed him towards the discovery of a new domain, which would become that of power. So much so that what we were already looking for last time was a kind of transition that caused us to shift from knowledge to power, and we proceeded by making a series of remarks, the most concrete remarks possible. And for today, we suggested closely examining a particular passage, because it is a mysterious passage from The Archaeology of Knowledge. You see, ultimately, it’s always been very embarrassing to answer the question, if this question was ever asked, but come on, come on, come on, come on: give us an example of a statement. At least now we know why we were embarrassed, such that we’re no longer embarrassed. It is very, very difficult to give an example of a statement.
In fact, statements are distinct from words, sentences and propositions, but at the same time, they are completely immanent to them. I cannot give an example of a statement without passing through what the statement is not, namely words, sentences and propositions. So, every time that I am asked for an example of a statement, I will only be able to provide a sentence or a proposition and explain why the statement is different from the sentence itself. But since it doesn’t exist outside of a sentence, it is very difficult for me to give an example. Accordingly, if someone held fast to their demand—give an example, an example of a statement! — Foucault would respond as he does in The Archaeology of Knowledge, and that’s what should be the focus of our session today. He would respond: A Z E R T, azert.[1] So, obviously, memories flood back to us. We say to ourselves, ah yes, the Stoics, for example, had a secret word, this secret word was blituri, it was a wonderfully magical word. [Pause] For the Stoics, “blituri” refers to a word that has no meaning. So, is “azert”, in this sense, the secret statement, A Z E R T? That is why pages 109-114 [82-87] from The Archaeology, which I asked you to read, if possible, should be analyzed very closely.[2]
On page 109, I will quote Foucault’s first remark. The point is to demonstrate that a statement does not necessarily imply a grammar or syntax. And, to demonstrate this, he tells us: an equation is a statement. And he adds: a curve is a statement, “A graph, a growth curve, an age pyramid, a distribution cloud are all statements” (p.109) [82]. An equation and a curve are statements. Can we say the inverse? We want to say the inverse. In any case, I immediately want to say the inverse. But under what conditions is this legitimate? Every statement is a curve. It would be important for us, it would be interesting, very interesting for us, because it would be a way of insisting on the irreducibility of the statement to the sentence. Maybe if the curve were stated it would involve a sentence, but this wouldn’t be the same sentence, it would be the curve of the sentence. But what is that, the curve of a sentence? Alright, let’s move on.
On pages 113-114 [85-86], he tells us a little more. Because he tells us what a statement is and what it isn’t, using even more unusual examples than the previous ones. A curve is a statement. This tells us a little more, because he is saying what a statement is and what it isn’t. What does not constitute a statement? Letters picked at random. Literally, a handful of letters. A handful of letters is not a statement. Letters that are picked at random literally do not constitute a statement. Um, you see this intelligent game: Scrabble, yeah? You take a handful of letters, you have a handful of letters in your hand: that’s not a statement. Fine. [Pause] However, if you copy the letters that you drew at random down on a sheet of paper: that’s a statement. If you copy them down, then it becomes a statement. We have to go slowly because he’s telling us something funny. If I write down this handful of letters, if I reproduce them on a sheet of paper, then it is a statement. A statement of what? A statement of a series of letters with no law other than that of chance. We have to remember this: a statement of a series of letters with no law other than that of chance. But my handful of letters is not a statement.
Or else — and you can see that this example is equivalent — the letters on the keyboard of a typewriter. A Z E R T. These are the first letters on the keyboard of French typewriters. These letters on the keyboard of a typewriter are not a statement. If I copy them down or say them, then it becomes a statement. A statement of what? It is the statement of the order of letters on a French machine. There you go. Suddenly, questions abound, and we think we understand. [Pause] Fine, but before you even begin to think you understand, you have to see how it ends. Look, you recall, right? My handful of letters is not a statement, but if I copy them onto a sheet of paper, then it becomes a statement. If I copy them onto a sheet of paper or say them, it’s a statement. A Z E R T on the keyboard is not a statement. If I say “A Z E R T” or write it down on a sheet of paper, then it is a statement.
On p. 117 [89], he concludes: “A series of signs …” — In fact, the letters from Scrabble or the letters on a keyboard are already a series of signs, they’re not yet statements – so, “A series of signs will become a statement on condition that it possesses ‘something else’,” — “on the condition that it contains ‘something else’,” — in parentheses: “‘something else’ (which may be strangely similar to it, and almost identical, as in the example chosen)” — “‘[S]omething else,’ (which may be strangely similar to it …)”, this is very, very strange, and that is pure Foucault, “’something else’ (which may be strangely similar to it, and almost identical …” — So: “A series of signs will become a statement on condition that it possesses ‘something else’ (which may be strangely similar to it, and almost identical, as in the example chosen), a specific relationship that concerns itself.” You see: a series of signs, A Z E R T, becomes a statement on the condition that it contains “something else”… What is the “something else”? The same signs on the keyboard, which are not themselves a statement, yet the statement is strangely similar and almost identical.
Why do I say “this is pure Foucault”? It’s pure Foucault, because if there’s something, a problem, a running problem, or an amusing problem… a fascinating problem that haunted him; everyone has problems that fascinate them, and for Foucault it’s the problem of the double. What is a double? And we won’t be able to get anywhere in our attempt to explicate Foucault if we don’t pass through this issue, which is the issue of the double, and the problem of the double. And that haunts him, it haunted him from beginning to end. What is a double? What does it mean to have a double? So, yeah, something else “strangely similar” and yet other, “strangely similar and quasi-identical.” This is a first for us: the first time we see the existence of the double surface in Foucault. The statement is the double of something that is “strangely similar and quasi-identical” to it. AZERT. The statement “AZERT” is the double of “AZERT” on the keyboard. And yet one is not a statement, while the other is a statement.
Well, then we say to ourselves, OK, we get it, he’s saying something banal. What would it be like to say something banal? It would be, for instance, saying something like: for there to be a statement, well, it has to be spoken or written. So, the letters on the keyboard are not a statement, but if I say the letters on the keyboard or write them down on a sheet of paper, then I am making a statement. To make a statement would imply: speaking or writing. In other words, it would be talking about something that exists. In other words, it would mean: for there to be a statement, there has to be a copy. I have to copy the sequence of letters as they are on the keyboard, or I have to copy the letters that I picked at random. At that time, there would be a statement. This is stupid. Why is it stupid? If only because the keyboard itself, containing the letters that are not themselves a statement, is itself is a copy. Every French machine copies the French model of that machine. So if there is a copy of the conditions of enunciation, then we’d have to say: the letters on the keyboard are already statements. So that doesn’t work.
Would it be better to say: well, yes, we understand that in order for there to be a statement, there has to be a designation, and when I copy the letters from the keyboard I have a statement, because I have an instance where something is designated, and what is this something? Well, it is an instance that designates something else that is strangely similar and quasi-identical, namely: the letters on the keyboard.
So, I would say, at this point: yes, there is a statement when there is something that designates. Or, what comes down to the same thing from this point of view: there is a statement when there is something that signifies. And I would say: the designated “A Z E R T” on the keyboard is not a statement; however, when I copy it on paper…You see, I am no longer defining the statement by the condition of copying, I am defining it by the condition of designating, because the second series designates the first strangely similar and quasi-identical. That would be stupid too. Why? It would not be stupid on one condition, which would be if I were able to define the designation or signification without presupposing anything about the statement. Maybe it’s possible, I don’t know at all … [Interruption of the recording without any missing text]
Maybe it’s possible, I don’t know at all since, the statement is presupposed in traditional definitions of designation and signification. So, I cannot define the statement by designation, nor by signification, for the simple reason that these are dimensions of the statement that presuppose the statement itself. The statement itself is what designates. The statement itself is what designates. My second remark, after having thought that I understood the matter too quickly, well, it collapsed.
Someone will say to me: well, the statement must be defined by what all the other dimensions presuppose, both the designation and the signification, namely: the statement must be defined as a signifying chain, because the signifying chain does not presuppose the statement; it is constitutive, or it can pass for being constitutive. That doesn’t work this time either! Because if I define the statement by the signifying chain, what will prevent me from saying that the signifying chain is already on the keyboard? And now I’m back to square one. And, going back to square one is like saying… I take my head in my hands and I say to myself: what is this “something else”? If the statement is fundamentally related to something else that is strangely similar and quasi-identical, then this “something else” is neither a designated, nor a signified, nor a signifier. What could it be? We’re starting over. So, this long path we took—how did it serve us? We used it to create impasses. We know that we can’t, we can’t… And then a word emerges, a word that Foucault… to which Foucault attaches a lot of importance, but that, oddly enough, he barely comments on.
It’s enough simply to take a look at the table of contents at the beginning of The Archaeology of Knowledge, and we see that, basically, the first major part—since it is the title of a part and not of a chapter—is called “Discursive Regularities”. Discursive regularities. And then we see that, in the last part, in a chapter, Chapter 2 of the fourth part is called “The Original and the Regular”. And what is the theme of this chapter on the Original and the Regular? Basically, it consists in saying that when you want to define a statement, and he doesn’t say that it’s unimportant, but when you want to define a statement, there’s something that is unimportant, which is the criterion of the original versus the banal. When you want to know… again, you shouldn’t go too fast, because you shouldn’t conclude that Foucault is saying that a proposition said a thousand times, or a statement said a thousand times is a new statement. No, he says that the criterion of novelty or banality doesn’t constitute the statement itself. A banal statement is no less a statement than an original statement. In other words, “banal/original” is not a relevant distinction when it comes to knowing what a statement is.
This matters to us. Why? Because we saw that the statement was related to “one speaks”. [Pause] Well, the “one” is no more banal than original. The “one” is not the “one” of banality. The distinction between banal and original is not relevant to the “one speaks” as a condition of any enunciation. Good. So, the statement is neither banal nor original, it is regular. “Regular”, what does that mean? That it obeys rules. What are these rules? We saw that these are very specific rules when we commented on the nature of the statement… we felt the need… I felt the need to borrow [William] Labov’s term “facultative rules” as opposed to “obligatory rules”.[3] Thus, these are strange rules. Not every rule is a rule of enunciation. No doubt the statement involves very specific rules, the very rules that have been called “facultative rules”. The fact remains that the statement is a regularity. Alright. That would be like saying: alright, it’s a regularity, not just any regularity. So, it seems that, when I said, “facultative rule”, it implied for us that these rules are defined in relation to something, but in relation to what? Which would not be the same as obligatory rules. In other words, the enunciative rules are defined in relation to what? In relation to what? Not the original, not the banal.
The rules themselves are defined in relation to what? Now we’ll move on, maybe. Maybe…I’ll say: enunciative rules are rules that are defined in relation to singularities. They’re in relation to singularities. Ha! I seem surprised that I’ve made so much progress. And yes. Because: wouldn’t this already be a way of confirming the distinction between facultative and obligatory rules? Facultative rules concern singularities that they regularize. While obligatory rules always concern the universal. That would be convenient. That would be a solid confirmation. But we’ll leave that. Enunciative rules would concern singularities. I will say right away that Foucault rarely uses the word “singularities”, but he does use it. For example, in “The Discourse on Language”, you’ll find the following sentence: the “logos […] elevating singularities into concepts,” to the level of the concept.[4] And understand, even if you cut out the context, he is critiquing the logos, because the concept is universal. The logos elevates singularities to the level of the concept, that is, it transforms them into universals.
Then, here and there, Foucault uses the word “singular”, or “singularity”, but at the same time we cannot say that he makes a big deal out of their terminology. I think what happens is even more beautiful. You know, among philosophers there are always two kinds of terms at the terminological level. There are terms whose importance is explicitly noted, [Pause] such as “statement” for Foucault. In this case, he explicitly tells you: Take note—what I mean by “statement” is not what it is meant by “sentence” or “proposition”. And then there are terms that the philosopher doesn’t feel like s/he needs… that s/he uses and does not feel the need to tell you to “pay attention”. And s/he slips it into some part of a sentence, like it’s no big deal. It’s up to you to sort out for yourselves. At that moment, these concepts are no longer explicit concepts to this philosopher, they’re implicit concepts. They are no longer concepts that announce “look!”, but are concepts that “wink” or “glance.”
I’m coming back to my theme. There you go. First of all, [Pause; sound of Deleuze moving over to the blackboard] let’s make it simple, yeah! I mean that sounds like mathematics, but it’s not, it’s not… [Pause; he writes something on the board]. There you go. What did I do? I made an emission, an emission of singularities, three singularities, there you go. Or, as they say in mathematics, I marked three singular points. These singular points, as they are…good, I drew three singular points on a plane. One of the singularities is not on the same line. I could have made another emission of singularities, I would have… [he writes on the board]. It was a completely different emission of singularities. That’s where I made my emission of singularities, that one. Then that’s all. You will notice that my singular points are indefinite. [Pause] Ahhh. One more attempt.
I’m going to do something else. [Pause] I’ll do that, and then… Alright. What did I do? There’s a second figure. To put it simply, I would say: I have united the three singular points. I drew three lines. Good. We’re going very slowly because I think there are tons of traps. The more you think you’ve understood something right away, the less you will have understood it. At the same time, if you no longer understand any of this, it’s annoying. So, what can I say other than “I have connected my singular points?” I can say: I have regularized them. I have regularized them. Indeed, each of the lines is a line of regular points. A line of regular points unites one singularity to another or, if you prefer, I’ll introduce another term because it will be very useful to us, a regular line, a line of regular points, or, if you prefer, so I’m multiplying words, a series of regular points goes from the neighborhood — this is a mathematical notion, and you don’t need to know mathematics to know that it holds great mathematical importance — it goes from the neighborhood of a single point to the neighborhood of another single point. This is my regular line.
And the idea of a series appears. So, regularity is a series of points that go from the neighborhood of one singularity to the neighborhood of another singularity. Good. Can I… can I conceive…, at this level from before…, in my first figure, over there, my singular points were indefinite. Over there, when I regularized them, they receive a determination, namely, vertices of a triangle. But earlier, as indefinite points, they existed as singularities. They were indefinite. I had three singular points. Can I… Is it necessary that the regularization be triangular? I would say that triangular regularization is a regularity, that is, a form in which I regularized my singular points. Is that the only one, or—and let’s keep trying to fix the words, the words that will be extremely important terminologically—was another series possible with the same singular points than the “triangle” series? Oh yes, well yes, another series was possible. We see one right away…. [Interruption of the recording] [34:28]
Part 2
… And, in this second regularity — you see, it’s a second regularization — what will my third singular point be? My third singular point will be determined—it will be another determination—will be determined as follows: a point located outside the line AB through which I will draw a parallel line to AB. It will be [he writes on the board] another regularity. You can see that regularities are infinite. Because, from that point on, I might as well suggest drawing a secant. Are you alright?
Alright. I’m sticking to my two series. Practically, you see that I might conceive an infinite number of series. What is the relationship between my two series? Convergent or divergent? Meaning, the same family or different families? In other words, you can already see that in my triangular regularization, I actually have three series. But, strictly speaking, since these three series are convergent, I can consider them a single series. I’m moving on to regularization. I’m also treating it like one series. What is the relationship between the two series? Can I extend one into the other? I don’t know in advance. Maybe, maybe. On what condition? On the condition that a third series is set up, which includes the two previous ones.
Well, you know what it is. If you remember your elementary geometry, you know on what condition the two series will be extended. The answer is not given. Maybe they aren’t extended. In some cases, they aren’t extended They’re extended on the condition that you introduce a new regularization. What will it be? It’s… [he writes on the board] If you use one of the vertices of the triangle to raise the parallel to the opposite side. This will be the condition under which you will demonstrate that the three angles of a triangle are equal to two right angles. There you go. You will have made your series converge. Fine. Ah. It’s perfect. We’ve found almost everything. I mean, you can’t go wrong in this matter. What can we say now? Well, that’s it: a statement is a regularity; Foucault says it explicitly.
A statement is a regularity. What does that mean? What is it regularizing? It regularizes singular points. That is why it is a very particular regularity that is called a “facultative regularity”. It regularizes singular points, and to regularize means “to constitute a series that goes from the neighborhood of one singular point to the neighborhood of another singular point.” These series can be multiple. There will be as many statements as there are series. Will these statements converge? There’s no ready-made answer. You have to wait and see. Yes, if the series converge. No, if the series diverge. We saw this with “what is a family of statements?” If there is a convergence of the series, the statements will be from the same family. In other words, the statement is a regularity, but the emission of singularities is not a statement. The pure emission of singularities is not a statement. The statement presupposes it. If there is not an emission of singularities, then there is no statement.
Something else strangely similar and almost identical, what is it? The statement refers to something else that is strangely similar and almost identical: the emission of singularities. And, in fact, my indefinite singular points are strangely similar and almost identical to what the statement will be. The statement simply adds a regular line from the neighborhood of one of these points to the neighborhood of the other point. In other words, the statement contains “something else strangely similar and almost identical,” and yet this “something else” is much different from the statement. The statement is the regularity, it’s the series. All statements are serial.
Parenthetically, this is a great confirmation of a kind of anti-structuralism since Foucault never stops wanting to substitute the perspective of series for the perspective of structure. So, you can see that he is offering us a completely different solution. A Z E R T on the keyboard of the machine is not what the statement means at all. A Z E R T on the machine’s keyboard are the singularities that the statement will embody, the emission of singularities. When I copy A Z E R T onto the sheet of paper, I am doing much more than copying or designating what is on the keyboard. I’m regularizing the singularities. I’m making my series. The same thing goes for the handful of letters that are picked randomly in the game of Scrabble, and when I copy them down, I’m doing something other than designating: I’m embodying singularities, I’m regularizing them.
In other words, the small difference, the “something else”, which I’m calling a small difference—something else that can be strangely similar and almost identical—the small difference does not pass between the statement and what it is supposed to designate, nor between the statement and what it is supposed to copy, but between the regularity that it constitutes by itself and the singularities that it embodies [Pause] or it actualizes. [Pause] There you go.
Take a break. I mean, that has to be clear. What he did was… he did something wonderful, in my opinion. Within an outdated system of representation, the copy, designation, signification, and the signifier, he erected a kind of vertical dimension that redistributes everything, creating a new kind of distribution. So, obviously, what is troubling is that he does not explicitly emphasize this notion of singularity as much as he would have liked to do. In a way, it’s because it’s too close to him.
And we will see — I mean we have to wait — and in my opinion, we will see that in his entire canon, including at very different levels that no longer have anything to do with the statement, the notion of singularity is fundamental and for a very simple reason: it is the element of multiplicities. A multiplicity is a set of singularities, an emission of singularities. And it seems to me that all of his hatred of the universal and all his criticism of the universal remains incomprehensible if we do not see what he means, namely that things proceed in singularities. So why doesn’t he develop it? I think because he is able to consider a notion that has already been attained in mathematics and physics. There you go.
I would like for you to reflect upon it yourselves. Because, if it’s not absolutely clear, I’ll go over it again, yeah! I’ll go over it again: everything else depends on it, so stay alert for two minutes of intense reflection. What seems very, very curious to me, and what seems to go without saying, but, you know, if you want to skip it…you have to fully understand how this no longer has anything to do with a relation of designation.
If I say: the statement is a regularity that embodies or actualizes singular points, then you must understand that the serial conception of the statement immediately comes out, along with all of the problems that accompany it: if we take two statements, can you say whether they are from the same family or not? Well, they’ll be from the same family if you can extend the series of one into the series of the other. If you can’t, they won’t be from the same family. You see, everything depends on this. Ultimately, it’s a vertical construction, and a construction above all… You have: singularities…there, you could even put them in a kind of sky, we’ll see in what sense elsewhere, but singularities are in a kind of sky. Simply put, universal ideas are not what is in the sky, but singular points, small stars; it’s in this vertical construction that you have indeterminate singularities. It is a kind of Platonism of singularity. And then you have the statements that actualize them, that embody them by constituting series of regular points that go from the neighborhood of one singularity to the neighborhood of the other and that can go in multiple ways. Think about it. [Long pause]
Listen, that might seem radical, but if every statement… In fact, what I just tried to show is that every statement was a curve, and we could successfully draw the reciprocal out from Foucault’s formula: “A curve is a statement.” But conversely, every statement is a curve, a curve what? A curve that unites singularities. But nothing enables me to claim that all curves are convergent. If they are not convergent, I cannot say: the totality of curves is convergent. If that were the case, I would have irreducible families of statements. It is not at all certain that every series will come together into a single series, even if that series is infinite. [Long pause] No comments?
A woman student: [Inaudible remarks]
Deleuze: Oh, well, yes, political examples, we’ll get to that soon. Yeah, political examples are constant here. But I can’t give them yet, because… I can’t give them yet, but I promise you that I will. Actually, we are going to create a schema of non-mathematical curves… of curves that…yes, that’s right.
But I don’t want to leave mathematics right away because I need it. I need it a little bit. Because there is a great philosopher of mathematics named [Albert] Lautman, L-A-U-T-M-A-N. And in one of Lautman’s books I came across this: he comments on a famous text by [Henri] Poincaré, which says, “On curves” … A text by Poincaré called “On curves that are defined by a differential equation…”[5] It doesn’t even matter which “differential equation”; we don’t even need to understand. You will see how we can read mathematics extremely well, even at a very high level, without losing anything… [Laughter] But… while… above all, there is reason to be even more modest, because… But you can sense that the differential equation is essential for philosophy even without having done mathematics. That’s what Lautman says. “The theory of differential equations highlights two absolutely distinct realities.” Now I understand this again. Good. He announces that differential equations give rise to two heterogeneous, absolutely distinct realities. “There is a field of directions and topological accidents that can occur here, like the existence of singular points, for example.” A little further on, he says, “The existence and distribution of singularities in a vector field defined by the differential equation.”
What don’t we understand from that passage? Even if we admit that we don’t understand anything, there’s no need to… no need to… “Existence and distribution of singularities,” let’s say, okay, singularities—I’m really doing a minimal job to understand—points, points, points on a plane, like I did with my three points. I distributed singularities, I made them exist, I distributed them. “In a vector field defined by the differential equation.” The vector field, where did the vector field come from? It appeared when I had to choose between two organizations of singularities. Was I going to put each of them in relation to the other two or was I going to put the third one solely in relation to other two? These were two vector fields. You see, I don’t need to comment on “vector field”, I just need to recognize it myself.
So, I have an existence and distribution of singularities in a vector field. That’s one thing. And Lautman, along with Poincaré, tells us: and now there is something else and this other thing is the form of integral curves. As little as you know, you know that this relates differential calculus to integral calculus. We are told: there is the existence and distribution of singularities, which is the case in differential equations, but beware, the form of the integral curves is relative to what? Not to the differential equation, but to the solutions of this equation. Alright. And what is the form of integral curves? It is what determines singularities. And that is Poincaré’s main theme in this note. Singularities exist and are distributed in a vector field, but as indeterminate points. Singularities receive their determination from integral curves that pass through their neighborhood, and everything depends on what the integral curve does to the neighborhood. And I just have to go back to my example, which was rudimentary, but was shown to be totally consistent in this respect.
And I would say: in both my cases, I don’t have the same integral lines at all. So, in my two cases, the singularities are not determined in the same way. Since in one case they are determined in a triangle that I will call an “integral figure”, which passes through the neighborhood of my three singularities. In the other case, I have a completely different figure: parallel lines. For the two series to link up with one another and continue in the same series is possible, but that’s not my concern for the moment. It will be…because, effectively, a third series is needed in order for them to continue, as we have seen, and on that point Poincaré makes his great…the points will be determined based on the form that the integral curves take as they pass through the neighborhood of points. And I’ll reread Lautman: Poincaré distinguishes between a saddle, which is a determined singularity, nodes, foci and centers. Saddles, nodes, foci and centers—the words are very pretty, they are the names that singularities take on when they are determined by integral curves that pass through their neighborhood. Saddles, nodes, foci, and centers. Saddlesare defined in the following way: saddles are that by which two curves (and only two curves) are defined by the equation. So: two curves and only two curves. Nodes are where an infinite number of curves intersect. An infinity of curves. Foci are what curves rotate around while endlessly approaching them like a spiral. Centers are where curves appear in the form of closed disks.
Well, it doesn’t really matter. You see, we could christen singularities according to our needs and we will have to christen them. It’s only under the condition of the form of integral curves, or the equivalent of integral curves, passing through the neighborhood of singularities that singularities determine themselves or are determined. And, indeed, if we go back to my story and recreate my pure emission of singularities…There. [He draws on the board]. There. Three indeterminate singularities. Let’s suppose that we can do better. [He indicates the drawing on the board] I’m constructing an emission with only one singularity. But that doesn’t tell me anything about how the curve that passes through the neighborhood will look. There we are, one case. I would say, my singularity here is determined as a vertex. And we can conceive of another case. [Pause as Deleuze draws on the board]. This is even prettier. Yeah? I can conceive of yet another case… [Pause] another case. [Pause] What matters to me is that the singularity in these three cases will be determined differently according to the speed of the curve passing through the neighborhood.
Conclusion: a statement is not a structure, it is a function. It’s a function. A function that consists of what? It consists in regularizing singularities by drawing a curve that passes through the neighborhood of these singularities. And I see that it is very complicated, and given that the statement is a function, I can conclude, even immediately deduce, that the statement is serial. [Pause] With this one-time question: how far will a series go? Thus, the problem that Foucault posits in the beginning of the Archaeology, from the introduction of The Archaeology of Knowledge, and his deep interest in modern history as modern history, at least in one aspect under [Fernand] Braudel’s influence, has built a whole method that is referred to as “serial”. Establish series of capacities, variable temporalities, once it’s stated that any series is spatiotemporal. Good. That’s the point we’ve reached.
Still, it bothers me. We just distinguished two dimensions. It goes without saying that one does not exist independently of the other. Finally, singularities without an integral curve, or an integral curve that passes through their neighborhood: indefinite. Conversely, there is no curve that does not pass through a neighborhood of singularities. So, one is in the other, but that doesn’t mean they don’t differ in nature. The two differ in nature. One does not exist without the other; there is reciprocal presupposition, there is everything you want. We’re rediscovering all of our themes from the entire trimester. There is reciprocal presupposition, yes, but there’s still a difference in nature. There is immanence, yes, and yet there is heterogeneity. So, what would interest me would be managing to say a little bit more about indeterminate singularities.
What can I say about them? What can I say about them? I can’t even say what they consist of. I can’t say: they’re vertices of a triangle, since what constitutes them as the setting for a triangle is the regularity that embodies them. But can I say something about them in-themselves? What can I say about AZERT on the keyboard? Ah, what am I going to say about AZERT on the keyboard? So, I tried because this is much trickier than mathematics. I assure you, I did everything I could; I wanted a typing manual, I went so far as to call Pigier,[6] and they are rather unpleasant. I absolutely had to… I asked to speak to a typing teacher… I did all of that. I mean, I should have gone there myself because I came up with nothing, with nothing… So, these are assumptions I’m forced to make. Hypotheses. But any typist may know…if any of you… Anyway, what I’m about to say is wrong, but it’s easy to correct. You can replace what I say with the truth, if you find it, [Laughter] but it won’t change anything, absolutely anything.
I tell myself: AZERT, well, why is AZERT on French machines? You see, I’m talking about AZERT on the keyboard. So, I’m putting myself on the level of “pure emissions of singularities.” What does this depend on? In the case of Scrabble letters, it’s very simple, it’s very simple because I would say: it’s an emission of chance. There is a relationship between the letters I draw, and I would say that this relationship is random. Okay, you’re going to understand right away what I’m getting at. But a random relationship is a relation of force. Picking letters at random is a force relation between those letters. If I draw letters at random, and I have, for example, A K E, I can’t say that these letters are unrelated; they have a relationship, a random relationship. The random relationship is a force relation between the letters. I’ll keep that in mind in case there’s anything to be gained from it. Alright.
The French machine says: AZERT…ah. In my opinion—this is where I’m cautiously adding information where it’s lacking, since they didn’t want to give the information to me—in my opinion, you have to take into account… I can’t say that A Z E R T is with relation… There are tons of relations, what are the relations this time? These are not random relations. I think there are two things to consider if you want to understand the keyboard of a typewriter. It is necessary to take relations of frequency or attraction into account—which comes down to the same thing—relations of frequency in groups of letters, or attraction of one letter in relation to the others. On this point, linguists have carried out very detailed studies for each language on the power of attraction of one letter on the others and on the frequencies of a particular group of letters within a language. For example, WH has a high frequency in English. The frequency in French is zero. [Pause; Deleuze hums] Hmmm. G, the letter G in French: what does it attract? It attracts U and N with a relatively high frequency. It doesn’t matter if all that’s true or false, yeah. Alright.
The letters will be spread over the keyboard. The typist is supposed to achieve the ideal, i.e., to type with both hands. Notice that it’s already a vector field, huh? The keyboard has two sides. The keyboard is vectorized. It is vectorized into two sides—at a fluid border, but a left side and a right side. What does that mean? It already means that if you have a letter—BTW, I’m presupposing all of this—if you have a letter, if you have, for example, two high frequency letters, let’s say (and grant me this even if it’s not this and is something else) G and U. So, when you type a G, there is a considerable chance or a large number of cases where the letter will be followed by a U. It is obvious that it is a good idea to distribute the G and U across both sides. [Pause] Because, if you put, for example—look at our keyboard—if you put the G here and the U just below it on the same side, you would have to hit G and U with the same finger, which would be a huge waste of time. Thus, there are factors: time, the relationship between both hands, finger spacing and frequency relations specific to a language between letters that will determine your emission of singularities, which constitute the keyboard. Allow me to call this, this set (hand relations, finger relations between each hand, frequency relations, and relations of attraction of letters), force relations between letters and fingers. [Pause] Relations of frequency of letters and dynamic relations of fingers. I would say: this is what governs the emission of singularities on a keyboard. That’s what Pigier should have said to me. Alright.
On this point, Foucault says that the statement is a regularity. That is to say: as soon as—if only fictitiously, and it can be completely fictitious—as soon as I make a curve pass through an integral that goes from the neighborhood of one singularity to another, even if it follows the same order as the keyboard, even if I seem to copy down AZERT, I create a statement since I have embodied the singularities in an integral. Why isn’t the first AZERT a statement, while the second AZERT is? It’s because the first AZERT envisages the pure emission of singularities in a vector field defined by force relations, while the second AZERT embodies these same singularities in integrals, even if the integrals only happen to be fictional. [Pause] I integrated the force relations, and thus constructed a statement. [Pause] If you understand that, then everything is in place. Now, the transition from knowledge to power. And what will Foucault call “power”? What he will call “power” — and it is time to say it once and for all… no, we’ll repeat it — what he will call “power” is any force relation, no matter what it is, except he does not use the term “force relation” for just anything.
What is a force relation for Foucault? That’s very important. But one letter has power over another, and if you don’t understand that, then you won’t understand anything about Foucault’s political philosophy. Either a letter has power over another letter, or it doesn’t…a letter will have a power of attraction over another letter. G, assuming that G commands U quite frequently, or N; in English, let’s assume that W commands H quite frequently, you will say: W has a power of attraction. Good. All force relations are power, and power alone consists of force relations. Between two terms there are force relations, and you can say: one exercises power over the other or both of them exercise power over each other.
How do we get from knowledge to power? We have our answer, at least. We move from knowledge to power insofar as the statement forms knowledge, is an integral, operates the integration of singularities, and it is only at the end that we realize that these singularities as such maintained power relations and force relations amongst one another. In other words, knowledge is the integration of force relations, in the most general sense, which is force relations between things, between people, between letters, between light, between shadow and light, between everything you can think of. This is why Foucault will be able to create a political ontology. [Pause]
I would say that we are now able to distinguish between force relations that constitute power and the relations of form that constitute knowledge. Relations of form are the appearance of integral curves that actualize, and what do they actualize? They actualize singularities, the singularities that sustain force relations between them. [Pause] Hence the claim from before becomes particularly urgent: an example from something other than mathematics or linguistics, and notice that we have given another one, a linguistic example with AZERT, where we see that letters exercise… [A brief jump on the tape] The statement varies in force relations that the statement will regularize.
Are you alright? It’s, it’s… this thought seems really extraordinary to me. Right now, if you want, it’s…if I were asked…oh there are so many utterly novel things in Foucault, but this is one of the most remarkably novel points is this analysis, and it seems to me that this analysis is very extraordinary. Then why…? Yeah, fine…Well, you get some rest…What time is it? 11:00… [Interruption of the recording] [1:18:32]
… So, someone just made a very fair remark. He says this is all well and good, but if we already introduce the vector field at the level of singularities; in other words, if we already take singularities to be force relations, then it’s almost the same thing as taking them at the level of integral curves that pass through a neighborhood. It’s not untrue, it’s very…the embodiment is very… it doesn’t prevent force relations… it’s not yet the appearance of curves that pass through a neighborhood. But there is a kind of intertwining of the two. But the appearance of curves won’t be defined by force relations. Moreover, you can see that when I invoked relations of frequency between letters in a given language, for instance, or relations of attraction between one letter and another, it was a question of any relations whatever, independent of a given integral curve.
But that doesn’t mean there’s not a kind of… exactly as we saw between the visible and the statable. If you will, everything that has been said throughout this term about the two forms of knowledge, that is, the perpetual interweaving of the visible and the statable, which although they differ in nature, one does not cease to arouse the other, to capture the other. Foucault will say exactly the same thing about power and knowledge. Here too, there will be a mutual presupposition of the two, but a relation of forces is not a relation of forms. For the very simple reason that a relation of forces—we’ll see this after the new year—is fundamentally non-formal. While integral curves always define forms. But ultimately, we will only be able to see this very gradually.
So, the main thing is that you intuit…I always do, I always appeal to your intuition because this is all new material…if it wasn’t new, I mean… and there’s a mode of philosophical intuition without which you wouldn’t understand… [Interruption of the recording] [1:21:15]
Part 3
… of pseudo-mathematics. Although, I think that when it comes to philosophy, the fact that mathematics includes a mathematical theory of singularities in its most important chapters is one of the great intersections between mathematics and philosophy. And this has always been the case. To me, it seems impossible to understand a philosopher like Leibniz without taking into account the dual affiliation of philosophy and mathematics in the notion of singularity and singular points. You see, every philosophical theory that reacted against the universal could only do so in the name of singularities, and singularities already understood in the mathematical sense, and we will see why it’s so important to Foucault to carry out a criticism of the universal as far as possible. Very good, well, well, well…So, yeah, there you go.
There are going to be all kinds of practical problems, because, you see, the method—The Archaeology of Knowledge is a book about method—so if I summarize it by saying: well, yes, it’s about constructing integrals, integral curves in the neighborhoods of singularities, then I think you’ll understand that I’m not claiming that Foucault is doing mathematics. Because he will directly apply this method to all other domains. And he’ll have the right to do so, since he will have revealed the conditions under which this method is not limited to mathematics.
So, let’s take an example. What kind of example? A social example this time, an example from the social field, because I made it clear that there is no reason to keep force relations confined; to repeat, there are force relations between the letters of the alphabet. Well, let’s take an example from the social field this time. Are there any singularities in the social field? Obviously, there are singularities in the social field. What are singularities in the social field? Are there any singularities in the aesthetic field? Yes, there are… there are plenty of them in the aesthetic field. After all, isn’t that the definition of thinking? To think is to emit singularities. If that were a definition of thought then we would understand [Stéphane] Mallarmé’s “A Throw of the Dice” better, we would understand Nietzsche’s call to play dice. To think is to roll the dice.[7]
What are singularities? These are dots on the face of the die. The face that results from a dice throw. So, does that mean we can think anything? Nothing at all. In fact, it means that we can’t think just anything at all, since the singularities I emit must form beautiful integral curves that I don’t know in advance…there are always risks. Take a stupid thought… a moronic thought [Laughter], I roll the dice, but nothing comes out. It’s a bad toss. I would say… thinking is throwing the dice, which once again means that, yes, chance itself is a force relation, a force relation between the dots on the faces of the dice and what comes out of them? Well, maybe the integrals of philosophy are concepts, but a concept is not a universal; it’s an integral of singularities. Then there would be noetic singularities, singularities of thought? Yes, there would be noetic singularities, and to do philosophy would be to throw the dice. Alright, good. So, there will be a philosophical field with singularities. And that’s what I’d use to create concepts. Or else I wouldn’t create anything at all. Good.
But, so, let’s get back to the social field. There, I’m throwing something, I’m making a little constellation. So, of course, I am forced to appear to contradict myself, but it’s only pretense. Because I’m going to name these singularities, otherwise we wouldn’t understand anything. I am forced to name them, so I am already prejudging the integrals that will unite them. But you can correct it on your own. I’m going to make a little constellation…Oh, no. See, that’s how I do it. I’m throwing…I’m emitting a singularity. I call it “confession”. I’ll say: it’s a point of confession. There you go, it’s a singularity, a point of confession. [Pause] I’m emitting another one, there, a little higher up…a point of sacrament. [Pause] Down below, I emit a third, a point of guilt. [Pause] On the left, on the right, I’m emitting one last one, I can’t take it anymore: a point of memorization. There, that makes four, five. I can define force relations. I can define force relations between these points in a vector field. A first one: my point of confession is typically taken in a priest-user/confessor-confessed force relation. It’s a force relation in the broadest sense: it doesn’t mean the confessor slaps my face, does it? We already saw that that’s not what a force relation is. Notably, lines of attraction are force relations. Attractions are typically exercises of force. Sacrament… well, all that… Sacrament, guilt, memorization…
You see that I can make a curve pass through the neighborhood of all these points. I’m doing it hypothetically. I tell myself: oh, well, yes, there’s still something to see. What does the curve tell me? The integral will pass through the neighborhood of each of these points. Well, that’s how I start from the “point of confession” and then draw a line to the “point of sacrament”. In fact, I will have had to confess for communion, to receive communion. It’s more like the force relation of each point in the vector field, an integral curve that goes from the neighborhood of the first point to the neighborhood of the second.
Sacrament and confession: I can set up two regular lines that reach towards guilt. The sacrament is a way of redeeming primordial guilt. Confession is the declaration of secondary guilt. Well, again, if what I say is wrong, you correct it on your own, it doesn’t change anything. Memorization: the examination of conscience that precedes confession. Well, I can push my integral, my line that I’m calling the “integration” or “actualization” of singular points. How far can I… how far can I push it? How far? It is extremely variable.
Here’s the first case: I push it until I’m able to call it a special curve, that’s right, which would set the end of the series. I would say: a series ends if I can assign what mathematicians call an “envelope” (these are very convenient words) among and around the set of integral curves that actualize the series. Let’s put it this way: it’s a curve that envelops all the others. An envelope: all of these terms are pretty… it’s very pretty, an envelope of singularities. That’s good. And, simply, does such an envelope take place? There are cases where there is no envelope. It’s like mathematics, I guess. There are cases where there is no envelope. There are cases where there are. At the end of his life Foucault became more and more interested in what he called “pastoral power”. And I think the unpublished book, Confessions of the Flesh, analyzes the formation of this particular church power, pastoral power.
Pastoral power is an ancient idea that can be found in Plato, namely: grazing as the model of government. Grazing a flock is the whole theme of Plato’s Statesman…What is a good ruler? He’s the pastor of a herd. It may seem like nothing, but it is a fundamental political problem: is power pastoral? It goes without saying that Christianity takes advantage of the idea of pastoral power in its revival of Platonism with the Fathers of the Church and alters it in ways that obviously differ greatly from Plato, since they are Christian ways. And pastoral power becomes, above all, a new type of power that state power did not fulfil at that time, which, perhaps, prefigures the future State. It could be defined in the following way (how would the force relation of pastoral power arise)? Control of everyday life.
Control of everyday life. Management of everyday life. The pastor of the herd…the human multiplicity, the human community is likened to a herd such that the pastor must take charge of the everyday details of existence for each member of the herd. This is a type of power that has no equivalent. Pastoral power is completely different from royal power. The king does not care at all about the everyday life of his subjects. The pastor takes care of the everyday life of his herd and what takes place in the mind of the herd. The king doesn’t give a damn what’s going on in people’s heads. Well, I would say: here, the pastoral schema is encompassing, it’s an envelope. [Pause]
Can I say: the series ends, therefore, it is closed by the pastoral schema? Yes, yes, from a certain point of view. Can’t it be extended? From a certain moment in history, it’s likely that Foucault would situate the transition of the pastoral power of the Church to State power, which retains the model of pastoral power, at the end of the 18th century and the beginning of the 19th century. Specifically, one of the fundamental claims of the church’s power is to individualize those to whom it relates, to individualize its subjects, and, by the same token, to be able to control them in their everyday lives, and the State also makes this its objective through completely different means. There is, therefore, a kind of relay where pastoral power is taken over by State power, even if it entails very important changes. State power demands, or positions itself to demand, the individualization of its subjects. Good. At that moment, I can say that under this condition my series extends beyond pastoral power. There will be convergence between the pastoralized series and the state series. Between the Church series and the State series. Alright. We’ll settle for that.
What does that mean? What would the method for analysis of the social field be? The method of analysis of the social field is: to fix the singularities that are present in this field as they enter into force relations that constitute the vector field. So: to fix the constitutive singularities of such and such a social field, i.e., those that enter into force relations that correspond to this social field. The second point is to construct the institutional forms, i.e., the integral curves that actualize these force relations. The sacrament, the confessional, the power of the Church as an institution… well. Insofar as force relations and singularities are actualized, are considered to be actualized in these integral curves and in these institutions, they constitute real kinds of knowledge. [Pause] All this knowledge will be developed at the confessional level as casuistry, at the sacraments level, at the level of the Fathers of the Church, and at the level of what can be called “pastoral knowledge”, in general. And, to the extent that the force relations of singularities are embodied in these curves, statements emerge.
We’re discovering our solution again. I am looking for sexual statements in the 19th century. All we have to do now is start over again, but I think we have already taken a big enough step forward that everything is much clearer now. I’m looking to construct a corpus of sentences about sexuality and words that speak about sexuality at this particular period in time. How do I construct my corpus? I’m looking for singularities as foci of power. Focus is a bad word: there are centers of power, nodes of power, saddles of power, or whatever you want…they’re singularities. I assign my singularities, which are foci of power. I’m causing my curves to move. These curves are forms of statements. They are forms of statements that carry knowledge in themselves.
Third question: so, you see the two aspects of the serial method. First aspect… I’ll start again: you assign singularities and force relations according to where they are taken. That’s the problem with power. Second aspect: constructing integral curves, i.e., institutional integrations that produce statements. That’s the aspect: knowledge. I’m building my series. Second aspect. Third aspect: when does a series end? Variable response. It all depends on the level. Once again, there is a whole series that ends with pastoral power, but which, from another point of view, converges with State power. You can cut it off, depending on your goal, you can cut it off at some place that is closer or further away. Sometimes the duration will be short; since every series is spatiotemporal, you’ll have some series with short duration, and you can also construct series with long duration.
Note for the future that this is a problem for Foucault: Foucault has always preferred series of short duration. If you look at all of his books, except the last ones, you see that he studies periods of short duration and hates periods of long duration, because he is afraid that they will fall back into universal history. So, at most, they’re series taken over two centuries. The History of Madness; even Discipline and Punish is a 50-year series. They’re short series. And we’ll visit this problem again, except with The Use of Pleasure, where Foucault’s conversion to the long series, to long duration, bursts forth. It’s about tracing the history of something that begins with the Greeks. It is quite unusual for such a long period of time. From the Greeks to us, while passing through the Fathers of the Church. The History of Sexuality continues from the second volume of a long duration. What could have happened? If we have to start with something very specific to understand, we could even ask the question: How did Foucault change between The Will to Knowledge and The Use of Pleasure? I think that a good way to ask this question, because it involves concrete detail, would be: what could have converted Foucault toward dealing with a large series, a long series?
So, this is like the third aspect: when does a series end? And on this point, we can see quite clearly what Foucault owes to the historians of his time. We can see quite clearly what he owes to Braudel, since Braudel always dealt with series, constituted historical series and, furthermore, he distinguished series according to the length of time that they spread out. And Braudel’s entire conception of history, as you know—perhaps I will speak more precisely about it further along, later on—is to distinguish between three kinds of duration: short duration, medium duration, and long duration, which coexist with each other. We will have to ask ourselves what the distribution of durations is with respect to series in Foucault’s thought. That’s a lot of trouble, isn’t it?[8]
So, I gave an example. My example is: related to sexuality, how do the foci of power locate themselves within singularities, in force relations that will be actualized in processes of integration, these processes of integration being constitutive of knowledge? Well, that’s Foucault’s general theme. I’ll take two examples. Two other examples that Foucault summarizes in The Archaeology of Knowledge, the example of psychiatry (pp. 233-234) [179-180]. He says: “what made it possible” – psychiatry — “what made it possible at the time that it appeared, what brought about this great change in the economy of concepts [analyses and demonstrations], was a whole set of relations…” — be careful, you will see that the terms of these relations are not knowledge — “…was a whole set of relations between…” — now I’m starting my emission of singularities again — “…between hospitalization [1], internment [2], conditions and procedures of social exclusion [3]…” — procedures of exclusion are not the same as internment — “…the rules of jurisprudence [4], industrial labor norm [5] […] in short, a whole group of relations that characterized for this discursive practice the formation of its statements”.[9]
It can’t be said any better: it is the constellation of power foci, i.e., the constellation of singularities, which will make it possible to draw the curves that constitute knowledge. Good. That’s exactly… it should be said that the social field emits a roll of the dice. You’ll tell me: Oh well, the social field emits a roll of the dice, but isn’t that starting from scratch? No, it’s not starting from scratch. Undoubtedly, the roll of the dice for each social field is partly determined by the state of the forces of the previous field. What is it? I say that because we might see that again later. I’m saying it now for those who were here last year that it’s exactly what is referred to as a succession of semi-dependent events, or what is referred to as a Markov chain. Successive re-enchaining. Each time something is drawn at random, but according to the data of the previous draw. A succession of random draws that partially depend upon one another. Well, that constitutes a Markov chain. Social mutations can be conceived in the form of a Markov chain.[10]
It’s the same analysis that Foucault gives of pathological anatomy. Pathological anatomy is a form of knowledge, it is a kind of knowledge that is constituted and formed in the 19th century, at the beginning of the 19th century, and at the end of the 18th century, but what was there before that? Similarly, one can ask: what was there before psychiatry? There was no psychiatry, there was something else. It was a complete redistribution of the previous field that made psychiatry possible. A new draw. Before pathological anatomy, what was there? There was the clinic. The clinic was the conquest of the 18th century. It takes a complete redistribution of clinical foci for pathological anatomy to be possible as a form of knowledge.
On pages 213-214, he says this, which was demonstrated in much more detail in The Birth of the Clinic – [Pause] Hmmm, I’m not going to find… [Pause] — here it is: pathological anatomy will discover a new field, a new object that will be an object of knowledge, and that is tissue. Tissue is… tissue is a great discovery for biology, for medicine… Good. Pathological anatomy is formed around tissue, taking tissue as its object. But “the [preliminary] fields are constituted by the mass of population under administrative compartmented” — You will ask: how does this relate to tissues? Well, yes it does, The Birth of the Clinic clearly shows what the relationship is between the discovery of tissues and this kind of data — “…the mass of population administratively compartmented and supervised, gauged according to certain norms of life and health” — you see, those are all power relations — “…analyzed according to documentary and statistical forms of registration; they are also constituted by the great conscript armies of the revolutionary and Napoleonic period […]; they are also constituted by the institutions of hospital assistance that were defined at the end of the 18th and beginning of the 19th centuries…” etc., etc.[11]
You can see that Foucault proceeds each time by making his constellation of singularities, questioning the force relations that vectorize these singularities, and then constructs these series that are constitutive of knowledge.
Good. I can summarize in very broad terms by saying… except, if you followed me, we’ve only done half. We’ve only done half, because: what did I just demonstrate? Well, yes, curves—now I can say curve-statement with a hyphen, so I would say: any statement is a curve-statement—and I would say: curve-statements actualize the force relations or power relations between singularities. Actualize, embody, etc. We don’t know what word to use yet, we’ll only figure that out later. Yeah. Are you alright?
But, I’m saying, I’ve only done half of our task. Only we’re so tired that the other half is going to go fast. Because, as you recall, [Pause] knowledge has two irreducible forms. The production of statements and production of light. Knowledge is light no less than language is. Knowledge intertwines light and language and much more, and we wondered how it could do so since the light-form and the language-form have nothing to do with each other and are irreducible. We arrived at this cruel problem, since, last time, everything brought us back to this: if you stay in the dimension of knowledge, you will never understand how the two forms can be intertwined. But you can see that we have the solution. We have everything we need now. That’s good. But, at one point, we were too tired to be happy about it, as always. [Laughter] On the other hand, I can’t move on without saying the same thing.
On the other hand, it is also necessary that, on their behalf, luminosities integrate singular points that are taken up in power relations, in force relations. In other words: just as statements are curves, visibilities are tables. And yet, there is something that is bothering me, which is that Foucault uses the word “table”, which he uses so often in a much more general sense than suits both curves and visibilities. It doesn’t matter. It doesn’t matter. If we try to reserve a particular meaning for the word “table”, then we will have to say: well, yes, visibilities integrate singular points into tables and not into curve-statements. Into visibility-tables. Visibilities are tables no less than… That’s why visibilities are never things. As we saw last time, for Foucault, particularly with regard to Raymond Roussel, visibility is the label on the Evian water bottle, the letterhead of a grand hotel. Visibility is always a picture. Why? Because visibility is a being of light before it is a solid being.
Well, light, just like the statement, is an integration of singularities, singular points. And you can only define a light…and the path of a light as going from one singularity to another; that is, there are luminous series just as there are verbal series. I want to comment more on this point, but we can’t take it anymore. On this subject, I’ll point you toward two kinds of text. The famous description of [Diego] Velázquez’s painting, Las Meninas, which I’ll ask you to read from the following point of view—I am not saying that this is the only point of view to read this text, it is one possible point of view—from the following point of view: how do Velázquez’s lines of light unite and pass through neighborhoods of singularities? What will the singularities of Velázquez’s painting be? You will see that they are not reduced, they are multiple, they follow the very path of light, the way in which the path of light curves, has vertices, i.e., passes through singularities that distribute reflections, brightness, etc. And everything culminates in the force relations between two masterly singularities, two dominant singularities: the painter and his model, the king. I’m not saying that everything can be reduced to this; on the contrary, there is a whole development of an extremely varied pictorial field populated by singularities, but certain singularities are dominant.
The two dominant ones are: the painter and his model, the painter’s gaze that sees without us seeing what he sees, and the king’s gaze that sees without being seen. I would say that it is the relationship between these two singularities, the force relations, of the painter and the king. We can ask the question: which of these two forces is stronger? The strength of the painter or the strength of the king. It all depends on the point of view. In any case, that’s what will mark the closing of the painting. This will be the envelope of the painting: the standoff between the painter and the king, but this standoff passes through the distribution of the baby, the dog, the jester, etc., etc. And you have the light of the painting which is the integration of all these singularities in a mode, in a certain mode that is Velasquez’s mode. You can conceive of other modes. If you refer to the book on Raymond Roussel, around page 150, before and after page 150, as we saw when I commented on it last time, there’s a great passage where Foucault analyzes visibilities in Roussel. Here, you have a visibility regime of a completely different type than Velasquez’s, which this time proceeds bit by bit.
While describing the bottle of mineral water, the label of the mineral water bottle, Roussel proceeds in a kind of local construction that gradually spreads, where he is constantly saying: on the right we see this, a little in the background we see that, etc. just as Foucault says, it’s as if we were moving from one niche to another niche, as if we were slipping into a succession of tiny rooms, and this succession of tiny rooms will constitute the path of light, a whole other regime of light. If you wanted to do practical exercises related to what we’re doing here…maybe we’ll do it later, you’ll take painting regimes and you’ll wonder how they are different regimes of light and what kind of singularities they include, what kind of force relations these regimes of light have.
Because, after all, there are force relations, just like I said earlier, and you know there are force relations between the letters of the alphabet. In this sense, we can talk about a politics of language: yes, from the level where there are force relations between the letters of the alphabet. There are also force relations between colors. Force relations between colors…it is even…force relations between colors…colors…you can conceive of them as singularities united by force relations in a vector field. And, if you consider any kind of color scheme, you cannot define things like cold and hot, for example in terms of color, without using forces. If there is anyone who has shown this definitively, it is [Wassily] Kandinsky, who is only able to present colors according to the forces affected by each one, as the forces affected by each one already determine the relationship between two colors, and the painting will be the integration of this and that force relation between colors. This is why Kandinsky’s paintings are poorly referred to as being “abstract”.
So. I can say that…well…we have our solution. You’ll recall our solution: but how can the two forms be intertwined? How can the two forms of knowledge, the visible and the enunciable, intertwine when they have nothing in common? Our answer was: well, they can only be intertwined in an instance that arises from another dimension. Now we have it. We have it, the instance that is in another dimension and that explains the intertwining of the two forms of knowledge is: the distribution of singularities and force relations between singularities, which I would call a non-formal dimension, the non-formal dimension of force relations as opposed to the dimension formed by form relations, by form relations. However, we can only finish if we understand what it means to say “force relations are non-formal.” We still have a lot to do, because how “force relations are non-formal” is a mystery.
But above all, you can see that when Foucault uses the term “force relations”, he never, ever means “exercise of violence.” So, what does it mean, since the force relation is non-formal and does not consist of violence, that is, in the destruction of form? This will be our object of study when we return next term. All I can say in conclusion is that it’s very simple.
What is in common or what is, undoubtedly, the deepest point in common between Foucault and [Maurice] Blanchot? I would say the deepest point in common is to have established, in two very different ways, a set of intimate connections between the following three notions: the neutral or the “one”, as one part; the singular, as another; and the multiple, as yet another. The neutral, or “one”, is opposed to the person. [Pause] The singular is opposed to the universal. The multiple is opposed to one and the same. The three notions are strictly… [Interruption of the recording] [2:07:48]
Part 4
…are emitted. It is the form of the distribution of singularities. [Pause] When I said: to think is to emit a throw of the dice, well, the throw of the dice is always emitted into a “one”. “One” thinks. On the side of the singular, the singular is not opposed to the “one”, since it is pre-personal, it has nothing to do with a person. Moreover, it would not be difficult to show that a singularity is already a force relation. In other words, the real subject is force. And that’s where Foucault ends up being Nietzschean. And the only misinterpretation to translate is: there is violence everywhere. That is so little of Foucault’s thought. Foucault absolutely separated force relations from the effects of violence. We’ll see why.
Finally, the singular is not opposed to the multiple. We will call a “multiplicity” a constellation of singularities in the “one”. A distribution of singularities in the “one” is precisely what a multiplicity is. I think that Foucault assigned very precise relations and status to these three terms around which Blanchot revolved as the three components of his thought, which was not Blanchot’s objective.
So, we have our answer, once again: what is this other dimension that is the only one capable of ensuring the intertwining of the two irreducible forms of knowledge? It is force relations or power. They are the ones that are embodied in the visibility-table, as they are embodied in curve-statements. It is, therefore, in the non-formal elements of the force relation and singularity that the two forms of knowledge can find a reason for their intertwining. Hence the need to move beyond knowledge towards power, even though knowledge and power are inseparable from each other, such that Foucault speaks about an inseparable complex of the power-knowledge system. What could possibly happen so that at the end of his life he would still discover a third dimension and why did he need it? That will be our guiding question. There you go, have a good holiday, get some rest. [End of the recording] [2:11:15]
Notes
[1] Despite the emphasis that Deleuze places on this example from Foucault, Deleuze refers to it in Foucault only briefly, pp. 78-79. However, he introduces this example in the opening chapter, “A New Archivist”, p. 2, and returns to it precisely in the context discussed here on pp. 11-12. He also refers in passing to the example in the chapter on subjectification, p. 98.
[2] The page numbers are to the French edition of L’Archéologie du savoir (Paris: Gallimard, 1969), with English edition references in brackets, trans. A. M. Sheridan Smith (New York: Harbert & Row, 1972).
[3] On these rules, see Foucault, p. 147, note 18; in a note from the first chapter, “The New Archivist”, Deleuze cites Labov’s work in Sociolinguistic Patterns (Philadelphia: University of Pennsylvania Press, 1972).
[4] In L’Ordre du discours, the French text reads “le mouvement d’un logos qui élève les singularités jusqu’au concept” (Paris : Gallimard, 1971), p. 20, hence the English translation would be “a logos that elevates singularities all the way to the concept”. We follow the translation of this passage, p. 228, in the text, “The Discourse on Language” published as an appendix to The Archeology of Knowledge (cited above, pp. 215-237), translation by Rupert Sawyer.
[5] In fact, the title of this early text by Poincaré is Sur les propriétés des fonctions définies par les équations aux différences partielles (1879) [On the properties of functions defined through equations with partial differences]. As for Lautman’s book, the full title in French is Symmétrie et dissymétrie en Mathématiques et en Physique: Le proble du temps (Paris : Hermann, 1946) that Deleuze cites partially in Foucault, p. 143, note 13.
[6] This was a well-known commerce school.
[7] This is a reference to Mallarmé’s poem, “Un coup de dés n’abolira jamais le hasard” (1897) [A toss of the dice will never abolish chance”] On this aesthetic analysis and references to Proust, Velasquez, Magritte, and Roussel, see Foucault, pp. 79-81.
[8] On Braudel and the types of duration, see Foucault, p. 148, note 31, where Deleuze refers to Braudel’s Écrits sur l’histoire (Paris: Flammarion 1977).
[9] L’Archéologie du savoir, p. 233; The Archeology of Knowledge, p. 179.
[10] On these chains as they relate to discontinuities and mutations, see Foucault, p. 86, and also Chance, see Foucault, p. 117.
[11] L’Archéologie du savoir, p. 214; The Archeology of Knowledge, pp. 163-164.
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 and translation and time stamp were completed in January 2022, with additional revision of the translation and transcript in May 2022. Additional revisions were added in February 2024.
The part 2 audio file plays through stereo channel only. This coincides with the (approx.) 34:27 – 1:21:10 marks of the complete audio file.