November 18, 1975

When Félix plays on [the term, black hole], and this is true of our method as a whole, what we're formally saying is: these are not metaphors. And if we say the face, the eyes are black holes, if we say that consciousness, I-equals-I, is a black hole, then the problem is how to get out of this. As I said before: how do we break through the wall, how do we get out of the whirlpool of the hole? ... We will proceed neither by metaphor nor by metonymy. We will proceed by using an inexact term to say the exact thing. Which is: the eyes, the face, consciousness… And we will say: consciousness is a black hole, the eyes are black holes and many other things besides. Memories are a black hole.

Seminar Introduction

Deleuze's 1975-1876 seminars were filmed by one of his students, Marielle Burkhalter, as part of her masters project, "Filming Philosophy as it Happens." Enrico Ghezzi acquired the videos for broadcast on the RAI 3 cinema programme "Fuori Orario," after inviting Burkhalter to screen them at a festival he co-curated. Marielle Burkhalter, along with Stavroula Bellos, would eventually become the director of the L’association Siècle Deleuzien, and oversaw the French transcriptions of Deleuze's seminars at the Voix de Gilles Deleuze website at the University of Paris 8. 

Links to a number of these recordings that are available on YouTube are provided below, each of which includes subtitles in Italian. We are grateful to those who have uploaded these videos.

We have ordered the video material chronologically and divided it into 3 categories, providing short titles describing the topics covered for general orientation. The first (A) refers to the seminars that were filmed at Paris-8 Vincennes in 1975-76, around the time that Deleuze and Guattari began working on what would become “Mille Plateaux”. The second (B) includes seminars from 1980-87, after the campus of Vincennes had been demolished and relocated to St. Denis. The third category (C) features miscellaneous fragments or short edits of the above material.

Deleuze in Seminar
Deleuze in his seminar at the University of Paris VIII, Vincennes, in 1975.

A. Deleuze at Paris 8-Vincennes, 1975-76

Il Senso In Meno 1 Gilles Deleuze & Felix Guattari Vincennes, 1975 1976 Parti 1 2 3 4 (4:00:00)

Il Senso In Meno 2 Gilles Deleuze & Felix Guattari Vincennes, 1975 1976 Parti 6 7 8 9 (4:21:40)

deleuze su molteplicità molare e molteplicità molecolare (1:40:51)

B. Deleuze at Paris 8-St.Denis, 1980-87

The personal pronoun “I” (1980) (video link only partial), transcript and translation located in the second Anti-Oedipus and Other Reflections Seminar, 3 June 1980. [New translation provided of this segment added to the 3 June 1980 lecture]

Deleuze sur Hegel (3:34), Deleuze on Hegel (1980), follows the preceding clip, transcript and translation located in the second Anti-Oedipus and Other Reflections Seminar, 3 June 1980.

On Leibniz (1986) -- Session 3 (18 Nov 1986), transcript and translation located in the Leibniz and the Baroque seminar, 1986-87

On Harmony (1987) -- Session 20 (2 June 1987), transcript and translation located in the Leibniz and the Baroque seminar

C. Various fragments (in random order, clips that are segments from longer videos included above in A)

Gilles Deleuze - Vincennes 1975-76 (compilation) Although this long cllp seems to begin at 1:28:40, the viewer can back up to the start for the full length. However, the segment begins in progress, with students seeming to debate a burning question.  (3:05:57)

Gilles Deleuze, Pierre-Félix Guattari a Vincennes (1975-1976) (22:30)

Félix Guattari - Université de Vincennes 1975 (9:56)

Deleuze et le roman (9:35)

Deleuze sur le langage (1:19)

Deleuze sur la musique (1:04)

Deleuze - Boulez - Berg (5:36)

Gilles Deleuze, Lecture, Mille Plateaux 1 (7:00)

Gilles Deleuze, Lecture, Mille Plateaux 2 (9:26)

Gilles Deleuze - Morale ed etica (Lezioni a Vincennes, 1975/76) Interview with Richard Pinhas (7:03)

Gilles Deleuze à Vincennes, 1 (sub. ITA) (9:47)

Gilles Deleuze à Vincennes, 2 (sub. ITA) (9:51)

Gilles Deleuze à Vincennes, 3 (sub. ITA) (4:36)

Gilles Deleuze à Vincennes, 4 (sub. ITA) (9:27)

Gilles Deleuze à Vincennes, 5 (sub. ITA) (9:56)

Gilles Deleuze à Vincennes, 6 (sub. ITA) (9:50)

Gilles Deleuze à Vincennes, 7 (sub. ITA) (9:47)

Gilles Deleuze à Vincennes, 8 (sub. ITA) (9:49)

Gilles Deleuze à Vincennes, 9 (sub. ITA) (9:58)

Gilles Deleuze à Vincennes, 10 (sub. ITA) (9:48)

Gilles Deleuze à Vincennes, 11 (sub. ITA) (9:52)

Gilles Deleuze à Vincennes, 12 (sub. ITA) (9:35)

Gilles Deleuze à Vincennes, 13 (sub. ITA) (9:52)

Gilles Deleuze à Vincennes, 14 (sub. ITA) (9:54)

Gilles Deleuze à Vincennes, 15 (sub. ITA) (9:56)

Gilles Deleuze à Vincennes 1975 (9) italian sub (51:51)

Gilles Deleuze a Vincennes 1/3 (sub ita) (4:00:00)

Gilles Deleuze a Vincennes 2/3 (sub ita) (4:21:40)


English Translation


In a dual lecture including both Deleuze and Guattari, the authors introduce different terms developed in their ongoing collaboration which will result, in 1980, in A Thousand Plateaus.

Gilles Deleuze

Deleuze at Vincennes, 1975-76

Part 1 -- Surfaces of Redundancy, Black Holes, Language and Orders

Translation: Graeme Thomson and Silvia Maglioni


Félix Guattari: … to get an order of the world to function, systems of orders that order the world - the same type of promotion of invariants that constitute the co-ordinates of a single social, cosmic and affective plane…

Gilles Deleuze: Excuse me… Can everyone hear? Don't you want to sit here? If you speak from the front everyone will be able to hear.

Guattari: So I want to return to the example you proposed of the societies studied by Clastres... [Tape interrupted] [0:48]

Guattari: … the supposed capitalisation of information through distinct units like letters, a highly purified articulation of phonemes with a policed syntax. But there's also a whole series of semiotic components that contribute to something that isn't a transmission of information but an expression of the libidinal life of the group. And this can express itself both through words and through mime, dance, tattoos, rituals and so on. This is what I call the different semiotic components

At the end of the chain we arrive at an order whereby we can transmit something through a series of messages that are conveyed by computer. Today, to designate someone means taking a certain type of data and passing it through a computer. This will tell us not only a person's physical location but also information about their various behaviors, degree of freedom, earning and spending power etc.  I would say that there's been a semiotic collapse in the sense that collective modes of enunciation which inextricably wove and articulated together the various semiotic components have found themselves reduced to the point where they can always be translated in terms of a quantity of information.

The learning of language - the passage from infant to school to professional language and so on - is a process whose aim is to make individuals, however polyvocal their desires, capable of arriving at this possible reduction, which is essential to systems of production and exchange that can permit the circulation only of people who are translatable in terms of information. Otherwise, they must be mad, marginals, poets and what have you. Special tools are used to treat them and they are institutionalized in facilities specifically created to deal with these marginal phenomena.

This process of semiotic collapse makes it possible nowadays to enunciate any element of one's desires or lifeworld, provided that it is compatible with the informatics machine of the system as a whole, whether that be socialist, capitalist, bureaucratic, all the state systems - to borrow Clastres' classification. So, all of a sudden, the only subjectivity possible is that which renders you compatible with the human species in general, the species of state controlled societies and of citizenship. Therefore, you don't have the same rapport with the particularization of enunciations which consisted in saying.... for example, Amerindians… Amerindian tribes say that the only “people” are Amerindians, and when they see white people they say they are not “people”. Their sense of belonging to a community of expression is delimited by the place where there is a mesh of the various semiotic components.

Excuse me for this awful generalization. I know I'm taking two extremely different points in the social field. So what happens? One can submit to this system of computerized reduction only insofar as territorialities are constituted, what I call surfaces of redundancy, where we can articulate this type of opposition. To give an example: choosing a woman in the kind of societies described by Clastres isn't simply a question of trying to have sex, or of reproduction, or of possessing somebody... it's always a matter of the meeting of two social subgroups, implying different systems of exchange, systems of multiple semiotic composition.

Today, we can say that the selection of a sexual partner - which seems to be a free choice - is in fact determined by systems that compel people to correspond to specific socio-economic profiles. To the point that there are some who try to facilitate this through a computer program that matches people's preferences. However, one no longer chooses someone from a certain clan with all that this intricate semiotic dance implies. Perhaps one no longer chooses - and here I'm jumping ahead - a body, the possession of the other's sexual organs - if we want to define it like that - but the possibility of finding a certain type of redundancy: redundancy of survival, redundancy of faciality. We look for someone in the enunciative field that allows us to say something of the order: Tristan-Isolde, Isolde-Tristan. But in a much more sad and sinister way, as in an endless conjugal scene that consists only in the person one is talking to.

When I return, announce my name, and my objective, who is it that resounds the echo of what I'm saying? On which surface of redundancy can I express myself?  In territorialized societies you have large surfaces of redundancy and multiple possibilities of semiotic composition. In the case of conjugal relations, in a deterritorialized, miniaturized subjectivity, one can only latch onto this someone and say: “Do you recognise me when I speak to you?”, “I'll be back at such and such time”, “That's what I'm doing”. But the same system operates with children. Even in rapports of faciality, there's this dissociation, this type of imprinting that ethologists refer to. The rapport of visual contact, the “eye to eye” contact of which Spitz and a number of other Americans speak.

There's always a need to frame things. I exist only as long as there is a certain point that serves as a surface of reference and black hole where I can go on articulating my enunciations. If this last surface of redundancy or reference is taken away from me, the whole system of my informational coordinates literally collapses.

Strangely, we see that this thing psychoanalysts call the partial object - the eyes, the face, the eye-nose-mouth triangle  - constitutes an extraordinary safeguard with respect to different individuals, since not only have the various semiotic components (dance, mime, gesture, group rituals and so on) largely disappeared, but also the possibility of the sexual act itself. We get to the point where one can conceive of falling in love with a trait of faciality in a woman, without having the chance to bring into play the various semiotics of love, sexuality and so on. The essential thing is that there be this last hold, this hook of territorialization. So there you have it.

The second thing we can sketch out for the moment is this: either you have informatics redundancies that latch onto surfaces with black-hole systems such as faciality, state power... In every system of power there's always a system of black holes – the eyes of Giscard d'Estaing, the eyes of the leader... Something like, “In any case he's one of us”, “There's a chance, at least it's France!”, “I recognize myself there because I've already seen him”. I can continue speaking, producing meaning... 

Either you have this system of redundancy-black holes – which is precisely the arborescent system, because wherever there's a black hole there are trees – and, as we said in the case of dreams, you have the dream's umbilicus, a place where everything is organized around a central point, a blind spot. Or we can think of another system of redundancy, though in this case it wouldn't be a redundancy of orders or of subjectivity but a rhizomatic system where there is no black hole or else where the various black holes are bypassed in such a way that we return to a semiotic polyvocality - one no longer subjected to arborescence and the stratification of the various semiotic components, but where every type of expressive element (verbal, bodily, dance, sexual and so on) rather than contributing to the organization of a subject or a couple produces something that I think you (Gilles) are going to unpack.

This would be another type of semiotic organization, one that is no longer informatics-based but which we could roughly call diagrammatic. One that wouldn't be centred around a black hole or pass by way of the mediation of a subject or relationship with another person, but through direct connections between the different semiotic components…  [Tape interrupted]

Deleuze: … since these redundancies of resonance with their faciality, the traits of faciality that compose them, find their outlet, their organization, in this whirlpool system that I called holes... Here, Félix adds a qualification, that we should call them black holes. And I remind you, for those who may not know this... but then you should try to find out more for yourselves, because that's how we can work together... I remind you that black hole is an expression taken from astronomy.

What black hole designates - and Félix will correct me if I'm wrong - is more or less, in the field of relativity… though it depends on which theory of relativity… it’s a star that has passed beneath its critical radius and in so doing, by diminishing this critical radius, surpassing a threshold, has acquired a strange property: it captures what are in this case photons passing nearby and retains them. It captures them and nothing gets out any more. And this is the reason that it doesn't emit... it doesn't emit light, it doesn't emit photons... hence the expression black hole.

When Félix plays on this, and this is true of our method as a whole, what we're formally saying is: these are not metaphors. And if we say the face, the eyes are black holes, if we say that consciousness, I-equals-I, is a black hole, then the problem is how to get out of this. As I said before: how do we break through the wall, how do we get out of the whirlpool of the hole?

It's not a question of saying black hole in a metaphorical sense. It's a question of stealing a word, in this case a word from astronomy, ok we'll take it, we'll pick it up and we'll keep it for our own uses but not as a metaphor. We will proceed neither by metaphor nor by metonym. We will proceed by using an inexact term to say the exact thing. Which is: the eyes, the face, consciousness… And we will say: consciousness is a black hole, the eyes are black holes and many other things besides. Memories are a black hole.

As Félix said, with regard to Proust's madeleine when somebody raised the matter... “What is this business of the madeleine”? Félix put it admirably. He munches on his madeleine. It's a redundancy, a type of redundancy perhaps. He has plunged into a black hole. The black hole of memories. How will he get out? Don't think this is a victory for him. You recognize Proustians by the way they regard the story of the madeleine as a victory. And we recognize the anti-Proustians, those who really love Proust, who suffer for him, saying “Oh no, what a mess he's getting himself into. How is he going to get out?”

Well, he'll get out in his own way, by making a hell of a din. He does what Félix has just said: He makes a line of flight out of the black hole, he bursts, springs out from the black hole exactly the way others manage to break through the white wall, the way Gherasim Luca does. So I think that in the last part of what Guattari said what will be fundamental for us, and he will have to develop it further, is both the topic of faciality and that of black holes… [Tape interrupted]

… In the conception of power that we require everyone, in a certain sense, is a messenger. You're right, everyone's a messenger, there's no front line, it's clear. So it’s language itself that is the messenger of a pre-existing order, which nonetheless doesn't exist outside of language.

Yolande Finkelstein: Regarding the categorization I was thinking of a parallel with what your friend whose name I don't remember… who recites his poem...

Deleuze: Yes, I'm happy to repeat that. I would add that, not by chance, he's called Gherasim… a well-known first name… Luca, L-U-C-A. He's written many poems, a number of wonderful collections published by Soleil noir: Le chant de la carpe (The Song of the Carp) and another whose title escapes me… [Tape interrupted]

Today I want to begin like this. I've quoted some fairly unremarkable examples of redundancy. But we're not even sure that this is redundancy. Units of information that are independent and then probable, faced with which we are obliged to make choices. So at the other pole what is there?

At the other pole you have what computer scientists call noise. Already we're wary and with good reason. We're not doing computer science here. So it's good to be diffident. We know that when computer scientists use the word noise, they're talking about a very particular type of noise. It goes without saying that they oppose to information a noise that is presumed to be non-informative, or that contains the minimum of information, like radio or TV interference. But the noise of a beast that hides in the undergrowth is a noise that is rich in information. Even the noise of interference contains a minimum of information. For example it could be the enemy seeking to disturb a transmission. Even if we assume that it's a random noise burst, it's never completely fortuitous.

We might consider the pole noise as the opposite of the pole maximal information. Redundancy is presented as the diminution of theoretical information that is supposedly a priori by right. But something else appears at the same time. Redundancy is the only way of fighting against noise. Which is to say it's the only way to save the information from crumbling, from disintegrating into noise. Whether it be at the level of letters or that of phonemes makes no difference.

A language contains a more or less large frequency of any given letter or a certain phoneme that it uses. For example the frequency of a given letter or phoneme are not the same in French as they are in English. So we can already begin a table of the comparative frequencies of letters or phonemes in a certain number of languages. We'll call it a zero-order estimation. After which, the following estimation is already a redundancy: the frequency of a letter or phoneme in a given language.

And then there is another estimation that we can call the first-order estimation. This time we study the frequency of a letter in a given language with respect to the preceding or subsequent letter. In this case too, depending on the language, the frequencies are different. I'll give you a random example. In French what is the frequency of... we could also imagine a computer performing this kind of research... What is the frequency of the group B-A? That is an A preceded by a B. We could do similar studies in the case of phonemes.

For the second-order estimation I can look for frequencies concerning three groups of letters. This series of estimations will therefore define a certain type of redundancy… [Tape interrupted]

The diminishing of absolute theoretical information and the struggle against noise… why are these two functions linked? Nothing would prevent absolute theoretical information from falling into pure noise if there wasn't this regulating force of redundancy, which assures the struggle against noise by diminishing the level of absolute theoretical information. So we have a first schema, a schema of redundancy, where at the top we will put “maximal theoretical information” and below “noise” and between the two “redundancy”.

And Martinet concludes: “The presence of redundancy is a way to permit the transmission of signs”, and this happens through the series of estimations of which I just spoke. Then Martinet identifies a second type of redundancy. He says: “Not only is redundancy a way to permit the transmission of signs but there is nothing to prevent it itself from becoming a sign”. There… you see how it constitutes a sign in its own right. Redundancy not as regulator of the transmission of signs - a regulatory process in the transmission of signs, or code for transmitting information - but it itself as a sign. Why is this?

He says: “It must be so that the user can seek out the redundancy for itself.” So users can look for the actual redundancy. “A means of expression for the individual, of manifestation of group consensus. We will therefore identify a principal function of combating noise, without which any communication would be impossible”. And this is our first case of redundancy.

And then there are secondary uses, which are left at the subject's disposal. Means of expression, of action upon others, of enchantment. Reducing the second type of resonance to a secondary use of the first type may be correct from the perspective of informatics, but we're not sure that what we are dealing with has anything to do with informatics. And perhaps we can even say that it isn't correct from any perspective, particularly since it actually forms part of very different systems of signs: the subjective redundancies of resonance and the signifying redundancies of frequency. It's by no means sure that they form part of the same system of signs. It's by no means sure that the latter are simply a secondary use of the former.

And there's another thing that disturbs me and that might disturb us all, but this time it's not related to the second type of redundancy but to what popular informatics tell us about the first schema. For the moment I'm only going to speak vaguely about this… because I'd like to focus on it later. So you can clearly see that what I want to say about redundancies is that we still aren't even sure what all these redundancies are. So far so good… [Tape interrupted]

There's a lot that can be said about the informative nature of language. A number of linguists have already spoken extensively about it. It's interesting because the idea that language is by its very nature informative is one that corrupts us to such an extent that... I think of a case like that of Sartre, who at a certain point felt the need... and I don't think he would say this now... to identify what it was that characterised language, or to be precise, poetry or literature. And he said that literature and poetry begin when there is information. Barthes too once said something similar.

It's very odd to make this kind of affirmation. What is it that has completely corrupted, compromised, putrefied even, the question of language-power relations? It's the bad choice we've been left with, which is to say the very conception of power that has been proposed to us.

When we were told “language is either infrastructure or ideology, we were already pretty irritated, we were irritated for language's sake, we were really irritated for language to be stuck with this alternative, even if it was presented in the most sophisticated way imaginable. Although sometimes it wasn't presented in a very sophisticated way, yet in the end the more sophisticated the presentation the worse it turned out. Because we realised that this question of language was extremely complicated. It's not infrastructure. No, it can't be infrastructure. Language doesn't produce anything. It produces only words. It doesn't produce goods. No aspect of infrastructure coincides with language. So we wondered if it was rather a superstructure. In other words is it the state apparatus that decides on language? Difficult to say.

As Stalin said: “No, we've changed everything but not Russian, not much at least. Of course we've perfected it. But within certain limits”. So it's not the state apparatus. We don't change language the way we change a constitution or a police force. So is it ideology? “No”, he said. “It may be the vehicle of ideologies, but it can equally well be the vehicle of other things besides ideology.” So even ideology isn't a strong candidate. So what is language?

We always have to go back to this text because it's both short and rather wonderful, the text Stalin wrote on linguistics, where he says: “Comrades, you're wrong. There are those among you who say that language is infrastructure and they are mistaken, they are not looking at the question rationally. Others say that language is ideology and there is a language of the people, a proletarian language, a bourgeois language. I say that this not the case and that you don't really see what is at issue, comrades.”

And Stalin goes so far as to say that language is the common good of a nation, and that it ensures the communication of information. It suited him to say that of course, since it implied a conception of power related to our well-known themes of infrastructure, superstructure, ideology and so on. If we were to say instead that language has always been a system of order and not of information... that it is orders we are given, not information that is communicated, it would seem to us that we were saying something obvious.

We turn on the TV news and what do we get? In the first place, we don't receive information, we receive orders. And at school, what goes on? Here too it's obvious. At school the children don't receive information... The example of school is... Félix puts it well in a text he wrote... we put language in the mouths of children exactly the way we put shovels and picks in the hands of workers. Which doesn't mean that language is an infrastructure but that it relates to the field of orders.

When the teacher gathers the children together it's not to inform them about the alphabet, it's to teach them a system of orders. And here, to settle our accounts with the Chomskians - which we'll have to do - we have to add that their famous phrase markers are above all markers of power. And that a syntax is a system of orders, a system of command that will allow or force individuals to form enunciations that conform to dominant enunciations. And the function of school is primarily this.

So language must be conceived first of all not in terms of information but in terms of order. Not in terms of communicating information but of transmitting orders. Clearly, for us, this implies - and I would say this is simple and obvious - that we have to look for another conception of power, because, after all, Stalin is quite right in his remarks. There's no one who decides syntax. Which means that power is undoubtedly something completely different to the properties of individuals or groups in a given moment.

For the time being let's just say that language is a formalization of expression. Not all expressions pertain to language. Language is a particular formalization of expression whose function is to transmit orders in a society. We know that this implies giving power another conception different from the Marxist conception. In this sense, language, including syntax, is… we can't even call it an instrument… it's an element and component of power. So in one sense it's not informative. And yet, in a certain way it is. Which is to say that it provides the minimal information and guides the minimum choice necessary for the correct understanding of relative, limited information, relative to the orders that are given.

It goes without saying that when someone on the street shouts “There's a fire!” (Au feu!), it's better if the kids don't understand it as “Go and play!” (Au jeu!). So there is of course information and there are choices and approximations, but which are nonetheless relative to the orders communicated by language.   

There's a letter… you know how Lewis Carroll used to write letters to little girls, never to little boys. And there’s a famous letter to a little girl he wrote - which is well known and well translated so we don't need to read it English which would be difficult for some, myself included. This admirable translation by Jacques Papy that works by equivalences, you can find in various editions of Carroll's Letters. It describes a situation that corresponds perfectly to what we've been discussing.

During one of his lessons in a high society context, the teacher is at the bottom of the garden. And there's a first servant who repeats his questions. Everyone knows that a teacher's questions are really orders. When a teacher asks a panicked child what 2 and 2 are, it's clear he's not asking for information, he's giving an order. And you will tell me, “But the child would have to be informed first”. Indeed the child has been informed by a previous system of orders. It's always the information that presupposes the order and not the other way round. I don't mean order in the sense of organization or ruling but in the sense of a command.

So the first servant repeats the teacher's question, then there is a second servant who repeats the question repeated by the first, and then a third and so on. And to mark the hierarchy that runs through language, the pupil is at the other end of the garden. And then he sends back his answer. And in Carroll's letter, which I deliberately didn't bring with me... you can look for it yourselves if you're interested… the question is passed down the line, each time completely transformed, because the first servant didn't hear properly. It begins with “what are 2 and 2”, the first servant deforms it and the question changes, the second deforms it further and the third even more. So what the pupil hears is a completely different question to which he anxiously responds, and the answer goes back up the chain.

Here we have a whole system in which the choices are always wrong but are nonetheless determined by a chain of order and command. And so the obedient response is sent back up and is more and more wrong. The conditions of information are conditions of reception of orders and commands… [Tape interrupted]

The schema has three heads -- Thank you for the chalk -- The schema has three heads. Maximal theoretical information... noise which completely disturbs the emission and reception of information... redundancy... there... fighting against the noise, letting us defeat the noise which diminishes at the cost of also diminishing the maximal theoretical information.

The two go well together, since without redundancy the maximal theoretical information would itself be noise. We are trying, in a crafty, underhand way, to replace this with another schema. The orders-commands schema… You see immediately where I'm heading. I still don't know exactly how but I feel... no, we feel... we all feel, that orders-commands - I'm not saying that these are different –that orders-commands contain, comprise and in fact result to be the same thing as redundancies.

An order doesn't need to be repeated. If we repeat an order, it's because the order is already redundant. So one type of redundancy is the form of the order as such, the form of command. And generally speaking, it shouldn’t surprise us if this is the case. We shouldn’t be surprised when an order is repeated, if this is the way an order is given and received, for no other reason than to show that I have understood well. Like when someone says: “Come on, go and do such and such!” and I reply: “Yes, I'll go and do such and such!”

In this case I am superfluous. I redund, so to speak. But what do I redund? Redundancy… it's the order itself. “Go and do such and such!” or “Go and play!” (Au jeu!) no… “There's a fire!” (Au feu!) no… “Open fire!” (En joue le feu!)

“Open fire!” The order has to pass down the line. So the general says “Open fire!” or “Weapons at the ready!”, or something similar, and the captain says “Weapons at the ready!” and the sergeant says “Weapons at the ready!” until the order arrives at the poor guys who have to prepare their weapons. But if the order is repeated, it's because, in itself, it is redundant. The order, the command are forms of redundancy in themselves. Maybe…

So we were saying that language is not information but order or command. And we also said that the pure order, the order-command in its pure state, is pure redundancy, absolute redundancy. And this is what we put at the top of our schema. Redundancy is the same thing as an order. So to repeat: if order becomes redundancy, if it is repeated, it is because it is in itself redundancy. So repetition as a practice with respect to the order would simply be a consequence of redundancy being identical in nature to order. It's not certain, but that's the sensation we have... there's nothing we can do about it.

Here, between the two... the two what? I don't know. Between the two... it's just to make a well-balanced schema. We'll see. Between the two we'll put “information”, which is always relative. If it's true to say that redundancy is the absolute form of order, we will say that relative information is the limited content of an order in as far as it differs from another order. ABC is not the same thing as BCD. Ok, so we have relative information…

You see that the form of order is absolute redundancy. So what follows is relative information. If the order didn't communicate a piece of relative information, “Do this rather than that”, we would be in the same situation as Lewis Carroll's teacher and pupil where we have the order that is passed down the line - an execution of the order - and an act of obedience which is completely different. The information is merely the relative condition whereby the execution of the order corresponds to the order itself. It's like an inversion of the informatics schema. Actually, it's even worse. It's a completely different field. So anyway. Here in this third position what can I put?

It's clear, and it’s also clear that this is how it works. It's not noise that is an informatics abstraction. It's silence. Silence. What does that mean? Why introduce silence here? Silence.

Silence is ambiguous. Because it could be the state of the person who obeys - but that is a silence of language, a silence of language itself that is included in language itself… it's what happens between the reception of the order and the response to the order. The captain says: “Load your rifle!” and what follows is a silence filled with the sound of a rifle's bolt action. And then the soldier says: “Ready, captain”. But there's another silence, a silence that consists in something quite bizarre: an escape from all this. We know this silence. When the captain says: “Ready, take aim, fire!” And there’s nothing. Silence… [Tape interrupted]

There's no great composer who doesn't have their techniques of silence. Here we can say something that everyone who listens to music will already be familiar with. Which is that music is traversed by a sort of vector of abolition. A vector of sonic abolition. As though it was completely intrinsic to music. The will and the movement of turning off, and turning us off with it. Like a kind of tracing of abolition. And sounds - which are not a language, even if in music too there are orders but music flows underneath these orders that are properly musical - trace a line of abolition that is fully part of the music… and they do so in many ways.

This first system of signifying redundancy, or redundancy of frequency, inscribes itself on a semiotic wall. And the attempt to pierce this wall... we'll call this attempt exiting from the signifier, from signifiance. Just as Gherasim Luca does.

And now I'll make a brief mention of the second type of redundancy. Obviously, once again, these are not secondary uses of the first type. This time we're dealing with redundancies of resonance.

Of course, we already know something in advance: that everything mixes, everything is mixed. There are no frequencies that don't have resonance and vice versa. So it's not a duality, even if it would be easier to proceed as if that were the case. Though if you tell me that it's not a duality, I'll say, no it isn't. But at the same time, yes it is. And if you tell me it is, I'll say no it isn't. 

Let's try to work things out from there… Redundancy of resonance: what is it? As we saw, it's not a secondary use. To use the language of linguists we would say that it's neither a phoneme nor a morpheme, nor is it a word. So what are we dealing with? Not surprisingly, the last two examples of redundancy that I quoted refer to something linguists call... something we've talked about at length… something they call “shifter”. Which is the redundancy – I-equals-I – of the personal pronoun. And the Tristan-Isolde / Isolde-Tristan redundancy of the proper name.

As you know, according to linguists, shifters are terms that designate those who are literally their bearers. The “I” designates the person who says “I”. This is not the case for other words, it's not the case for common names. The dog in this case is not the word “dog”. But the word “I” designates the one who enunciates “I”, and we speak of this as a  “shifter.”

The proper name designates its bearer. Therefore, the proper name - like the personal pronoun, the “I” - are very special things that animate the second of the categories of resonance. And we've seen that our two examples of categories of resonance are probably closely connected - just like our two examples of categories of frequency. We have this consciousness that says “I-equals-I”, or we have this couple who say “I love you - I love you” or “I hate you - I hate you” or “I love you - I hate you” or “I hate you - I love you”. A redundancy. Or “Tristan - Isolde”.

Is this schema the same? Once again, what we had in the first schema was: redundancy of frequency, signifier inscribed on a semiotic wall, and the problem was how to pierce this wall to arrive... at something that might no longer even be semiotics. In any case, it will be an unformulated semiotics.

In the other case – here I take up a hypothesis that Guattari is currently developing… is it the same case? Perhaps, but for our purposes it's better to make distinctions. This time the schema seems to me a little different. -- Oops, I've lost the chalk… Thank you! --

Tristan - Isolde. Or else we have “I-equals-I”… I would say that this time, in the redundancy of resonance, the resonance is assured by the elements in question – I and I… or in more technical terms “the subject of enunciation” and “the subject of the enunciated”. Tristan and Isolde, successively “subject of enunciation” and “subject of the enunciated” in their respective mouths. “I Tristan enunciate you Isolde”, but also “I Isolde enunciate you Tristan”.

The subject of enunciation and the subject of the enunciated in the two cases of redundancy of resonance were as if attracted in a kind of vortex, a real black hole. They started to whirl around and catch fire in this hole. I, I, I... Tristan - Isolde, Tristan - Isolde, Tristan - Isolde, and the boat sails on... Towards what? Towards death.

The cogito doesn't have a very bright future ahead of it. In appearance, it's a formally different system. I had, and here I would like to end… redundancy of frequency, signifier, operating on a wall, a semiotic wall, with the crucial question of how to knock down the wall. And here I have redundancy of resonance or of subjectivity that refers to a hole. [End of recording


French Transcript


In a dual lecture including both Deleuze and Guattari, the authors introduce different terms developed in their ongoing collaboration which will result, in 1980, in A Thousand Plateaus.

En cours de developpement


For archival purposes, the translation was completed in April 2020.