interview with Gilles Deleuze
Conversation of September 13, 1983, with Pascal Bonitzer and Jean Narboni, asset down and amplified by the participants
Cahiers du Cinema 352 (October 1983)
Your book's presented, not as a history of cinema, but as a classification of images and signs, a taxonomy. In this respect, it follows on from some of your earlier works: for instance, you made a classification of signs when writing about Proust. But with The Movement-Image you've decided for the first time to tackle, not a philosophical problem or a particular body of work (that of Spinoza, Kafka, Bacon, or Proust, say),but the whole of a particular field, in this case cinema. And also, although you rule out producing a history of cinema, you deal with it historically.
Well yes, in a way it's a history of cinema, but a "natural history." It aims to classify types of images and the corresponding signs, western, whole, as one classifies animals. The main genres, western, crime, period films, comedy, and soon, tell us nothing about different types of images or their intrinsic characteristics. The different sorts of shot, on the other hand-dose-up, long shot, and soon-do amount to different types of image, but there are lots of other factors, lighting, sound, time, which come in too. If I consider the field of nine whole, images, it's because it's all built upon the movement-image. That's how it's able to reveal or create a maximum of dpi images, and aboveall to combinethemwithone another through montage. I There are perception-images, action-images, affection-images, along with many other types. And in each case there are internal signs that characteristic extremely rich classification of signs, relatively independent linguistic model. It was particularly tempting to see whether the most matter introduced by cinema was going to require a new understanding of images and signs. In this sense, I've tried to produce book on logic, a logic of cinema.
Isn't this just what you get in a filmmaker like Dreyer, who inspires some very fine passages in your book? I recently saw Gertrud again, which is going to be re-released after twenty years. It's a wonderful film, where the modulation between different levels of time reaches a subtlety only, sometimes, equalled in Mizoguchi's films (with the appearance and disappearance of the potter's wife, dead and alive, at the end of Ugetsu Monogatari, for instance). And Dreyer, in his essays, is constantly saying we should get rid of the third dimension, depth, and produce flat images, setting them in direct relation to a fourth and fifth dimension, to Time and spirit. When he discusses The Word, for example, whats so intriguing is his explanation that it's not a story about ghosts or madness, it's about a "profound relation between exact science and intuitive religion. " And he invokes Einstein. I quote: " Recent science, following upon Einstein's relativity has brought proofs of the existence outside the world of three dimensions which is that of our senses of a fourth dimension, that of time, and a fifth, the psychical. It has been shown that it is possible to experience events which have not yet taken place. New perspectives have been opened up which make us recognise a profound relation between exact science and intuitive religion. " . . . But let's return to the question of "the history of cinema. "You introduce an order of succession, you say a certain type of image appears at a certain moment, for instance after the war. So you're not just producing an abstract classification or even a natural history. You want to account for a historical development too.
In the first place, the various types of image don't already exist, they have to be created. A flat image or, conversely, depth of field, always has to be created or re-created-signs, if you like, always imply a signature. So an analysis of images and signs has to include monographs on major auteurs. To take an example: I think expressionism conceives light in relation to darkness, and their relation is one of struggle. In the prewar French school it's quite different: there's no struggle, but alternation; not only is light itself motion, but there are two alternating lights, solar and lunar. It's very similar to the painter Delaunay. It's anti-expressionism. If an auteur like Rivette belongs these days to the French school, it's because he's rediscovered and completely reworked this theme of two kinds of light. He's done wonders with it. He's not only like Delaunay, but like Nerval in literature. He's the most Nervalian, the only Nervalian, filmmaker. There are of course historical and geographical factors in all this, running through cinema, bringing it into relation with other arts, subjecting it to influences and allowing it to exert them. There's a whole history. But this history of images doesn't seem to me to be developmental. I think all images combine the same elements, the same signs, differently. But not just any combination's possible at just any moment: a particular element can only be developed given certain conditions, without which it will remain atrophied, or secondary. So there are different levels of development, each of them perfectly coherent, rather than lines of descent or filiation. That's why one should talk of natural history rather than historical history.
Still, your classification's an evaluation. It implies value judgments about the auteurs you deal with, and so about those you hardly notice, or don't mention. The book does, to be sure, point toward a sequel, leaving us on the threshold of a time-image that goes beyond the movement-image. But in this first volume you describe the breakdown of the action-image at the end of, and just after, the Second World War (Italian neorealism, then the French New Wave. . . ). Aren't some of the features by which you characterize the cinema of this crisis (a taking into account of reality as fragmentary and dispersive, a feeling that everythings become a cliche, constant permutations of what s central and peripheral, new articulations of sequences, a breakdown of the simple link between a given situation and a characters action) . . . isn't all that already there in two prewar films, The Rules of the Game and Citizen Kane, generally considered to be founding works of modern cinema, which you don't mention?
I don't, first of all, claim to have discovered anyone, and all the auteurs I cite are well-known people I really admire. For example, on the monographic side, I consider Losey's world: I try to define it as a great sheer cliff dotted with huge birds, helicopters, and disturbing sculptures, towering over a little Victorian city at its foot. It's Losey's own way of recreating the naturalist framework. A framework of which you get different versions in Stroheim, in Bunuel. I take someone's work as a whole, I don't think there's anything bad in a great body of work: in Losey's case The Trout was disparaged, even by Cahiers, because people didn't take enough account of its place in his work as a whole: it's a reworking of Eva. Then you say there are gaps, Welles, Renoir, tremendously important auteurs. That's because I can't in this volume deal with their work as a whole. Renoir's work seems to me dominated by a certain relation between theater and life or, more precisely, between actual and virtual images. I think Welles was the first to construct a direct Time-image, a Time-image that's no longer just derived from movement. It's an amazing advance, later taken up by Resnais. But I couldn't discuss these things in the first volume, whereas I could discuss Naturalism as a whole. Even with neorealism and the New Wave, I only touch on their most superficial aspects, right at the very end.
One gets the impression, all the same, that what really interests you is naturalism and spiritualism (say Bunuel, Stroheim, and Losey on the one hand, Bresson and Dreyer on the other), that is, naturalism's descent and degradation, and the elan, the ascent of spirit, the fourth dimension. They're vertical motions. You don't seem so interested in horizontal motion, in the linking of actions, in American cinema for example. And when you come to neorealism and the New Wave, you talk sometimes about the action-image breaking down, and sometimes about the movement-image in general breaking down. Are you saying that at that point it's the movement-image as a whole that begins to break down, producing a situation where another type of image that goes beyond movement can appear; or just the action-image, leaving in place, or even reinforcing, the other two aspects of the movement-image: pure perceptions and affections?
It's not enough just to say that modern cinema breaks with narrative. That's only an effect whose cause lies elsewhere. The cinema of action depicts sensory-motor situations: there are characters, in a certain situation, who act, perhaps very violently, according to how they perceive the situation. Actions are linked to perceptions and perceptions develop into actions. Now, suppose a character finds himself in a situation, however ordinary or extraordinary, that's beyond any possible action, or to which he can't react. It's too powerful, or too painful, too beautiful. The sensory-motor link's broken. He's no longer in a sensory-motor situation, but in a purely optical and aural situation. There's a new type of image. Take the foreign woman in Rosselini's Stromboli: she goes through the tuna-fishing, the tuna's agony, then the volcano's eruption. She doesn't know how to react, can't respond, it's too intense: "I've had it, I'm afraid, it's so strange, so beautiful, God . . . " Or the posh lady, seeing the factory in Europa 51 : 'They looked like convicts. . . " That, I think, is neorealism's great innovation: we no longer have much faith in being able to act upon situations or react to situations, but it doesn't make us at all passive, it allows us to catch or reveal something intolerable, unbearable, even i in the most everyday things. It's a Visionary cinema. As Robbe-Grillet , says, descriptions replace objects. Now, when we find ourselves in these purely optical and aural situations, not only does action and thus narrative break down, but the nature of perceptions and affecions changes, because they enter a completely different system, from the sensory-motor system of "classic" cinema. What's more, we're no longer in the same type of space: space, having lost its motor connections, becomes a disconnected or vacant space. Modern cinema constructs extraordinary spaces; sensory-motor signs have given way to "opsigns" and "sonsigns." There's still movement, of course. But the movement-image as a whole comes into question. And here again, obviously, the new optical and aural image involves external factors resulting from the war, if only half-demolished or derelict spaces, all the forms of "wandering" that take the place of action, and the rise, everywhere, of what is intolerable.
An image never stands alone. The key thing's the relation between images. So when perception becomes purely optical and aural, with what does it come into relation, if not with action? An actual image, cut off from its motor development, comes into relation with a virtual image, a mental or mirror image. I saw the factory, and they looked like convicts. . . Instead of a linear development, we get a circuit in which the two images are constantly chasing one another round a point where real and imaginary become indistinguishable. The actual image and its virtual image crystallize, so to speak. It's a crystal image, always double or duplicated, which we find already in Renoir, but in Ophuls too, and which reappears in a different form in Fellini. There are many ways images can crystallize, and many crystalline signs. But you always see something in the crystal. In the first place, you see Time, layers of time, a direct time-image. Not that movement's ceased, but the relation between movement and time's been inverted. Time no longer derives from the combination of movement-images (from montage), it's the other way round, movement now follows from time. Montage doesn't necessarily vanish, but it plays a different role, becomes what Lapoujade calls "montrage." Second, the image bears a new relation to its optical and aural elements: you might say that in its visionary aspect it becomes more "legible" than visible. So a whole pedagogy of the image, like Godard's, becomes possible. Finally, image becomes thought, is able to catch the mechanisms of thought, while the camera takes on various functions strictly comparable to propositional functions. It's in these three respects, I think, that we get be ond the movement-image. One might talk, in a classification, of "chronosigns", "lectosigns", and "noosigns."
You're very critical of linguistics, and of theories of cinema inspired by that discipline. Yet you talk of images becoming "legible" rather than "visible." Now, the term legible as applied to cinema was all the rage when linguistics dominated film theory ("reading a film, " "readings" of films. . .). Isn't there a risk of confusion in your use of this word? Does your term legible image convey something different from that linguistic conception, or does it bring you back to it?
No, I think not. It's catastrophic to try and apply linguistics to cinema. Of course, thinkers like Metz, or Pasolini, have done very important critical work. But their application of a linguistic model always ends up showing that cinema is something different, and that if it's a language, it's an analogical one, a language of modulation. This might lead one to think that applying a linguistic model is a detour that's better avoided. Among Bazin's finest pieces there's one where he explains that photography's a mold, a molding (you might saythat, in a different way,language too is a mold), whereas cinema is modulation through and through. Not just the voices but sounds, lights, and movements are being constantly modulated. These parameters of the image are subjected to variations, repetitions, alternations, recycling, and so on. Any recent advances relative to what we call classic cinema, which already went so far in this direction, have two aspects, evident in electronic images: an increasing number of parameters, and the generation of divergent series, where the classic image tended toward convergent series. This corresponds to a transition from visibility to legibility. The legibility of images relates to the independence of their parameters and the divergence of series. There's another aspect, too, which takes us back to an earlier remark. It's the question of verticality. Our visual world's determined in part by our vertical posture. An American critic, Leo Steinberg, explained that modern painting is defined less by a flat purely visual space than by ceasing to privilege the vertical: it's as though the window's replaced as a model by an opaque horizontal or tilting plane on which elements are inscribed. That's the sense of legibility, which doesn't imply a language but something like a diagram. As Beckett says, it's better to be sitting than standing, and better to be lying down than sitting. Modern ballet brings this out really well: sometimes the most dynamic movements take place on the ground, while upright the dancers stick to each other and give the impression they'd collapse if they moved apart. Maybe in cinema the screen retains only a purely nominal verticality and functions like a horizontal or tilting plane. Michael Snow has seriously questioned the dominance of verticality and has even constructed special equipment to explore the question. Cinema's great auteurs work like Varese in music: they have to work with what they've got, but they call forth new equipment, new instruments. These instruments produce nothing in the hands of second-rate auteurs, providing only a substitute for ideas. It's the ideas of great auteurs, rather, that call them forth. That's why I don't think cinema will die, and be replaced by TV or video. Great auteurs can adapt any new resource.
Verticality may well beone of the great questions of modern cinema: it's at the heart of Glauber Rocha's latest film, The Age of the Earth, for example - a marvelous film containing unbelievable shots that really defy verticality. And yet, by considering cinema only from this "geometric, "spatial angle, aren't you missing an essentially dramatic dimension, which comes out for example in the problem of the l00k as handled by auteurs like Hitchcock and Lang? You do, in relation to Hitchcock, talk about a "demarque," which seems implicitly to relate to the look. But the notion of the look, the very word itself, doesn't once appear in your book. Is this deliberate?
I'm not sure the notion's absolutely necessary. The eye's already there in things, it's part of the image, the image's visibility. Bergson shows how an image itself is luminous or visible, and needs only a "dark screen" to stop it tumbling around with other images, to stop its light diffusing, spreading in all directions, to reflect and refract the light. 'The light which, if it kept on spreading, would never be seen." The eye isn't the camera, it's the screen. As for the camera, with all its ropositional functions, it's a sort of third eye, the mind's eye. You cite Hitchcock: he does, it's true, bring the viewer into the film, as Truffaut and Douchet have shown. But that's nothing to do with the look. It's rather because he frames the action in a whole network of relations. Say the action's a crime. Then these relations are another dimension that allows the criminal to "give" his crime to someone else, to transfer or pass it on to someone else. Rohmer and Chabrol saw this really well. The relations aren't actions but symbolic acts that have a purely mental existence (gift, exchange, and so on). And they're what the camera reveals: framing and camera movement display mental relations. If Hitchcock's so English, it's because what interests him is the problem and the paradoxes of relation. The frame for him is like a tapestry frame: it holds within it the network of relations, while the action is just a thread moving in and out of the network. What Hitchcock thus brings into cinema is, then, the mental image. It's not a matter of the look, and if the camera's an eye, it's the mind's eye. So Hitchcock has a special place in cinema: he goes beyond the action-image to something deeper, mental relations, a kind of vision. Only, instead of seeing this as a breaking-down of the action image, and of the movement-image in general, he makes it a consummation, saturation, of that image. So you might equally well say he's the last of the classic directors, or the first of the moderns.
You see Hitchcock as the prototypical filmmaker of relations, of what you call thirdness. Relations: is that what you mean by the whole ? It's a difficult bit of your book. You invoke Bergson, saying the whole isn't closed, it's rather the Open, something that's always open. It's particular sets of things that are closed, and one mustn't confuse the two . . .
The Open is familiar as a key notion in Rilke's poetry. But it's a notion in Bergson's philosophy too. The key thing is to distinguish between particular sets of things and the whole. Once you confuse them, the whole makes no sense and you fall into the famous paradox of the set of all sets. A set of things may contain very diverse elements, but it's nonetheless closed, relatively closed or artificially limited. I say "artificially" because there's always some thread, however tenuous, linking the set to another larger set, to infinity. But the whole is of a different nature, it relates to time: it ranges over all sets of things, and it's precisely what stops them completely fulfilling their own tendency to become completely closed. Bergson's always saying that Time is the Open, is what changes-is constantly changing in nature-each moment. It's the whole, which isn't any set of things but the ceaseless passage from one set to another, the transformation of one set of things into another. It's very difficult to think about, this relation between time, the whole, and openness. But it's precisely cinema that makes it easier for us to do this. There are, as it were, three coexisting levels in cinematography: framing, which defines a provisional artificially limited set of things; cutting, which defines the distribution of movement or movements among the elements of the set; and then this movement reflects a change or variation in the whole, which is the realm of montage. The whole ranges over all sets and is precisely what stops them becoming "wholly" closed. By talking about off screen space, we're saying on the one hand that any given set of things is part of another larger two- or three-dimensional set, but we're also saying that all sets are embedded in a whole that's different in nature, a fourth or fifth dimension, constantly changing across all the sets (however large) over which it ranges. In the first case we have spatial and material extension, but in the other, the spiritual order we find in Dreyer or Bresson. The two aspects aren't mutually exclusive but complementary, mutually supportive, and sometimes one's dominant, sometimes the other. Cinema's always played upon these coexisting levels, each great auteur has his own way of conceiving and using them. In a great film, as in any work of art, there's always something open. And it always turns out to be time, the whole, as these appear in every different film in very different ways.
Deleuze, Gilles, Negotiations, 1972-1990 / part two CINEMAS/Gilles Deleuze : translated by Martin Joughin.
Columbia University Press New York Chichester, West Sussex
Pourparlers @ 1990 by Les Editions de Minuit Translation copyright@ 1995 Columbia University Press