April 24, 1984
It’s the movement-image which derives from the time-image and not the reverse, and that is why movement is only presented in the form of an aberration of movement, and that is why that this is the reign of the jump cut (faux-raccord). And you will tell me, but there were also jump cuts before. Obviously there have always been jump cuts with the cinema, only everything changes; it is a question of evaluation, they knew it. Once again, Epstein said the deepest things about aberrations of movement in the movement-image itself. But what is important is that as far as these authors went in the aberrations of movement, aberrations of movement were related to movement-images; they were accidents of movement-images. As a result, the time-image remained the indirect image of time concluded from movement-images, taking account of aberrations. What changed? What is it? It is that between the beginning of the war and the end of the war, in the period of the Second World War, the cinematographic image indeed undergoes a mutation, namely the aberration of movement is no longer related to movement . The aberration of movement is valid for itself; the only movement that exists is aberrant, and henceforth, the aberrant movement depends on the time-image.
Seminar Introduction
In contrast to his rather apologetic return at the start of year 2 to the Cinema material discussed in year 1, Deleuze commences year 3 with a forthright proposal to discuss the intersection of cinema with the theme of truth, time and the falsifier. Adopting this topic, that constitutes the focus of chapter 6 in The Time-Image, means that Deleuze intends to situate these thematics within the broader framework of the concepts introduced in years 1 & 2 as well as those that inform his development in The Time-Image.
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

Gene Kelly in Stanley Donen’s “Singin’ in the Rain”
Stating his annoyance at losing the two following Tuesdays (1 & 8 May) to national holidays, Deleuze summarizes the year’s work to date, with the session’s third segment to be a double inquiry: to see how a direct image of time or time-image is constructed, as well as “time” as a concept, and to understand the relations between this concept as philosophical and this image as aesthetic. Taking issue with Robbe-Grillet, Deleuze asserts that cinema’s time is instead an indeterminate temporality, and that the more that aberrations of movement gain independence, so too the time-image ceases depending on movement. To demonstrate that montage becomes the operation through which time relations are determined in a direct time-image, Deleuze first points to Welles, Fellini, and Dreyer and considers the problem of depth of field as a cinema technique (cf. Bazin and Jean Mitry; the history of painting; and then diverse filmmakers). Referring to “Citizen Kane”, Deleuze considers depth of field as a function of temporalization, inserting the viewer into a direct time-image (cf. Visconti, Robbe-Grillet and Resnais’ “Marienbad”). Deleuze then shifts to the “flat image” that causes an experience of time in a pure state (cf. Dreyer and Ozu), i.e., cases of an indeterminate direct time-image. This leads Deleuze to the emergence of pure optical and sound situations, a “clairvoyant cinema” (cinéma de voyant), also a “cinema of strollers” in any-spaces-whatever (cinéma de balade). Deleuze proposes decors in American musical comedy as developing their own value, hence dancers emerging within these pure optical and sound situations, i.e., the rhythmic relations of time-image (cf. Donen). Deleuze closes on the importance of non-localizable links evident in dance cinema, between pure optical and sound situations and world movements, i.e., direct time-images or rhythms creating circuits across different planes of actual images linked to virtual images, to be discussed at the next session.
Gilles Deleuze
Seminar on Cinema, Truth, and Time: The Falsifier, 1983-1984
Lecture 17, 24 April 1984 (Cinema Course 61)
Transcription: La voix de Deleuze, Lucie Picandet (Part 1), Adrien Pequignot (Part 2) and Pablo Gomez (Part 3); additional revisions to the transcription and time stamp, Charles J. Stivale
Translation: Graeme Thomson & Silvia Maglioni
Part 1
We’re in the dark… here… so there’s nothing to distract us anymore! But then there is something both pleasant and unfortunate… something that’s never happened to me before… Mayday falls on a Tuesday. This has never happened before. So, I was vaguely expecting it, but now that it’s happening, it’s too late. I’m not even happy about it anymore. But then, what bothers me frankly is that I think, I think… Does anyone know… the 8th is also a holiday, right? So, I’m thinking, oh no, this is too much, this is really too much! Therefore, we’ll have to… So, that’s going to be 15 days that we’ll have to skip, two whole sessions. I’ll have to extend my contract until June, even if I’ll be on my own, right? I’ll be coming here. So, we won’t be meeting again until… which bothers me, because today, I don’t know what to do because I need some kind of continuity, but let’s see… So, our next session will take place on…
Student: May 15.
Deleuze: May 15? Yes. So, May 1st and May 8th… is it true that the 8th is a holiday?
Student: Yes, yes, yes.
Deleuze: Does it say that? Does it say Victory? VE Day…[1]
So, last time, we had finished with a large group of concepts… If we have to summarize the major themes of what we did, roughly speaking, we can say that during the first term we covered the force of time-powers of the false and the crisis of truth, while during the second term we focused on the philosophical reversal of the relation between time and movement. I hope you’ve understood that the two are connected… Okay.
And what remains for us to do, in this third term, is a two-fold research program, at the same time, to try to elaborate, or to see, how an image of time, a direct image of time, a direct time-image, as well as a concept of time itself, is constructed. What would be the relationship between this concept as a philosophical concept and this image as an aesthetic image? Perhaps at the very end, one would be able to say, yes but these three parts are… [Deleuze doesn’t finish the sentence] So, my question refers to this whole second part that occupied us for an entire term – movement, time and the reversal of the relation between time and movement – and that we ended last time, quite arbitrarily, with Kant… Are there any questions, any points that you find problematic? Are there any…? All this is very difficult, so we mustn’t think we have to understand everything. When I say points, I mean… Are there any clarifications you’d like me to make? Was that okay last time, on Kant? Fine.
So now we’ll just move on, and since I don’t know exactly what I said in the first term, you can tell me what we covered and what we didn’t cover. So now I’ll move on by going back to the question of the image and how, in terms of the image, we run into the same adventure as the one we have tried to analyze at the level of the concept.
Indeed, we can say that for a long time, the image of time was derived from the movement-image. I remember that we covered this well during the first term. And how was it derived from the movement-image? In the case of cinema, it was through the intermediary of montage. It was the montage of movement-images that gave… that produced an image of time, an indirect image of time.
What do we mean by this? Immediately we can see that there is something odd about this.
Because such an operation is only valid if we think that the cinematographic image, in itself and directly, exists in the present. And it seems to me that for a long time, and even now, we have what is like a strange postulate, a pseudo-commonplace that is invoked by saying: cinema knows only one temporality, the present tense, the cinematographic image is in the present.
And those people who think this – or who pretend to think that, because people are very complicated… do they really think or do they pretend to think? – Those who pretend to think this, are obviously not perturbed – and I insist on this from the very beginning because we’ve already seen it a little bit – are obviously not perturbed by an argument that would consist in saying, but wait a minute, what about the flashback? Because once again, the flashback – and we saw this, if I remember well, we saw this on one occasion when I was speaking about Mankiewicz, which is to say, a director who makes extensive use of the flashback[2] – the flashback is a conventional procedure the reason for which lies elsewhere. The need for it must come from elsewhere. In itself, the flashback is exactly like a signpost. It’s a sign, it’s a sign saying: “Past this way!” It doesn’t transform the nature of the image. It’s what we would call… it’s what in philosophy we would call an extrinsic denomination. It doesn’t prevent the image from being in the present. It’s just as if there were a sign: “Pretend this is in the past”. It’s not because there’s a fade-out or because you have a little cross-dissolve that plays exactly the role of a sign: “Watch out! Conventionally speaking, we’re now in the past.”
So, the flashback argument doesn’t contradict this idea that the cinematic image is in the present. What compromises this idea from the start, what gives us to think that it’s a bizarre idea… is something else. You understand, however, that it’s all one: in the schema movement-image, montage, indirect image of time that derives from movement-images through the intermediary of montage, from this very vague schema, one passes quite easily to: image in the present, montage, synthesis of time, which distributes past and future in the succession of movement-images, in the succession of presents – it’s the same thing. In the same way that montage extracts an indirect image of time from movement-images, so too does montage extract a synthesis of time from images in the present that will distribute past and future in the passing present… No problem there.[3]
I would say this commonplace that cinema knows only one temporality, that the cinematographic image knows only one temporality, which is the present tense, meaning the present indicative tense, is something routinely stated but that nobody really believes. It’s false. It’s absolutely false. It’s absolutely false. I mean… I’m taking a step back; we had seen, and I’m not going to go back to this, that one filmmaker… for example, I find it astonishing that a filmmaker as modern as Pier Paolo Pasolini, in his theory – even though certain aspects of Pasolini’s theories are surprisingly modern, and even prophetic – on this point, sticks to the old scheme: movement-image, montage, indirect image of time, or what amounts to the same thing: image in the present, montage, synthesis of time distributing present and future in the succession of passing presents.
But I would say no, no, that’s not right. It doesn’t fit because, once again, what we need to immediately question is the idea that the cinema-image exists in the present. Why? I take the case of Alain Robbe-Grillet[4]. Because Robbe-Grillet is not only an extremely important author and filmmaker. Of all the great authors, he is probably the most cunning or the most malicious in his declarations. He very often says: as everyone knows, the cinematographic image is in the present tense. This is even the way he distinguishes himself from Alain Resnais. He says: Oh, myself and Resnais, yes of course, we made Last Year at Marienbad [5] but as for Resnais… what interests him is the architecture of time and so on, so this led to a lot of misunderstandings in our collaboration. And Robbe-Grillet said: But as everyone knows, there is no problem, there is no problem of memory in cinema. There is no problem of time because the cinematographic image is in the present tense.[6]
You cannot read a sentence like this without seeing that it is just one of the provocations Robbe-Grillet so loves to make. Why is this? Because the viewer’s immediate question is: if the cinematographic image is immediately in the present and naturally in the present, if this is an immediate fact of the cinematographic image, why on earth does Robbe-Grillet have to go to such lengths and use such complex means to produce cinematographic images in the present? Because the fact is that it’s not that easy, and that, independently of his collaboration with Resnais, Robbe-Grillet’s entire work will consist in the constitution of images in the present to the point that one can only think that his films cannot naturally be in the present, since it requires such tortuous means and such complex structures to produce these present tense images.
Consequently, if he requires very complicated means to produce images in the present, for example, phenomena of postcard production, very peculiar spaces, systems of repetition, of erasure (gommage), of destruction and creation as he says – an image being both the erasure/reduction and the creation of another image, and so on[7] – in order to obtain a pure present or to obtain pure presents, it seems to mark a clear confirmation that the present is not at all a natural given of the image. And much more, it even gives us a little something more, namely that, yes but of course, the present itself implies time. The present is no less a matter of time than the past.[8]
At this point, we could simply say: well actually, the cinematographic image is fundamentally temporal, and its dimension is not the present, it is an indeterminate time. An indeterminate time, that would suit us well enough…So, an indeterminate time, well yes, time is indeterminate. The cinematographic image would be time as indeterminate. But what does this mean? We don’t know. We don’t know yet.
Let’s try to proceed by an alternative route. Let’s start again, since we’re not the ones stretching the comparison. I was saying that the same thing happened in cinema in a much shorter period of time as to what occurred in the history of thought over a longer period of time and which we saw in the second term. Do you remember what our theme was? We claim to derive time from movement, but here we run into aberrations of movement. And the more these aberrations of movement become independent, the more the time-image ceases to depend on movement, or the concept of time ceases to depend on movement, and a reversal takes place, and it is now movement that depends on time. This doesn’t mean that there is no longer any movement; it means that through a kind of on-the-spot jump, the relation between time and movement is reversed.
Now, we have seen all this from the beginning, just as I said, but from the Greeks onwards movement presented aberrations such that time no longer derived from movement but, in relation to the aberrations of movement it would assume its independence or tended to assume its independence. Now the cinematographic image presents us with the same adventure. The more that aberrations of movement contained in the movement-image… because the movement-image is not separable from aberrations of movement, the movement-image in cinema is not separable from aberrations of movement, aberrations of all kinds, and this is true from the beginning.
Think of the simplest of these aberrations – I think I already mentioned it if I remember well… Ah, but you tell me, you intervene if we have to go back over it – the simplest aberration of movement, it seems to me, can be expressed very simply: it’s the fact that I as an immobile spectator can follow a movement. So, you’ll say to me, in cinema you often have a moving body that disappears; yes, you often have a moving body disappearing in cinema, but it also happens that I as an immobile spectator, I as an immobile body, follow a movement in the process of being made. At which moment I don’t consider myself as mobile or as being part of a moving body that accompanies the movement. I experience myself as immobile, and yet I am/follow this movement in the process of being made. This is an absolutely unnatural condition. In terms of the simplest thing, we could say about how cinematographic perception differs from natural perception, this already constitutes a starting point. I’m not saying that it’s sufficient, but it’s a basis. I am there, and I follow the movement. That is to say, I am given a tracking shot. Now what’s interesting is that I don’t put myself in the movement of the tracking shot, but it’s in being immobile that I follow the movement, such that the moving body doesn’t move away as the movement is made. What is this movement that is made and yet at the same time does not move away from a stationary point? Well, it’s an aberration of movement.
The perpetual changes in proportion of the image, I can say, are also aberrations of movement. The switch from a long shot to a close-up was an aberration of movement. So perpetual changes of proportion imply aberrations of movement. False continuity involves aberrations of movement. Yes, you’ll tell me, but false continuity implies montage. So, can we still maintain an idea of movement-image, montage, and the image of time that derives from it? No. No, we obviously can’t maintain this because first of all, not all aberrations of movement presuppose montage: there are aberrations of movement that occur within the same shot, independently of montage. And when there is montage, the aberrations of movement imply a new conception of montage.
The makers of modern cinema will maintain that montage is the fundamental act of cinema. You may have already guessed who I’m referring to… in my view, the director who invented the time-image, the first one to do so was Orson Welles. Welles always maintained, even contrary to those who believed there were profound mysteries to discover in the sequence shot and in depth of field – for example, the famous text by André Bazin, in which Bazin suggests that with depth of field montage assumes what is only a secondary role – Welles always disagreed very strongly with this, saying that depth of field or not, or whether a rapid shot or a sequence shot, montage remained for him the essential cinematic act. So, the question is not or, at least, not completely that of the disappearance of montage.[9]
But even when montage remains the quintessential cinematographic act, its meaning changes. Indeed, it is no longer the operation by which one draws an indirect image of time from movement-images. So, what is it now? It is the operation… it is no longer – here I’m being careful with my words – it is no longer the operation by which I extract an indirect image of time from movement-images. It becomes the operation by which I determine the relations of time in a direct time-image – you see? – by which I determine my relations of time in a direct time-image. Montage, as Welles says quite rightly, montage remains the quintessential cinematic act, but both its meaning and overall function changes.
Someone, I think, understood this very well… it’s in the issue of “L’Arc”[10] dedicated to Federico Fellini, an article by Robert Lapoujade, where Lapoujade says that in modern cinema, there’s a tendency to secondarize or even eliminate montage as much as possible – he refers to Marguerite Duras and Jean-Marie Straub – and another tendency where montage remains, but it’s no longer the old-style montage. And he proposes… he says – this is his own formula, but in my view, he doesn’t explain it, just uses it in reference to Fellini – he says that Fellini represents a case where we have montage, meaning there is no tendency to make montage secondary. There is fully montage, only it’s a new type of montage. And Lapoujade proposes to call this montrage. He says that what we have is the substitution of montrage for montage.[11]
Ok, but it’s still montage, and he would say this is true also for Welles. Here there is no tendency to secondarize or exclude montage; there’s simply a new conception of montage. He doesn’t explain this well, it seems to me, but in any case, I myself don’t quite understand what he means by montrage, as opposed to the old-style montage. What I would put in this… but at this point, I prefer another word to montrage. I would simply say, well, and I repeat: yes, the old montage consisted in constructing an indirect image of time from movement-images, whereas the new montage consists in determining the relations of time in a direct time-image.
So, my first conclusion – I’ll proceed like this, it’s a question of research and what we’re trying to say is… my first conclusion would be: actually, whether there is montage or not, in any case, it’s not montage that counts. What counts is metaphysics. And I’m not saying this as a philosopher. It’s because all the great filmmakers have said this. This is what all the major cinema auteurs have said, isn’t it? Technique is not a question of doing what you want with it, technique is something that follows. Technique follows. It all depends on what you want to say. That’s why – and I’m not the only one who says this – the great cinema auteurs are thinkers. And the great directors, the great directors have always said: we think and it’s because we think that we use a particular technique rather than another, so it’s not surprising that every great filmmaker jumps from one technique to another, according to their needs, according to the needs of what they have to say, according to the requirements of what they have to say.
So, all I can say is that my question of what constitutes a direct time-image is not related to whether there is montage or not. In fact, it’s related to something else. It’s not the difference between montage and absence of montage that is relevant here. Therefore, neither is it the difference between a sequence shot and a short shot. No, it’s something else. What is it? All we have for the moment as a starting point is: well, yes, it was necessary that aberrations of movement assumed a kind of independence from movement for time to become… [Tape interrupted] [29:11]
… Okay, what we have is no longer: movement-image / montage / indirect time-image. What we have is: aberration of movement / direct time-image, where movement subsists only in so far as it requires a direct time-image. It is the movement-image that depends on the time-image and not the reverse, and that’s why movement now only exists as an aberration of movement. And that’s why we are in the domain of false continuity.
But you will tell me, you will tell me, yes but also before, there was false continuity. Of course, there has always been false continuity in cinema. But everything changes here. It’s a question of re-evaluation, they knew that. Again, Jean Epstein said the most profound things about aberrations of movement within the movement-image itself.[12] But the important point is that no matter how far these directors went in exploring aberrations of movement, these aberrations of movement remained tied to the movement-image, they were considered accidents of the movement-image, so the time-image remained the indirect image of time that derived from these movement-images, taking into account their aberrations. What changed? What changed is that, between the beginning and the end of the war, which is to say during the Second World War, the cinematographic image underwent a mutation… That is to say: aberration of movement is no longer related to movement; aberration of movement exists on its own account. There are now only aberrant movements, from which point aberrant movement depends on the time-image, instead of an indirect image of time depending on a movement-image.
And perhaps this is why we weren’t able to see it before, and the reason it took this mutation for us to do so, so that what had already prepared it in the old cinema could become visible to us today. That’s why Dreyer was so misunderstood and why the false-continuity shots of Gertrud were met with such incomprehension. These are the conditions in which we find ourselves. A short while ago, I was speaking to someone about music, about the case of certain genius composers who find themselves stuck in a very curious situation, but how this impasse becomes the very essence of their work. That is to say, though they belong to an era, they are so far ahead of it that they lack the proper means. We were saying how this was the case – a very famous case in music – the case of Edgard Varèse[13], when he required certain means and he himself said that the means he required were ones he didn’t yet have. And it would be a long time before these means appeared in the obvious form of the synthesizer. Nonetheless – and this was the way Varèse lived. Though he lacked the means, it didn’t prevent him from composing perfect works, did it? Yet he saw himself as lacking the right means, and the richness of his work comes precisely from the fact that it prefigures the means that were lacking, it anticipates the means that were lacking. Well, Dreyer’s is a similar case. His case is quite similar. It’s quite obvious that Gertrud – I wouldn’t go as far as to say that it cannot be understood, because understanding is not so important – but it cannot be experienced… in terms of the nature of its images, it will only be able to be grasped when a revolution has occurred that provides the means that were missing, that were missing both for Dreyer himself and for his viewers.[14]
So now we are faced with the idea that the cinematographic image – you see I’m just summarizing the point I’ve got to – is neither in the present, I would say this is literally a joke, though it’s a good joke… It’s a good joke. I really like the page where Robbe-Grillet writes: as everyone knows, the cinematographic image is in the present. But it’s pure buffoonery, you know it’s clownery, deliberate clownery, I think, in the case of Robbe-Grillet. He loves making a fool of his readers, it’s one of his favorite tricks. Fine. But no, the only conclusion I can draw is: the cinematographic image is not the movement-image, nor is the cinematographic image in the present.[15]
And we are faced with the meagre conclusion that though the cinematographic image is of course in movement, once again, it doesn’t equate to the movement-image since its movement depends on time. So, there is a time-image, and this time-image, this direct time-image that might be considered the ground of cinema and regarding which we could say that once produced, once it appears… for the moment we could even add that it was there right from the beginning. Except that we couldn’t see it. Before, we could see it only indirectly. We could only see it as what derived, through montage, through the intermediary of montage, from movement-images.
Now, whether we have montage or not, there exists a direct time-image. Even when there is montage, there is a direct time-image that is directly expressed through aberrations of movement. At the same time as aberrations of movement become independent from movement, the indirect image of time gives way to a direct time-image… All this is extremely varied, it’s very… you understand, it’s… We’re trying to find a direction. That’s my first point. So, are there any questions? Are there any…?
So, what we’re now saying to ourselves is… yes, but the time-image isn’t something we can find just anywhere. It will become quite complicated since in some cases it passes through the new form of montage, while in other cases, it tends towards a secondarization of montage. In some cases, it emerges in the sequence shot, but in other cases it passes through chopped-up shots, subjected to this new form of montage that Lapoujade calls montrage and that I call the determination of the relations of time in the direct time-image, which suits us well, since we have seen that the direct time-image is precisely the non-present character of the cinematographic image, and therefore constitutes an indeterminate time-image, an indeterminate time. And indeed, one way of determining this indeterminate time-image would be to determine the relations of time in the direct time-image.
Is this okay for you. We can intuit many things now. Are you alright with this? You can stop me, if you’re not… and then I will start again. Everything okay? So, let’s try to make this more concrete. What does this mean? And we’re still at the level of feelings, impressions; these are just impressions.
So, yes, the first kind of impression, I see images in modern cinema that, at first glance, seem to be movement-images, yet at the same time, we as spectators cannot look at them without telling ourselves that they must be something else. I mean, they are movement-images in the sense that there’s a character who moves, or else there’s something that moves, but we as spectators realize that something else is at stake here. It could be because this movement-image no doubt represents or presents sufficient aberrations for us to say that what we have here is something other than a movement-image.
Fine, so what would this be? First case – and here I would like to distinguish two cases – faced with certain movement-images, we tell ourselves that it is no longer a question of traversing space. We are no longer concerned with a local change. It’s not simply a question of displacement in space, it’s also – and I insist on this “also” – or rather, it’s not only a question of displacement in space, but equally of plunging into, or exploring, time. It’s no longer merely a movement in space, it’s also an exploration of time. And the impression of strangeness the movement gives us derives from this. It’s also an exploration in time, or if you like, it’s an exploration of the past. It is in time that the character moves. It is a past that they explore.
So now I’m going to immediately jump to something that will help us move forward, though it’s still purely a hypothesis. I immediately say to myself that this is not because there’s a recollection, it’s nothing to do with the flashback, there is no flashback here. Fortunately, this impression is absolutely independent of that. We just saw or remembered how the flashback was a signpost. But here there is no flashback that functions as a signpost or that tells us: “watch out, this is the pseudo-past”. No, it’s much deeper than that. We have the impression that the character is moving through time, is exploring a past. I’m not saying that he has a recollection. Of course, you’ll tell me… Ah, but… You can tell me what you like but you can’t ask me what this means. All you can say to me is: do you realize what you’re committing yourself to?
And I say, oh yes, I realize what I am committing myself to. From now on I commit myself, to distinguishing and to showing the necessity of making a distinction between exploring the past and having a recollection. I would say that in certain images, the character gives us the impression of exploring a past, of exploring in time, and not of moving in space. It’s as if the place or places the character occupies in space were doubled by a place that he occupies in time, a place irreducible to the one he occupies in space. As Proust said, we occupy a place in time that is incommensurable with that which we occupy in space.[16] Well, it is as though these images were able to develop or deploy this place that the character occupies in time.
Now you will ask me for examples. Examples! First example: a depth of field shot, a sequence shot with depth of field in Welles. It seems to me there’s something here we haven’t commented on so far. I’m saying this, though I’m not really sure, but it’s my impression. At an important moment in film criticism, the question of depth of field was an object of controversy. Today it’s been largely put to rest because we’ve become sensitive to other aspects, but if I come back to this problem, and it was one of Bazin’s great merits to pose it, how exactly was the question posed? Well, it was posed in two ways. The question was posed in the form: to what extent, firstly, to what extent is depth of field new, meaning technically new? Bazin’s answer was that it was new, whereas Mitry’s answer was rather – but we’ll see that this debate becomes really complicated – “But wait a minute, it’s been used since the very beginning” – you see how here we find a problem somewhat analogous to that of our starting point. And the second question was not whether it was new or not, but what was the function of depth of field? To which Bazin’s answered: the function of depth of field is to give us more reality, an “additional reality”. While Mitry’s response was – but again this is just superficially because the matter is very complicated – that this isn’t so. Depth of field is as constraining as any other process and gives the viewer no additional freedom and no gain in reality.[17]
If I recall this polemic before coming to the examples, it is to say something that will have as much importance as this question of depth of field. My first remark concerns: is it new or not? I pass over the technical problems of film, which I’ll leave to the cinema experts who know more about it. Is it a new procedure or not? The answer seems to me obviously, the one given by Bazin: yes, with Welles there is something new in his use of depth of field. Why? Yet it’s true that, from its very beginnings, cinema employed and was indeed forced to employ – at that time, it had no choice – depth of field. I say it had no choice because it obviously had no other option as long as the camera was not mobile, when it took in all fields at once, or rather all shots at once. All this goes without saying.
But what do we call depth of field? I think that here again we mustn’t confuse – just as we mustn’t confuse indirect time-image and direct time-image – we mustn’t confuse depth in field and depth of field. Depth, in terms of the depth in the field, exists since the beginning of cinema, and you already find a perfect example in Griffith. You’ll see that this is a story that’s exactly the same as what happened in painting.
If I describe this fragment of the history of painting – isn’t there a bigger piece of chalk I can use? – where we encounter… and that doesn’t mean they ignore it, does it? Let’s say that for a certain time, depth is something that was obtained through a superposition or juxtaposition of planes, from near to far. You have your painting, you have a foreground, a second plane, a third plane, a background and so on. You see… – and feel free to correct what I say – I would say that this is in no way an inferior solution to other solutions, absolutely not. It has its own perfection. This form of juxtaposition of distinct planes that produces depth is something you’ll find very well explained, for example, in the classic text by Heinrich Wölfflin… [18] “Principles of Art History” [19]. No? Yes? Never mind! In any case, it’s a great classic of art history.
Of course, this can take extremely complex forms, because the planes can curve and they curve differently, it can produce extremely complicated figures, figures of great complexity and concavity. But generally speaking, what do these figures refer to? You have real depth through the succession of planes, from foreground to background. But you will notice that these planes are autonomous. Which doesn’t mean that they don’t have a harmony and that they don’t collude with one another. Of course, there is harmony, it’s even a fundamentally harmonic type of painting. But, in certain sense, whatever harmonies there are between the planes, each plane is only concerned with itself and with its own function.
But what does it mean to say that each plane has its own function? It means that characters who are in some kind of relation will be on the same plane, side by side. So, for example, you’ll have a foreground with two characters: again, there can be an extremely interesting curvature; in Raphael, you find some extremely beautiful curvatures of the foreground, wonderful curvatures. That’s why it’s not a question of saying that this is a solution that has noy yet been perfected. Of course not, it is a solution to the problem of depth. It’s what I call depth in field.
So, let’s imagine you have Adam and Eve… Adam and Eve side by side. This is very important because there are some very well-known images by Welles in Citizen Kane, which are precisely images without depth of field where the couple, Kane and his wife, are pictured side by side.[20] So you also have this in cinema. And in a completely different context, which we’ll come back to, you have Dreyer’s famous side by side shots. The characters there are really side by side. I say in a completely different context – because in Dreyer’s case, there is only a single plane[21], so this was another solution, where there isn’t a succession of planes. When there’s only a single plane or shot, we have a type of image that we’ll encounter with the problem of the time-image.
But let’s go back to painting, where then you’ll have a second plane, for example, which will be an open interior, that is to say: columns supporting a house, with windows and so on. Here we have a second plane that performs its own task, where windows match other windows, and columns match other columns. In the background, through the windows or through the openings between the columns, you have a landscape that itself performs its own task. The collusion in the painting obviously derives from a whole system of harmonies. As I said this type of painting has a profound sense of harmony, a whole system of harmony between the different planes. And the formula is: each plane performs its own task, abiding by the law of existing side by side. Depth is obtained through the succession of distinct planes and through curvatures, the curvatures of each plane which already constitute a kind of aberration, which would be like the equivalent of an aberration of movement in a cinema image. There may be all these aberrations which can lead us quite far, but none of this prevents the whole from abiding by this formula: that depth is given through a succession of distinct planes, each of which performs its own task. Okay.
Now what happens in the 17th century in this respect? What is the mutation that occurs in the 17th century? The 17th century mutation… I can express it in several ways. Something that is almost unimaginable happens: Adam and Eve are no longer side by side. That’s one way of expressing it, Adam and Eve are no longer side by side. In a famous Tintoretto painting, you have Adam who is here… therefore necessarily with his back turned, and you have Eve who is there, and what happens here? You have the diagonal of Adam stretching towards Eve, the two figures not being on the same plane.[22] The diagonal establishes a direct relationship between two planes, between two distinct planes; the distinct planes have ceased to perform their own task. Above all, you can forget the idea that this is old; it’s not old. You see how there is a depth here that is no longer depth in the field, there is a depth of field.
Is this the only form? No. In Rubens, who in this respect is particularly significant, in Rubens you can have something that seemingly maintains the old structure. You have a character placed alongside another character, and just behind there is another character side by side with yet another… You think this is the side-by-side configuration, indeed you can assign to each character the person who is beside him. Except that between the characters on the left and those on the right, a gap emerges and imposes itself, a gap that would be unimaginable in the old model. Everything plunges into this gap that prevents me from relating characters on the same plane to one another. Everything courses, everything rushes through the gap. As Wölfflin says, the painting is “internally hollowed out.” Here we have a second case, it’s this hole through successive planes that prevent the planes from performing their own task. And this is a constant feature of Rubens’ paintings… [Tape interrupted] [59:00]
Part 2
… Third example: Vermeer. I think three is a good number… You have this celebrated element in Vermeer… It’s that there will be a lateral expansion of the foreground, a lateral expansion of the foreground accompanied by a narrowing, a radical narrowing of the background, into an extremely sharp vanishing line, a decrease in… no, sorry, I mean an increase in the dimensions of the foreground and a decrease in the dimensions of the background, so there is a visible ceiling that is necessarily caused by the compression of the dimensions of the background. I insist on this, so you will immediately be able to see where I’m heading: an apparent ceiling, I said. Also, before, there were visible ceilings, but they didn’t convey the same sense, they didn’t produce the same spatial effect. You have light in the background, the famous Vermeer light, there, along with the possibility of the foreground being occupied by shadow, or by shadows. Okay.[23]
In this third example, we see a direct communication between foreground and background or between near and far. Foreground and background communicate directly and are in connection with one another. This is a third way of… Here we break with the autonomy of the successive planes. In other words, what we have is a new type of depth. And I would say that this is what we would call depth of field. Depth of field does not refer simply to an image where there is depth but refers to an image where depth is treated in such a way that it does not allow any autonomy to each of the planes, making one plane communicate immediately with another, that is to say, it establishes a system of diagonal relationships. Otherwise, there is no reason to speak of depth of field, although there may be depth in the image.
So, I’ll take up the question again, this time at the level of… but I might add, how should we understand it? When Paul Claudel analyzes Rembrandt, which he does splendidly – and this will be my last example – and, in particular, Rembrandt’s treatment of depth, he tells us how Rembrandt attains a kind of vibration, which we could call an invitation to recollection. It resembles an invitation to recollection.[24]
So now we have a small part of the answer, at least, to the first question. Okay, depth existed from the beginning, only, as Mitry reminds us – and Mitry is absolutely right on this point – only here, it was perhaps a depth in the field, it was not a depth of field in cinema. For there was a succession of planes, from near to far, with each plane performing its own task. The good thing is that Mitry says this without acknowledging that the same thing happened in painting. He says it when he analyzes a famous scene from Griffith’s Intolerance, the conquest of Babylon. He says that here we have a depth, only this depth is such that each shot remains independent. In the first plane, you have those facing the assault of the conquerors. You have a second plane where the second line of defenders are busy. And there’s a third plane – I don’t know – where there are women who are helping to prepare weapons, etc. Each plane performs its own task.[25]
So, this is the very example Mitry proposes, which shows that, in a certain sense, he is right, he is right in relation to Bazin. Yet there is a novelty in Welles’ use of depth of field, but why? Because Welles is not, as he himself says, a man of the Middle Ages… nor is he, as Bazin says, a man of the Renaissance: in this respect, he is a man of the 17th century. The depth that he establishes, the depth of field in his work, is, it seems to me, exactly the same as the depth of Vermeer. It is the same as Vermeer’s depth. Why is this, and how does he obtain it? Because he redoubles this depth with the help of wide-angle lenses. And what are these wide-angle shots? They are what allows him to laterally exaggerate the dimensions of the foreground while shrinking the dimensions of the background, all of which makes the visible ceilings necessary, the famous visible ceilings of Welles. And, in this respect, between the depth of field in Welles and the techniques of Vermeer, I see no difference, no difference at all. This is depth of field.
So, to answer the question: is this new? I would say, yes, from Welles onwards, it seems to me that there is clearly a depth of field that comes only from this procedure, or rather, one that is acknowledged as such. Not only would I speak about the resemblance to Vermeer, but I would also say as that with this new sense of depth of field you no longer have… you no longer have a plane that is sufficient in itself. There is always a character on one plane addressing a character on another plane. That is to say, you always have diagonal relations that replace… direct diagonal relations that come to replace horizontal relations, horizontal relations existing on the same plane.
Now, though I attribute this to Welles, that’s actually not quite true because everyone knows that he had in fact at least two very important precursors. Namely, this was already the case in The Rules of the Game[26] where you clearly have this space where characters who relate to one another are never on the same plane, and where you also have a constant direct communication between foreground and background. It is the famous depth of field that we find in The Rules of the Game. And finally, among the elders of cinema, the great directors of early cinema, the first to do this, the very first, I suppose one could say that he was the one who invented the procedure, was Erich von Stroheim. And it’s in Stroheim’s Greed that I think you find this very special depth of field where a character in the background directly calls – and not only through the voice, sometimes it’s only by way of the diagonal of the light – for example, to a character in the foreground. So that foreground and background have ceased to perform their own autonomous function and instead exist only in reacting to one another. You have this, for example, in the hero who enters the scene at the back of the room, and the woman in the foreground who is startled by his entrance. He opens the door and a ray of light make a diagonal connection from one to the other. And the woman is startled. You have a direct summoning from background to foreground.[27]
Hence, as regards the first question, it seems to me that there isn’t so much… But why am I telling you all this? We’ll see what happens regarding the time-image. Second problem, what is the function of such a procedure? This is the problem that interests us the most. What is the function of depth of field? Once again, I take up Bazin’s answer. It’s a gain in reality. That is to say, instead of imposing on us – I mean something very simple – instead of the image imposing a prefigured, pre-existing reality upon us, we are placed in the presence of a voluminous reality in relation to which we are free and are able to construct our own reality. Mitry objects that this is not the case, that depth of field, even when understood in this way, is equally restrictive, and obviously, we would wish to agree with him. We want to agree with him because when you have a diagonal, this is also restrictive. You are compelled to follow it. Tintoretto’s diagonal, where you have Adam in the foreground and Eve in the second plane is a diagonal that completely imposes itself in the painting. You are not in front of a voluminous reality, where you are free to trace your own path. The path is imposed upon you, no less than in the other case.
But Bazin’s thesis was much more complicated. He wasn’t just saying that depth of field corresponds to a function of reality or a gain in reality. Because he was the first to say that depth of field produces an excess of theatricality. And this is very interesting. It produces an excess of theatricality, that is to say – he wasn’t stupid, far from it – he knew very well that Renoir in particular, in The Rules of the Game, uses it in the service of a theatrical function. And so depth of field produces an excess of theatricality. Excess, meaning excessive in relation to theatre itself, an uber-theatre. Cinema adds this extra dimension to theatre by the very fact of its being cinema. Bazin offers an excellent example in William Wyler[28], who uses a fixed camera to film a closed set: This would be pure theatre, were it not for the fact that cinema, because it is cinema, is able to transform theatre. This is what constitutes the excess of theatricality. It transforms theatre, providing the director is good enough, providing he knows how to use this excess of theatricality that is typical of cinema. It’s not enough to film a play to obtain the excess of theatricality, you must have the spark of an idea.
Bazin analyzed an example in Wyler’s work, taken from The Little Foxes.[29] Here you have a fixed camera and depth of field, since you know that Wyler took up depth of field, but without using a wide angle. His depth of field is much more… Bazin says that there’s the motionless heroine in the center of the room, in the center of the salon, motionless and icy, motionless and hard. Then there is her husband, the greatest actor there ever was, that is, the most elegant, the only actor who attained a true elegance, namely Herbert Marshall[30]. I say this both for those of you who love this actor and for those of you who don’t know him, so you can go and discover some of his films… Herbert Marshall who is ill, who is having a heart attack. And he asks his wife, the inexorable fox, to go and fetch his medicine. But she remains where she is, she remains motionless. And Herbert Marshall goes out, in a truly cinematographic phenomenon he goes out of frame. You’ll tell me that he could be going backstage in a theatre. But in saying this, you must at the same time sense that the two things are in no way connected, because the question now is, how he going to come back in? Cinema doesn’t have this problem. He’ll reenter the scene in the background from the left, from the staircase that he mounts to go and get the medicine himself, and on which he collapses… [Tape interrupted] [1:15:17]
You’ll tell me, that this is a cinematic form of theatre, but even so, I don’t see exactly how we could obtain it. I suppose you could obtain it through a very special arrangement of the stage, where you have the backstage there, where the guy runs into a tunnel, so that he can come back in here and collapse, yes, but then, if you arranged the stage like this, you would say that the theatre was borrowing from cinema, that is, the stage would be treated as a frame.
So, well, Bazin’s thesis is much more complex, since he himself says that depth of field produces an excess of theatricality. Only he argues that this excess of theatricality, with respect to theatre itself, is at the service of the real. Hence the coherence of his thesis. He could say, in any case, that depth of field has a function of reality and that it gives us a gain in reality. Compared to actual theatre, the excess of theatricality that cinema conveys ensures a gain in reality. So, if you look at this thesis as a whole, you see it’s more complex than it seemed at first.
So why do I say this? It’s because here, I have just one comment to make. I’m not sure that everything in depth of field occurs between the function of theatricality and the function of reality, even in terms of the organization Bazin proposed, that is to say, an excess of theatricality that ultimately serves to give us an additional reality. Because there’s something that strikes me at the moment – we’re gathering elements to try to move forward – I notice that very often depth of field is linked, I won’t say to a recollection-image, but very often, the use of depth of field is linked to an effort to evoke a recollection. And generally, this happens in the form of a contraction: it is through contraction – notably the contraction of the background – that depth of field implies the affirmation of this link with the search for a recollection. Needless to say, Citizen Kane seems to presents itself… even if this is not the whole story it appears to be a search for recollections.
But there is more to it. There’s a famous depth-of-field scene in The Magnificent Ambersons[31] – this is the famous scene that Bazin commented on extensively.[32] But what surprises me is that he speaks about it but then doesn’t build upon his analysis. It’s the scene in the kitchen where – for those who can remember it – where there’s the boy, the little Amberson eating with his poor old aunt, and what’s is this scene about? It’s a rather complicated scene, but it’s shot in depth of field. As Bazin himself says, the young boy, while eating, has only one desire: to get his aunt to stir a recollection in herself. What he’s trying to do is to get his aunt to tell him whether or not his mother was accompanied by someone on a previous trip long before. So, this isn’t directly connected – I take up Claudel’s terms – with a recollection, but with an invitation to remember, it’s a typical scene involving the invitation to remember, or rather it’s in connection with an invitation to remember that this scene in depth of field is set up.
What concerns me here is not the recollection-image. What interests me is the prior effort. It’s in terms of a situation of invitation or evocation, of memorization, a function of memorization, and in relation to the other aspect of depth of field, what we have is not contraction, but, on the contrary, expansion, the kind of depth produced by the phenomenon of wide-angle photography. So, what’s going on here? This is also true of the great depth of field scenes where someone moves in Citizen Kane, for example the great scene where Kane moves to join his close friend the journalist which will mark the moment of rupture in their relationship. And he traverses a whole corridor in depth of field to get to the office where the journalist works. Okay. This is where I would say, rightly or wrongly, that we have the impression that Kane does not simply move in space, but in another way and at the same time, he explores… No, it’s not him who’s exploring, since he’s dead. He makes us explore. His movement, his movement causes us to explore a layer of the past, a region of the past.[33] Here it seems to me that the depth of field irrefutably marks a moment of rupture that has already happened: this was the moment of rupture, not this will be the moment of rupture. Depth of field causes movement in space give way to an exploration of the past as such.[34]
In other words, my answer, if I could add to Bazin’s, would be that depth of field is a function of temporalization. And it reverses – this is what I was trying to say before – it reverses – although it’s not the only way of doing it, as we shall see in a moment – it reverses the relation between movement and time. Movement in space is now merely the index of something deeper, namely the exploration of a sheet of past. It’s no longer “this is the moment of rupture” or “this will be the moment of rupture”, but “this was the moment of rupture”. This was the moment of rupture between the two men: that is the inexorable feeling that we get from depth of field. So, both as contraction and as expansion, it is a function of temporalization. At this level, I believe that theatricality and reality become secondary to … the fact that the character moves in time. The character makes us explore a sheet of past.[35]
I would say – and this is not difficult – it all amounts to saying that depth, when it is true depth, when it is depth of field and not depth in the field, is a temporal dimension. It is not a dimension of space. It places us in time. It is torn from space and propels us into time. Depth of field shows us the place that Kane occupies in time, a place incomparable to the one he occupies in space. And that’s why depth of field is literally torn from space. It substitutes space with time. It introduces us into the direct time-image.
On the same topic, I take another director, Visconti. Because it’s by accumulation that… because if you like, it’s by accumulating materials… we can only reach firm conclusions by accumulating examples. This time we have a tracking shot… Visconti is famous for his tracking shots… there’s a famous tracking shot at the beginning of Sandra. At the beginning of Sandra[36] there is a car moving on a road. We’ll say that this is a movement-image. Okay. What makes this image so odd? It is literally full of aberrations. You can see what I’m getting at right away: it’s precisely these aberrations of movement in the tracking shot that make it something other than a movement-image. It’s already a time-image on which movement depends, for which movement is only the index. And indeed, for most of the shot, we don’t see the passengers in the car. Only when the car stops, do we see the young woman get out, which happens twice. Once, she buys herself a black headscarf that she puts on her head. Another time, I don’t know what, I have the impression she buys some bread from the country or I don’t know what, she buys herself something else.[37] We understand, we will understand that she is coming back to the family home. It is a return. An invitation to remember. It’s not at all a flashback.
I am trying to explain that the invitation to remember is something absolutely specific, that recollection itself is a banality that is of no interest, but that on the other hand, the invitation to remember, or the exhortation to remember, the evocation of a memory, is a fundamental dimension of time and of the constitution of time. It pertains to the constitution of time. Recollection itself is literally useless. We’ll see why it serves no purpose. It serves no purpose, or even worse, it is harmful. But the search for a memory and the exploration of the past is something else.
Again, I come back to my theme: but then, you will tell me, isn’t exploring the past the same thing as recollection? No. But this will have to be shown. For the moment I’m sticking to it, I can’t say why yet. I don’t yet have sufficient material to be able to say how exploring the past has nothing to do with having a recollection. What’s more, you only have recollections when you have finished exploring, so you could well do without having recollections. Exploring the past is good. Having recollections is very, very, very bad. Explore your past but destroy all your memories. Suppress your memories or else you’ll suppress yourself. There was someone in the first term who told me that it’s the same thing, it is exactly the same. He said: “You suppress the things you adore and you suppress yourself. But the things you adore are memories, childhood memories, oh… To explore the past is a different kind of adventure because when you explore the past, what you explore is always the past of others, you are not exploring your own past. You have certain recollections of your past, and nearly always, by nature, they are lousy recollections. But exploring the past is always cosmic. Kafka said: “The history of my art undermines world history”. It means the same thing; it means exploring the past, not our recollections.
So, as I was saying, Sandra goes back to her home town. There will be flashbacks, there will be, oh there will be. I’m not saying there won’t be, but we have to see in what way. It’s all very precise, very, very precise. We have to see in what way the flashback intervenes, and how it is completely derisory in comparison to… It arrives always too late, of course, and is of no interest in itself. Everything has already happened, only once you’ve begun to explore the past, it’s inevitable that recollection-images in the forms of flashbacks start tumbling down like debris. But what interests me now has nothing to do with that. The fact is that while the long tracking shot journeying back to Sandra’s childhood home appears to take place in space, we know from the beginning that it’s an exploration in time, that it’s an exploration of a sheet of past, of a region of the past. And yet, there is no recollection-image. Or at least, no recollection-image is shown. What I want to say here is that the heroine plunges into time more than she moves in space. And Visconti’s image shows us this. That’s what this tracking shot is, it is this sinking into time. Of course, there is a displacement in space, obviously. But it’s like a petrified time. A displacement in space, but with these aberrations of movement it is only there to emphasize something deeper, namely the sinking into time. Here too, the beginning of Sandra is already a direct time-image.
And there’s a very minor short film by Visconti, which is not very well known in France, at least that’s my impression. You can see that it’s one of his obsessions, and that the tracking shot, the way he uses it, has this function. It’s a film which lasts only a few marvelous, marvelous, minutes and which is called Appunti su un fatto di cronaca.[38] Have you seen it? Appunti su un fatto di cronaca… There is something very moving in it, and you can see it clearly, in its pure state, here. The story is about a little girl who was raped, murdered and thrown into a well. And the few minutes of Visconti’s film also consist in a very slow tracking shot in which no one appears, which reconstructs, which simply follows the path the little girl took, that is to say, from her house, along the road, a very impoverished road, to a vacant lot, and the well is in this vacant lot. There’s a foleyed soundtrack, a muffled sound as if evoking distant steps, a struggle, the body thrown in the well, and that’s it. It’s an amazing film. It’s a kind of… it’s wonderful, much more convincing than… more moving than if he had shown the little girl being raped and thrown into the well. It’s very powerful.[39]
Well, you would have to be a Visconti to manage this. It’s not enough to have the idea, again you have to be able to do it. Here, we can feel how much the pseudo-movement in space becomes an index of the exploration region of the past. It’s a zone of the past that is being explored. It’s a direct time-image which at the same time has… it’s an invitation to remember, and at the same time, there is no recollection-image. There is no flashback, which would be a catastrophe, which would show the little girl, for example, being thrown into the well. There is a voice-off commentary. Here again, everything I’m saying, you see how everything I’m saying is something we have to add to our collection of materials, it’s clear that the voice-over, in terms of the question of the direct time-image, the voice of a commentator – we might think of Welles – the voice of a commentator will take on a very specific meaning, which will not be the same as commentaries that might accompany the classic movement-image. Okay, fine.
I could cite other examples of the same nature, such as in Last Year at Marienbad, where the spatial movement becomes completely detached from space, here too through processes that essentially concern and pass by way of the sound-image, for example, the hero who traverses long corridors, long tracking shots – here conceived by Resnais, since he too is a great director of tracking shots – where there is no direct sound. As the voice-off says… here too, as the voice-off commentary says, and puts it admirably, I don’t remember the exact phrase, the footsteps are muffled, it’s as if the ear of the person walking was too distant, the ear of the person walking was too far away, that is to say, that literally, he is not in the same time.[40] The shift, that is to say, the soundless character of the movement, will produce a kind of unmooring, where the aberration of movement will give free rein to a direct time-image. It is in time that the character moves.[41]
So, as I said, regarding my first case of direct time-image – and I repeat – it is certainly not in the flashback that we will find it. Though I suppose we will still have to take a closer look. All this is hypothesis. But what I’m saying is that recollections presuppose something much more profound and that this something is the exploration of the past, the exploration of zones of the past. You will say to me: but how can you explore zones of the past without recollections? And I will tell you conversely, how can you have a recollection, where will you go to look for it, if not by exploring zones of the past? Let’s leave it at that: what comes first, the recollection-image, the memory, or the exploration of zones of the past? What are the differences between the two? Let’s leave it at that for the moment. Okay.[42]
I’ll move on to my second case. As I said, my first case of direct time-image would be the invitation to remember, that is, the exploration of zones of the past or sheets of past as something absolutely different from recollection, Now, under certain conditions the cinematographic image is capable of offering this type of exploration. That’s all.
Second major case: here it is no longer a question of images that make us explore a zone of the past. I would say, it’s the images that give us – and this is not quite the same thing – a little time in its pure state. So, is this a second case? We’ll have to see. Indeed, it is likely that there are many kinds of time-image. Just as there are very particular types of movement-image, as we saw all last year, so too there are also very different kinds of time-image. That is to say, it’s no longer a question of exploring a sheet of past or a zone of the past; it’s something else. What does this concern? It concerns, let’s say, contemplating a bit of time in its pure state. So, in terms of such images, there is all the more reason for the reversal, all the more reason for a radical reversal of the rapport between time and movement. So, what does this consist in? Well, in my view, there is often – though not always – no technical rule. This second case is much more accomplished by a new type of image, which this time is the flat image, the image without depth. It is the image that presents itself in its flatness.
And obviously, if questions relating to depth of field have lost a lot of their relevance in cinema, it’s not only because we’ve become more aware of their relative importance. Even in Welles’ work, it certainly wasn’t the main question. Welles never stopped saying, regarding depth of field versus cut-up shots, that in the end, for him, it amounted to the same thing. It is secondary requirements that make him choose between one or other of the procedures. And in my view, what he means by this is that, whatever the case, he will construct his time-images. either through depth of field or through a new conception and manipulation of montage, what Lapoujade calls montrage.
But here, I’m speaking about something else again, since the flat image is something we find everywhere, this first formula I was talking about earlier, where there’s only a single plane, or almost only a single plane. There is a single plane with a minimum of depth. This reduced depth, this flat image, this flatness of the image is something you find in many of today’s directors. You find it in the work of Hans-Jürgen Syberberg[43]. What’s more, you have directors who have employed both. I’m thinking of Renoir, who employed the flat image, for example, in The Golden Coach, while he used depth of field in The Rules of the Game. So as regards the flat image… fine.
Someone who can be considered almost one of the founders of the flat image – just as I was looking for the founder of depth of field in the true sense of the word – is Dreyer. He inaugurated this image that repudiates depth. And Dreyer says over and over: what does he want to attain through the flat image? What does this repudiation of the third dimension mean to him? It’s the possibility of directly revealing a fourth or fifth dimension by cancelling the third dimension. Indeed, depth is very ambiguous, depth, we can say as well that as introducing depth in the field, it puts time in space, it makes it a sacred dimension of space. As depth of field, it frees time from space. It is a hinge.[44]
So, Dreyer is quite justified in thinking that it is better to suppress depth. Let’s suppress depth, because, in any case, it remains too ambiguous, that is for him, according to him. If we suppress depth, and if we proceed by way of a flat space, then with Dreyer’s famous characters placed side by side, and the false continuity shots between one plane and another – aberration of movement – the false continuity shots, the multiplicity and constancy of these false continuity shots, we will obtain a direct communication between the two dimensions of the flat image, and a fourth and fifth dimension, the fourth and fifth dimension which Dreyer will call time or spirit. What the flat image will produce and present is a little time in its pure state… [Tape interrupted] [1:46:19]
Part 3
… in a completely different way, after all, we had already seen this, we had seen it in a different way in Ozu’s cinema. If you remember in the first term,[45] I grouped together Ozu’s achievements, the role the still lifes played and in what sense we proposed to distinguish them from simple landscapes. Because what do Ozu’s famous still lifes, which also present us with flat or almost flat images, do in terms of their function? They are pure forms through which we evaluate a change. For example, a fruit bowl, a fruit bowl, a piece of fruit, golf clubs, a bicycle propped against a wall, all of which ensure the flatness of the image… the bicycle propped against the wall. Referring to these still lifes by Ozu… referring to these still lifes by Ozu, we evaluate a change. For example, the famous still life with a vase, where first we see the daughter smiling as she looks at her sleeping father. Then a long shot, a flat long shot, a flat long shot of the vase, then back to the girl who has tears in her eyes.[46] Depending on the still life, we evaluate a change. In this case, the passage from smile to tears. The still life is the form of a change.
As Kant says and as Ozu shows – I apologize for this assimilation – the form of change does not itself change, for a simple reason, once again, it’s that if the form of change – and this is a very Japanese kind of reasoning – if the form of change were itself to change, it would have to change into a form which, in turn, would not change, and so on ad infinitum. And if this second form were to change, it would have to change into a form which in its turn would not change, and so on. The form of what changes does not itself change. Kant’s definition of time, as we have seen, was the unchanging form of what changes, the non-changing form of what changes, the immutable form of what changes. This is Ozu’s still life. A little time in its pure state, that’s what Ozu’s still lifes give us, the unchanging form of what changes, which is inseparable from the actual change. Change is the passage from smile to tears, while the unchanging form of this change, is the still life, the vase. So Ozu is obviously being coy when he uses a perfectly mobile object to compose a still life, the still life with the bicycle, which is indeed a marvelous image.[47] You see a bicycle propped against a wall, but don’t think it suffices to prop a bicycle against a wall to compose a still life. That would be like thinking it’s enough to draw a fruit bowl to be Cézanne.
So, I would say this is another type of time-image. This time, it’s through an excess of flatness that the direct time-image will present itself, no longer in the form of an exploration in time, but in the form of a contemplation of time, no longer an exploration of a sheet of past, but a contemplation of a little time in its pure state. Again, a direct image of time.
And if I try to summarize this, I would say, so what is relevant and what isn’t here? What is relevant and what is not? What is relevant is not the difference between the shot… the sequence shot and montage. In the direct time-image, we may still have montage, simply it takes on a new meaning. Montage may tend to disappear, but that’s not what matters.
Second question, then, what is relevant? Is it a question of depth, the image in depth or the flat image? No, that’s not it either. Which implies that, if we talk about the birth of cinema, let’s say that it isn’t accomplished through any unity or complexity of shot or montage and so on. It is still elsewhere. It is still elsewhere. The shot, the organization of shots, the montage or absence of montage, all these are derived elements. Derived elements of what? They are derived elements of something deeper, namely the direct apprehension of time by the image.
I want to alert you, on the one hand, to an interview with Andrei Tarkovsky in “Positif”, number 249, entitled “De la figure cinématographique”[48] and then in Cahiers du cinéma, number 358, the April [1984] edition, there’s a very good commentary on this interview by Michel Chion[49], and the whole theme is Tarkovsky trying to explain what a cinematographic figure is. Well, he says, it’s neither a matter of the shot, nor of the montage of shots, nor of the sequence shot. It’s something else. There’s something that runs through all this and that prevents – this is what interests me in particular – that prevents us from being able to treat cinema as an organization of language-like units. Which amounts to saying that cinema is modulation. Cinema is modulation; if it’s a language, it’s an analogical language. It’s not, it’s not a binary language, it’s not a digital code. So, those who are very advanced will immediately say to me: ah but what about the video image, the electronic image? The electronic image would seem to me to confirmation this, because in the electronic image, what is performed are grafts, grafts of digital code, grafts of digital code that will give this modulation unsuspected powers. And it won’t at all be… it won’t at all be an image articulated in the manner of a language.
So, in this interview, it seems to me that Tarkovsky touches upon something extremely profound. The cinematographic figure isn’t… it cannot be defined at the level of the shot, no more than it can be at the level of montage, it cannot be that. We can almost say that it cannot be defined in isolation. So, what is it? It seems to me that what he calls the cinematographic figure is, I would say, exactly what I call the direct time-image. It’s the direct time-image.
The direct time-image is very strange, since it is in fact a bit like Bergson’s theory, the immediate givens which, by definition, are not immediately given, since they must be found. I could say in a sense that the direct time-image is never directly given. Why do I call it direct time-image? I call it direct time-image because it does not result from movement. On the contrary, it is movement. I cannot grasp it without understanding that it is movement which depends on it, and that from that point on movement is guided by its aberrations. This is what the direct time-image is. This is the cinematographic figure.
So, it can appear in a shot, or it can arrive through montage, but once again, montage will no longer be the same montage that draws an indirect image of time from the movement-image. It will be a completely different operation; it will be the montage… a new montage, a neo-montage, insofar as it determines the relations of time in an indeterminate direct time-image. The indeterminate direct time-image is what I tried to show in both these cases, this bizarre exploration of the past that precedes any recollection that can be evoked, these explorations of sheets of past, or these presentations of time in its pure state. Whether this happens through a single shot, as in a still life in Ozu, or whether through montage, whether it’s through a flat image or through an image with depth of field and so on. In other words, the direct time-image, it seems to me, constitutes the most fundamental definition of cinema.
Except that we haven’t made much progress, because all we have done is to advance a set of hypotheses: that’s all. In order to… not to say what this time-image, which would be the most important in cinema, is but to try to help us understand in what direction… that the first point I wanted to make today. So, let’s start with… Do you want a little pause? Perhaps you want a rest? Are you too hot? Okay, fine… [Tape interrupted] [1:59:13]
… You can make the connection yourself. I mean, the link goes without saying, because these are things we’ve already looked at, but we need to… It bothers me that we are going to miss two classes, because they would have been necessary for us to be able to develop these points. You’ll just have to remember this, okay? You’ll have to make an effort to remember for when we meet again.
So, you remember, I may as well present the former situation and the new one, in the following way. I would say, in the old situation, finally, what determines everything is the sensory-motor schema. What is the sensory-motor schema? It’s the sequence of situations and actions in which the characters are caught up. So that the formula would be exactly – and you find at this level everything we’ve just seen – the formula would be exactly: the sensory-motor situation or situations are such that an indirect image of time is created by montage. What do we call “montage”? Well, this time, it will literally be sensory-motor montage… [Tape interrupted, very long pause] [2:01:23]
… Montage will be the set of sensory-motor linkages such that the sensory-motor situation generates an indirect image of time. This is what we can call a cinema, to use a complicated word, of the actant, since I need a term that is not simply actor – there are always actors, or not as the case may be, but it doesn’t matter – a cinema of the actant. So, what happens at this level? Well, we saw that it corresponded completely to what Bergson called sensory-motor recognition, or habitual recognition, or automatic recognition. We go from an actual object to another actual object that is situated on the same plane. There is a situation, an action, a new situation, another action: situations are linked up on the same plane. Good.
I said, now, the act, what is the fundamental act of the new cinema? Of course, there was already this – I’m not going over this again – there was all this in the old cinema, but there the conditions were not given for it to emerge in its pure state, just as from the beginning, there were aberrations of movement. What’s happening here? I hear something, there’s something making a whistling sound, an awful sound, like it’s going to explode. When it whistles… I’m the one who gets it in the ear… Anyway, so what’s going on in the new cinema? As we’ve already seen, what happens – so I’m going very quickly – is the emergence of pure optical and sound situations.
And I was saying, what we have here is a cinema of the seer. You have characters who are caught up in situations but the problem for them is no longer how to react. They don’t believe in that anymore, that’s not the problem. The problem is that they see something, it is to see. So they may move, and once again they move a lot. We saw this two years ago, but it’s no longer an action, it’s a stroll/ballad [balade]. It’s a stroll, a wandering. Instead of triggering an action that will react to the situation, the pure optical and sonic situation triggers a stroll in the double sense of stroll and ballad, because the stroll/ballad, [bal(l)ade] can be sung. So, it’s a cinema of the seer.[50]
So that I could say exactly – and this is where you will be able to make the connection yourself, between the previous point of view, and this point of view, the current point of view, which here I take up again – I could say first: sensory-motor situation ➛ indirect image of time obtained by sensory-motor montage. Second case, pure optical and sound situation ➛ direct image of time. For what does the seer see? They see time. And here, once again, time is no longer subordinated to movement as it is in the sensory-motor situation. It no longer passes through sensory-motor linkages. It has value in itself, and it is now the movement, which is to say the stroll/ballad, that depends on time. Action was movement insofar as time depended on it. The wandering stroll/ballad, on the contrary, is movement insofar as it depends on time.[51]
And you see how complicated this is, because you see to what extent it constitutes a new image. In terms of the status of the old image, the sensory-motor status, sound images and the visual images exist in a certain relation. And what is this relation? The relation between the sound and the visual component of the image is obviously determined by the sensory-motor schema. If I define the new cinema by the collapse of sensory-motor schemas, it goes without saying that sound and vision will form complete different and unusual relations with respect to the old cinema. So, it’s a cinema of the seer, where time no longer depends on movement, but movement depends on time, and at this point, movement is only the movement of the stroll/ballad, a wandering in indeterminate spaces, in any-space-whatevers, in disconnected spaces and so on. These are all things we’ve already looked at.
From the point of view of recognition, you remember how this new status – the pure optical and sound situation that will give us a direct time-image – referred not to automatic sensory-motor recognition but to what Bergson called attentive recognition. And what happened in this attentive recognition? This is what happened. I would start with an actual image. And what is an actual image? It is a pure optical and sound situation. Also earlier there was an actual image, but in this case the actual image was a sensory-motor image. The sensory-motor image was an actual image that passed from itself to another actual image. What we had was a sensory-motor linkage of actual images on the same plane.
Here, on the contrary, now, with the pure optical and sound situation… I won’t go back to the examples; for those who don’t remember, think of… Yes, I say, there’s something irrelevant there too, in trying to decide what is relevant or irrelevant. It doesn’t matter… confronted with a cinema of pure optical and sound situations, it doesn’t matter whether we’re talking about a studio set or exteriors. That’s not where the difference will be.
In pure optical and sound situations, it’s true that the set obviously takes on a completely new value compared to sensory-motor situations. Because, you remember, in sensory-motor situations, everything that constitutes a description, everything that takes the form of a set, refers to a supposedly independent object, supposedly independent. In a sensory-motor cinema, I can make a set, that is, instead of shooting on location, construct a set. This set stands for something that I can conceive as independent of it… For example, the street in the studio stands for the actual street. The set stands for something else, even if that something else is not given, since it is the set that gives it. It gives it in a certain way, in its own fashion, but here the description of the situation always refers to an object that is supposedly independent of the description. So, there can very well be a set, yet this set still presupposes the independence of its object. There can also be exterior locations. In this case, from these exteriors the camera takes images that stand for a supposedly independent exterior. But when you have a pure optical-sound situation, it’s no longer like this. Though here again, we can proceed either by way of set or by way of exterior location.
So, whether one shoots on a set or on location is not the point. It’s well known that Italian neo-realist cinema began by shooting on location. But this didn’t matter. What counts is that it obtained and created pure optical-sound situations. For example, the key situation for Rossellini – I’m just reminding you of this since we’ve been over it many times – here, we can identify a new type of image that American cinema was absolutely incapable of assimilating – well, at first at least, though it caught up later, namely an optical and sound situation where the character doesn’t know what to do. He knows even less what to do insofar as the whole situation invites him to see. But to see what? This is the whole question, to see what? What kind of voyance?[52]
And Rossellini’s characters, the foreign woman in Stromboli, the bourgeois woman in Europe ‘51, the traveler in that other film…[53], the boy in Germany Year Zero, all these figures are caught up in pure optical and sound situations and, according to what the situation is, they see. But what do they see? They don’t see the situation. It is the situation that makes them see. They do not react. All they can do is wander. But in the case of neorealism, it’s the exteriors that produce pure optical and sound situations. In other cases, if there was… if there was even a… It’s strange, if in classical cinema we could think of a precursor to these pure optical situations, what would it be? Well, in the cinema… you’d have to look for it where sets present themselves as pure sets.
When a set claims to stand only for itself, how can this be relevant to what I’m saying… it can’t be relevant, can it? When a set is presented as only standing for itself as a set, when it doesn’t claim to refer to an object or situation supposedly independent of it… [Noise of someone leaving the room] That’s annoying, damn! Anyway, let’s proceed. What happens to a set that stands only for itself as a set, that asserts itself as a set, that is, that produces a pure optical and sound situation? It’s needn’t necessarily be a studio set. Once again, neo-realism will… But this happens in the musical. Which is very odd because it’s all a zig-zagging movement, it’s very complicated. It seems to me that in American cinema, the musical was the first time they discovered an art of pure optical and sound situations that did not extend into motor activity. You might say, that it didn’t produce an external motor action apart from that. No, because the pure optical and sound situation gave birth to a dancer in the character, because, after all, what is dance? Well, it can be many things. It may be a motivity, but it’s not a motivity of the person. It’s what we could call a supra-personal motivity.[54]
The set, that is to say, the set as set, the set that stands for itself, no longer extends into movement. It is a pure optical and sound situation. And if it extends into something it will be through dance, a dance that is above all rhythm, that is to say, time, time-image. What constitutes dance is a movement that is pronominalized, a depersonalized movement. And, of course, there is the genius of the dancer; naturally, you have the genius of the dancer. But what is the genius of the dancer? The genius of the dancer or the genius of dance is to cause a movement to arise from a person that exceeds him or her and that is ultimately a movement of the world. So, the sensory-motor schemas are broken by, or in favor of, a relation between pure optical and sound situations, the set, direct time-image, dance as rhythm. What arises is a dancer who replaces the actant, the character who acts.
If you take the films of Stanley Donen – remember he’s the director who made Singin’ in the Rain[55] – if you take Donen, there’s a constant procedure in his films: the city becomes a set, it’s already becoming a postcard image. It wasn’t Robbe-Grillet who invented the postcard image in cinema, nor was it Daniel Schmid, it’s not… they take it up and use it in a completely different context, they will use it in a completely new way. But those who invented it were the makers of American musical comedies. In Donen’s films, for instance, you often have an image of a city that flattens out to become a postcard-image, and at that point we have an instantaneous passage from a sensory-motor situation – image of the city – to a purely optical and sound situation – postcard image.
And it’s a function of the flattened image, which is an absolutely flat image, which is a pure set, a set that claims only to stand for itself as set, and which will unfold not through movement, but through dance. You will have a pairing between pure optical or sound situation – or what amounts to the same thing, postcard/flat image – and the rhythm of the dance that will give life and depth, that will produce a whole world, that will constitute a world around this flat perspective. And in the musical, you have a strange thing going on besides the use of colors in the set, where the optical situation is fundamentally full of color.
But this is very important, because if you think about it, what is a dancer? I’ll tell you, a dancer… a dancer is a question of center of gravity, a question of center of gravity. There are two ways to lose one’s center of gravity. You have to lose your center of gravity. The center of gravity is movement. Ahhhhh! You could define movement in terms of the ordinary path taken by the center of gravity. What defines movement are the laws of gravity. The ordinary paths and latitudes of gravity are what movement is. But if you want to dance, what do you do? You have two procedures, you have two methods, if you want to dance, two infallible methods – actually there is a third which is very recent but we’re not going to consider that here. So you have two old reliable methods, okay?
If you want to dance, you have to make your center of gravity sink into your body like a bullet, or like a plumb line that you let go. Who was it that did that? Oddly enough, it was Kafka. Kafka… for me, he’s a dancer. A prodigious art of postures. You can’t understand The Castle without seeing the extraordinary dance of… the land surveyor. But… I would say, this is how Kafka puts it. And Kafka talks about how the center of gravity sinks into his body like a, like… yes, like a rifle bullet, and then? It’s something like that… you should look at the Diaries. It’s a big book, it will take ages to find the passage in question. I swear he says this. It’s a method, it’s the method and at that moment, from your body a dancer is born. Not easy to apply of course, but infallible if you succeed. You have to see Kafka’s body, right? It’s enough to look at the images, the pictures of Kafka. But there is another method: to make the center of gravity float around you, or at least, around the perimeter – which is not easy – around the perimeter, at the border between you and the outside world. Here too, you make a dancer arise within you. So, these are the two ways, plus the more contemporary third method which, in fact, renews a lot of things.
And so, you understand… do you understand what I am describing? There are two great… basically, you have two great dancers in the American musical. Just two. Fred Astaire and Gene Kelly. How does Kelly dance? How does he dance? Kelly has a powerful stocky little body, he’s a prodigious dancer. If you see him, if you see him in Invitation to the Dance[56], he literally gives birth to the dance… what is exciting in the musical is the way in which, in a thousand different ways, he goes from a sensory-motor movement, that is to say, from ordinary steps to dancing. Sometimes it happens in the form of a decisive break, sometimes surreptitiously, yet neither of these procedures corresponds fully to the style of Astaire or Kelly. In the surreptitious passages, we see very clearly that dance is not a sensory-motivity. Dance is the constitution, a kind of rhythmic constitution of a movement of the world as an extension of pure optical and sound situations. It is the opposite of a sensory-motor extension. Sensory-motivity must be cancelled for the dancer to be born.[57]
Which is to say that, as I was insisting from the beginning, there is movement in the time-image. Obviously, movement has not disappeared, but now it is movement that depends on time. Movement depends on rhythm. So, what I want to say is, how does Kelly dance? Well, you can see that he does everything to lower the center of gravity in his muscular, stocky body. He even adds to it. And as he launches his movements, almost like a pendulum movement that brings down the center of gravity, so the dancer is born. Astaire is a completely different case. Astaire, on the other hand, has a kind of slim body, completely flattened. It’s a flat image. And, from his very first steps, it’s as if the center of gravity was at the border of the figure and traced the border of the figure, or even – it only takes one more, small step – was in its shadow. Hence, the summit of Astaire’s genius is the famous dance with his shadows, where he dances with his own shadows, he dances with his own shadows because it is in the shadows that we find his center of gravity.[58]
So you have these two opposing styles. I think… for those of you who remember the wonderful text by Heinrich von Kleist, “On the Theatre of Marionettes”, Kleist’s great text on puppet theater… He tells us that there are two types of grace. Obviously, this isn’t going to please Astaire, but you can modulate what I’m saying. There is the raw grace of a man devoid of intelligence and the divine grace of pure self-consciousness. The bear and the god. This corresponds exactly to the difference – if you remove the pejorative sense of “devoid of intelligence” – I think Kelly performs a dance, that does indeed refer to a grace of being devoid of intelligence. But, here, devoid of intelligence becomes a huge compliment, it has enormous value. Astaire, on the contrary, performs a dance of pure self-consciousness. These are the two extremes of Kleist, the two ways of being a marionette. Fine.
Why am I saying all this? I’m saying all this to assert… you see, the set is… whether interior or exterior, that’s not what counts. What counts is being able to obtain pure optical and sound situations. This long excursus on dance is something I won’t need to go over again. You need only remember the elements, even if you don’t agree fully, since I made it only to say that in the musical, you do not have sensory-motor linkages. You have, on the contrary, connections that I will call – and the name will be useful – non-localizable connections, following the expression used in physics or chemistry in speaking of non-localizable liaisons. You have non-localizable connections, because where and when does the dance begin? You have non-localizable connections between pure optical and sound situations, on the one hand, and a movement of the world, on the other. In other words, the sensory-motor schema has collapsed, so that these movements of the world are in fact time-images, they are rhythm. That’s why I wished to develop this, but I would add that it shows how pure optical and sound situations can be both simple sets and also exterior locations, as occur in the early phase of Neorealism.
So then, what’s going on here? I would just like to finish on this point, what is happening? What I’m saying, is that this has nothing to do with sensory-motor recognition, where I pass from an actual image to another actual image while staying on the same plane, through a series of sensory-motor linkages. Here, I have an actual image. And in fact, it is cut off from its extension. I can no longer pass from an actual image to another actual image. I have an actual image that is the pure optical and sound situation. There is no longer any sensory-motor linkage that would let me pass to another actual image. So, what happens? Well, this actual image, this purely optical and sonic situation, will make me see: I become a seer. The dancer is a seer of a certain type. Look at Minnelli, the identification you find in all of Minnelli’s work between the dancer and the seer. This is a fundamental element in Minelli. And this would provide me with some very strong arguments regarding… well, never mind. What was I saying? Yes, the pure optical and sound situation is not linked to any extension. I can’t pass from one actual image to another. So, I have an actual image. My question is: what happens then? How does it continue? By sequence, by non-localizable connections. Non-localizable connections, you see how this implies a time-image. I’m stuttering more and more, it must be these time-images!
Well, but… you see I always return to the actual image, I’m stuck there, in the pure optical and sound situation. I’m caught in an actual image. How does it continue? And through what? I have no choice: it can only be prolonged through the virtual. Instead of an actual image passing into another actual image on the same plane, there will be a linkage of an actual image to one or more virtual images on different planes. Oh, but what is the virtual image? I’ll return to my eternal example: the bourgeois woman in Europe ‘51. She sees the factory, a pure optical and sonic situation, the factory siren and so on, and she says: “I thought I was seeing convicts”, “I thought I was seeing convicts”. Notice that she doesn’t say, “I remember, or it reminds me of a prison”. I thought I was seeing convicts. A pure optical and sonic situation that makes her see something. What does Rossellini’s art consist in? The convicts are a virtual image. There is a circuit of actual image-virtual image. A bad filmmaker would have suddenly shown an image of a prison. That’s not possible, because an image of a prison would be another actual image. The image must remain virtual. I thought I was seeing convicts.
There will be, and this is where I would like to finish, the image, the pure optical and sound situation is an actual image that will enter into a circuit with one or more virtual images, passing through different planes. What are these virtual images? What are they? What are these virtual images that linked up with the actual image by means of non-localizable linkages? If I can answer that, we’re home and dry. We will know what the time-image is. So, from… [The recording stops abruptly, and then the start of the following recording appears here momentarily] … Oh fine, ah, ah. [End of the recording] [2:33:22]
Notes
[1] Deleuze refers here to VE (Victory in Europe) Day on the 8th of May, a holiday which celebrates the end of World War II and the liberation of the French people.
[2] See Session 7, January 10, 1984.
[3] On these questions of philosophy that translate into cinema, see The Time-Image, op. cit., pp. 34-37.
[4] Alain-Robbe Grillet (1922-2008) was a writer and film director who was one of the key figures of the French Nouveau roman a literary movement that set out to replace psychology, characterization and linear time-flow in the novel with pure surface description, temporal and spatial disorientation and a degree of meta-fictional self-awareness, elements which seeped into Robbe-Grillet’s filmmaking practice. His most well-known films include The Immortal One (orig. L’Immortelle, 1960), Trans-Europe Express (1965), The Man Who Lies (orig. L’homme qui ment, 1968), Eden and After (orig. L’Eden et après, 1970) and Progressive Slidings of Pleasure (orig. Glissements progressifs du plaisir, 1974).
[5] Last year at Marienbad (Orig. L’année dernière à Marienbad, 1960) is a film directed by Alain Resnais from a screenplay by Alain Robbe-Grillet and starring Giorgio Albertazzi and Delphine Seyrig. Partly inspired by Adolfo Bioy Casares novella The Invention of Morel, the film concerns a man and the woman he may or may not have met and had a relationship with (she seems to deny knowing him) in the luxurious though lugubrious hotel of Marienbad (or some such similar place). In the film, the couple appear to inhabit different regions of time, while their supposed relationship is further thwarted by a third character, a sinister master of ceremonies who may or may not be the woman’s husband. Along with Resnais’ previous film Hiroshima mon amour, on which he worked with Marguerite Duras, the film represents a quite unique collaboration between director and screenwriter, with Robbe-Grillet clearly having input also in directorial decisions.
[6] On the question of the present tense in cinema, see The Time-Image, op. cit., pp. 37-39.
[7] The reference is to Robbe-Grillet’s book For a New Novel (trans. Richard Howard), New York: Grove Press, 1965, where gommage is often translated simply as “destruction”, whereas it can mean both “erasure” and “reduction”. See Sessions 3 and 9 (1983) and Sessions 4 and 5 (1983-1984). See also The Time-Image, pp. 7, 12, 37, 69.
[8] On these questions in Resnais and Robbe-Grillet, see The Time-Image, op. cit., pp. 101-105.
[9] On Welles’ use of depth of field, see The Time-Image, op. cit., pp. 39, 105-106.
[10] “L’Arc” was a trimestral French cultural review that ran between 1958 and 1986.
[11] See The Time-Image, op. cit., p. 59. Here Deleuze quotes Lapoujade without giving the date of the issue of “L’Arc”, 45 (1971). See also Session 9 on January 24, 1984.
[12] On Epstein, see Session 9, January 24, 1984; see also The Time-Image, op. cit., p. 36.
[13] Edgar Varèse (1883-1965) was a French (later American) composer who pioneered the use of concrete sounds in music. His deployment of unusual, organic rhythms and often harsh percussive timbres had a great influence on the development of musique concrète, electro-acoustic, electronic and noise music as did his conceptual ideas on music as organized sound or noise, sound-masses and crystallisations. Though Varèse produced relatively little music in his lifetime, many of compositions had a paradigm shifting importance. His most celebrated pieces include Ameriques, Ionisation (the first classical music work to feature only percussion instruments) Ecuatorial, which Varèse conceived for theremin and later adapted for ondes Martenot (two early electronic instruments that anticipated the modern analogue synthesizer) and Deserts, one of the first works to make use of manipulated electronic tape sounds (anticipating today’s digital sampling).
[14] On Gertrud, see The Time-image, op. cit., p. 41.
[15] See Deleuze’s reference to Robbe-Grillet’s discussion of description in The Time-image, op. cit., p. 7 and especially pp. 45-46. For Robbe-Grillet’s reflections on the nature and temporality of the cinema image, see For a New Novel op. cit, pp. 21, 144-145, 149, and especially 151-152 in relation to the supposed “present-tense” of cinema. See also Deleuze’s analysis of Robbe-Grillet in sessions 18 and 19 of the Seminar on Cinema 1, May 11 and 18, 1982.
[16] Deleuze refers to Proust in The Time-image, op. cit., p. 288, note 20, in relation to a passage in Le temps retrouvé (Finding Time Again, trans. Ian Patterson, London & New York: Penguin Classics 2003, epub edition; also translated as Time Regained). The relevant passage here is: “[…] puppets steeped in the intangible colors of the years, puppets which were an expression of Time, Time which is normally not visible, which seeks out bodies in order to become so and wherever it finds them seizes upon them for its magic lantern show. As intangible as Golo had once been on my bedroom doorknob in Combray, the new, almost unrecognizable d’Argencourt stood there as the revelation of Time, which he rendered partially visible.
[17] On Welles and depth of field, see The Time-image, op. cit., pp. 38-39. On the debate between Bazin and Mitry, see The Time-Image, note. 16, pp. 298-299.
[18] On Wölfflin, see Session 6 of the seminar on Painting, May 19, 1981. See also the seminar on Leibniz and the Baroque, sessions 1 and 2, October 28 and November 4, 1986. Cfr. also The Time-Image, op. cit., p. 298, note 13, and The Movement-image, op. cit., p. 221, note 25.
[19] See Heinrich Wölfflin’s article, “Principles of Art History” in The Art of Art History: A Critical Anthology (Donald Preziosi ed.), Oxford: Oxford UP, 1998.
[20] In this regard, see The Time-Image, op. cit., pp. 105-106.
[21] In French, the word plan denotes both geometrical “plane” and cinematographic “shot”.
[22] The reference here is to The Temptation of Adam and Eve, circa 1550. However, contrary to what Deleuze says, in the painting it is Eve who reaches from the background plane towards Adam in the foreground.
[23] According to Deleuze, in note 13, p. 221, in The Time-Image: “Wôlfflin analyses the baroque spaces of the seventeenth century on the basis of Tintoretto’s diagonals, the anomalies of dimension in Vermeer, the gaps in Rubens, etc.”
[24] Deleuze quotes Claudel in The Time-image, p. 298, note 14: “Sensation has aroused recollection, and recollection in turn reaches and successively destroys the superimposed layers of memory”, The Eye Listens, (trans. Elsie Pell) New York: The Philosophical Library, 1950, p. 40 (translation modified).
[25] See on this subject The Time-image, op. cit., pp. 298-299, note 16.
[26] The Rules of the Game (orig. La règle du jeu, 1939) is a film by Jean Renoir starring Nora Gregor, Paulette Dubost, Julien Carette, Marcel Dalio and Renoir himself as a jocular but somewhat sinister master of ceremonies. Most of the action of the film is set during a gathering at a lavish country estate where, between rabbit hunting parties and various other episodes, including a masked ball, involving misidentifications and jealous misunderstandings in which both the gamekeeper and his wife are caught up, the story unfolds of a young pilot trying to win back his beloved from the rich landowner she has married. Initially criticized, cut and banned by the French government during the war for its critical portrayal of the callous French upper classes, the film was eventually restored and is now considered a masterpiece of world cinema.
[27] On Renoir and Stroheim, see The Time-image, op. cit., p. 108.
[28] William Wyler (1902-1981) was a German-born American director, one of the most successful of the classic Hollywood era, though more a craftsman than an auteur. Ranging across several genres, his most significant films include Jezebel (1938), The Little Foxes and Mrs Minniver (both 1941), The Best Years of Our Lives (1946), Roman Holiday (1953) The Big Country (1958) and Ben Hur (1959).
[29] The Little Foxes (1941) is a film directed by William Wyler based on a play by Lilian Hellmann, starring Bette Davis, Henry Marshall and Teresa Wright. The story, set in the Deep South, revolves around the scheming of avaricious siblings to gain control of the fortune of the sister’s dying husband which they want to invest in a business venture.
[30] Herbert Marshall was an English stage and screen actor who found fame in Hollywood after appearing in Hitchcock’s Murder (1930), Lubitsch’s Trouble in Paradise (1932) and Sternberg’s Blonde Venus (1932). He went on to appear in a number of key films of the 40s and 50s including Hitchcock’s Foreign Correspondent (1940), Wyler’s The Little Foxes (1941) and Otto Preminger’s Angel Face (1953)
[31] The Magnificent Ambersons (1942) is a film by Orson Welles, his second after the critical triumph of Citizen Kane. Starring Joseph Cotten, Ann Baxter, Tim Holt and Agnes Moorhead, the film is an adaptation of the novel by Booth Tarkington about the declining fortunes of a prominent Midwestern family and the concurrent rise of a man they once snobbishly spurned who becomes a wealthy and successful automobile manufacturer. Welles’ experience of making the film marked the beginning of his disillusionment with the Hollywood system, owing to the fact it was brutally cut and the ending changed by the studio after a negative preview screening while the director was in Brazil making the US government funded travelogue, It’s All True. Though the negatives of the removed footage were later destroyed, making restoration of Welles’ original vision impossible, it is still considered one of the greatest films of the classic Hollywood era.
[32] See The Time-image, op. cit., p. 112.
[33] Regarding the scene of the moment of rupture between Kane and his journalist friend Leland in Citizen Kane, it is worth mentioning that the scene has two mirroring parts, the first where Kane walks across the floor of the newsroom to Leland’s office where the drunk journalist has just written a review panning Kane’s talentless opera singer wife, the second where Leland himself walks across the newsroom to where an infuriated Kane, enlarged in the foreground, is rewriting the review in more favourable terms and who announces that he is fired.
[34] See Chapter 5 of The Time-Image, “Peaks of Present, Sheets of Past”.
[35] See The Time-Image, op. cit., pp. 109-11.
[36] Sandra (a.k.a. Vaghe Stelle dell’Orsa, 1965) is a film by Luchino Visconti starring Claudia Cardinale, Jean Sorel and Michael Craig, which revolves around the incestuous relationship between Sandra and her brother which is rekindled when she returns to her family home in Volterra from Paris with her fiancé. Trying to hide this secret from him, with tragic results, she at the same time tries to reveal the secret of her mother and stepfather who she believes denounced her Jewish father to the Nazis.
[37] Here Deleuze partly misremembers the details of the scene he evokes from Sandra, though this in no way invalidates his analysis of its function. The film actually begins with a short prelude scene at a high society party in a plush Parisian house where Sandra appears visibly troubled by the melancholy music the pianist is playing, Cesar Franck’s Prelude, Chorale and Fugue. This cuts to the scene of the couple’s car journey south which occupies the opening credit sequence, and which is shot from within the car itself so for most of the sequence we see only the passing landscape. The sense of the journey plunging into time is underlined by the continuation of the same music on the soundtrack. When we do eventually see the car, Sandra is already wearing a black headscarf while its approach towards Volterra, the town of her ancient family home, is shot from behind so we only see the backs of the couple’s heads, with the image focusing on the streets and buildings ahead.
[38] Appunti su un fatto di cronaca (1951) is a documentary short by Luchino Visconti.
[39] Deleuze comments on both this film and Sandra in The Time-Image, op. cit., p. 38.
[40] The phrase Deleuze refers to here, by Robbe-Grillet, is worth repeating in full, in terms of the mesmeric Peguy-like flow of its accumulating repetition, which in the film fades in and out of hearing as it accompanies the opening tracking shots like a loop: “Silent rooms where footsteps are absorbed by carpets so heavy, so thick that one hears no step, as if the very ear were far away, far away from this numb, barren decor, far from this elaborate frieze beneath the cornice with its branches and garlands like dead leaves, as if the floor were still sand and gravel, or stone slabs over which I advanced once again…” (text taken from English subtitles of Last Year in Marienbad; cf. the French text in Robbe-Grillet’s ciné-roman, L’Année dernière à Marienbad [Paris: Minuit, 1961], p. 27).
[41] On Resnais’ tracking shots, see The Time-Image, op. cit., pp. 116-125.
[42] On the zones of the past, see The Time-image, op. cit., pp. 54-55.
[43] Hans Jürgen Syberberg (b. 1935) is a German film director, most well-known for his lengthy works that combine film with Brechtian theatre techniques and operatic staging. These include the so-called German Trilogy comprising Ludwig: Requiem for a Virgin King (1972), Karl May (1974) and Hitler: a Film from Germany (1977) and his film version of Wagner’s Parsifal (1982).
[44] On this subject, see The Time-Image, op. cit., p. 178.
[45] See sessions 9 and 10, January 24 and 31, 1984.
[46] This scene occurs in Ozu’s Late Spring (1949) starring Setsuko Hara in the first of six films she made with Ozu along with Ozu’s longtime actor-collaborator Chishū Ryū. The film tells the story of a young woman who lives with her widowed father and who does not wish him to remarry, which he wants to do partly to compel her to find a suitable partner with whom to continue her life.
[47] The image Deleuze refers to here appears in A Story of Floating Weeds, a 1934 silent film set in the world of kabuki theatre which Ozu remade as Floating Weeds in colour and with sound in 1958.
[48] Deleuze gives the number as 249, December 1981, in The Time-image, p. 288, note 25.
[49] Michel Chion (b. 1947) is an experimental music composer and film and sound theorist, author of several seminal texts on the relationship between sound and image in cinema including The Voice in Cinema, Audio-Vision: Sound in Cinema and Music in Cinema.
[50] Deleuze speaks of strolling or wandering every year: see Sessions 17, 18, and 19 of the Cinema 1 Seminar (4, 11, and 18, 1982); Sessions 10 and 14 of the Cinema 2 Seminar (February 1 and March 15, 1983); and Session 10 of the current seminar, January 31, 1984. See also The Time-image, op. cit., pp 9, 41, 88, 103, 185.
[51] See The Time-Image, pp. 40-41, for this precise discussion.
[52] On Rossellini’s cinema, especially Europe ’51 and Stromboli see Session 18 of the Cinema 1 seminar, May 11, 1982; Session 23 of the Cinema 2 seminar, June 7, 1983; and sessions 8 and 9 of the current seminar, January 17 and 24, 1984. See also The Time-image, op. cit., pp. 1-2, 19, 23, 45-6, 171-2, 247-8, 251-2.
[53] Voyage in Italy (a.k.a. Viaggio in Italia, 1954) is a film by Rossellini, the last in the trilogy of post-war film he made with Ingrid Bergman, following Stromboli and Europe ’51. Also starring George Sanders, the film follows an estranged couple on a journey through Italy during which the woman wanders off alone to discover the country’s ancient history, until they meet again and go to look at the ruins of Pompei where they are shown the recently uncovered remains of another couple who perished there, an experience that prompts a seemingly miraculous reconciliation during a religious procession in Naples.
[54] On the musical, see Session 11 of the Cinema 1 seminar, March 2, 1982, and Sessions 9 and 14 of the current seminar, January 24 and March 20, 1984; see also The Time-image, op. cit., pp. 60-63.
[55] Regarded as one of the greatest of the Hollywood musicals, Singin’ in The Rain (1952) was co-directed by Stanley Donen in collaboration with its star, Gene Kelly, who directed and choreographed the film’s numerous dance routines including the spectacular “Broadway Melody” sequence.
[56] Invitation to the Dance (1956) is an anthology dance film directed by Gene Kelly following three films he co-directed with Stanley Donen. An unusual aspect of the film is that it has no spoken dialogue only music, dance and mime.
[57] On Kelly and Astaire, see The Time-image, op. cit., p. 61.
[58] Swing Time (1936) is a musical comedy by George Stevens starring Fred Astaire and Ginger Rogers. Astaire plays gambler and dancer trying to raise money to secure his marriage who on meeting a young dance instructor begins dancing with her and falls in love. In terms of its choreography, the film is considered by many to be among the finest of the ten musical comedies Astaire made in partnership with Rogers.
For archival purposes, the augmented and new time stamped version of the complete transcription was completed in July 2021. The translation and revised transcript were completed in December 2022. Additional revisions were completed in February 2024. [NB: The transcript time stamp is in synch with the WebDeleuze recording link provided here.]