ON DELEUZE'S CINEMA (1991)

On Deleuze's Cinema

(1991)

Near the end of his writings on the cinema [1] Gilles Deleuze almost regresses from his intellectual rigour over the new cinema of time, and mourns the passing of the silent cinema. He saw that, with the emergence of the talkie, we lost 'a kind of naturalness . . . the secret and beauty of the silent image' (TI:225), which presented us with 'the natural being of man in history or society' (TI:226). Yet, while in the talkie speech recovers 'discourse' (whose dead-end is filmed dialogue), leaving the images of film almost forgotten, with sound was brought a new dimension in the denaturalised image, readable 'like a musical score' (TI:235). This is the speed with which Deleuze moves from the love of one cinema, to the excited anticipation of its advancement, and even replacement. Here, in his book _Cinema_, moving through a very personal history of the cinema, we can see an example of the pull of his writing, and the radically cinematic way that he watches and reads film. As Keith Reader has noted, 'Deleuze gazes from a place very different to that learned by most of us'. [2]

What *is* important in this introduction is for it to be made plain that, for Deleuze, cinema is neither 'a universal or primitive language system [language], nor a language [langage]' (TI:262). What cinema does is more central than any 'meaning' to its system: 'It brings to light an intelligible content which is like a presupposition, a condition, a necessary correlate through which language constructs its own 'objects' (signifying units and operations)' (TI:262).

Deleuze is saying that films are not mere relays for meanings, that cinematic expression is complex, and that we should not attempt to distinguish 'implicit' and 'explicit' meanings as this distinction, in the end, does not exist. Deleuze also seems modest about his aims, asserting that his image 'classifications' are not so much an analysis of the image, as an indication of 'certain affects whose relation to the cinematographic image *remains* to be determined' (TI:265, my emphasis). What Deleuze maps in _Cinema_ is a movement from the conception of the screen as a window on another world, to that of a table of information.

This discussion will not be a 'book review', not an traditional author-based study, and any passages that I find either fruitless or uninteresting will most likely be left without too much attention paid to them. _Cinema_ is at once a great personal history of the cinema (from Eisenstein to Godard), a work of philosophy (with modern cinema recreating its concepts), and a theory of film, i.e., the grounding of, and the relations between, 'the concepts that cinema gives rise to' (TI:280). It is the latter position of the books that will be the concern of this discussion. Also, without the extensive viewing of the films that Deleuze discusses, the argument, to a certain extent, has to be analysed from its more essential conceptual pointers. The amount of classifications deny me the space to give cross-the-board examples from the cinema, and I can only give them where the concepts would not be self-explanatory. In reviewing the first volume, Dale Jamieson and Barbara James concluded that 'the trip through the funhouse can be rewarding. Just remember where you parked your car'. [3] My pleasure in this commentary is in easing out Deleuze's naked assertions from under the loose mass of film analyses, and the thick rhetorical body of dictionary straining neologisms, to find why Keith Reader notes that _Cinema_ 'will not prove possible to ignore'. [4] My pleasure (hopefully) is to *drive* through the funhouse.

Deleuze's conception of the movement-image takes up the whole of the first volume of _Cinema_ and begins with a minute break-down of the type of 'movement' we get from film. In what way is it an illusion, or a false or artificial movement; just because of the means by which we receive it? Deleuze asserts that 'cinema does not give us an image to which movement is added, it immediately gives us a movement-image' (MI:2). Deleuze derives this thesis from the philosopher Henri Bergson who opposed real movement and concrete duration (durée), with immobile sections and abstract time. Real movement in everyday life is indivisible, unstoppable, and thus in no way can it be reconstituted, while in the cinema what we get is false movement but which is not received by the viewer as in any way artificial. Thus, Deleuze sees cinema's sections ('immanent material elements' (MI:4)) not as immobile (even though it is immobile instantaneous photogrammes which make up celluloid film), but as mobile ones, which are given ('to us') along with an abstract movement or time, to make up the movement-image (l'image-mouvement).

Further, Deleuze describes the movement we are given as a function of equidistant any-instant-whatevers ( in the mechanism of cinema, Lumière's claws pulling the perforated film of Edison and Dickson), which can be ordinary or singularly remarkable and privileged. Yet all Deleuze seems to be able to offer us as an argument for our taking up of these classifications is the famous photograph, taken by Muybridge, showing a horse at the instant when all four of its hooves are off the ground at one time. This may seem well enough a 'privileged' instant, but how far, and to what extent, will this classification be of use to us in the future. We will thus find throughout _Cinema_ two types of semiological classification: the purely descriptive (even though in a philosophically incisive way), and the formally qualitative; signs that seem to come from the same theoretical idea or instinct, and nevertheless feed off each other.

Deleuze then extends the cinematic line of changing sections, positing movement as 'a mobile section of duration, that is, of the Whole, or of a Whole' (MI:8). Moreover, movement, being a 'translation in space' (MI:8), *expresses* a change in duration: just as when a sugar lump is placed in a cup of water there is a transition from this, water + sugar lump, to 'sugared water', and thus a qualitative change in the Whole. Deleuze gradually leads us to see that while the instant/movement development is an illusion, the movement/change one is real, and is something that we perceive. The Whole itself is constituted by flux and change, it is neither given nor giveable (it is in this sense that Deleuze refers to it as a 'spiritual reality' (MI:11)), and is not in any way determinate but in fact 'Open'. By this Deleuze seems simply to be referring to the complete text/body/whatever-you-will of the film, whose meaning is never closed, but which creates itself. Deleuze argues this by relating it to the premise that you cannot pluck movement from duration, such that if we were to find ourselves in, or confronted with, a duration 'we may conclude that there exists somewhere a whole which is changing, and which is open somewhere' (MI:9).

The Whole is not made up from objects but is defined by their external relations: with movement the objects or parts of a set (which are necessarily closed and are *in* space) change positions in relation to one another, and so qualitatively transforms the whole. (Confusingly, Deleuze changes from using an upper case W to using a lower case one, perhaps to differentiate between a completed text -- *Whole*, and a text still undergoing changes -- *whole*) Furthermore, 'movement relates the objects of a closed system to open duration, and duration to the objects of the system which it forces to open up' (MI:11). What Deleuze is getting at here seems to be that the set (closed) and the Whole (open) inform each other, the set, for example a sequence of shots, relates to other sets, informing and changing the meaning of the Whole, while the very growing of the 'wholes' *when perceived in this way* eases open the set relations for our greater understanding -- we, in effect, *see* more in realising that there is a whole that is changing ('before our very eyes').

III

Next Deleuze turns his attention to the composition of the frame and thus his varied assortment of image classifications. It is in the second volume of _Cinema_ that Deleuze in fact realises (or remembers) his basic movement-image regime: organic (kinetic). An organic description is one where the setting shown is totally independent of the camera's view, 'and stands for a supposedly pre-existing reality' (TI:126), in which we do not feel the attitude of the camera. One can find the pure descriptions of this regime everywhere where the director wants to return us to the real, but especially in the cinema of Italian neo-realism. Framing here is a closed system that includes everything that is present in the image (whether movement-image, or time-image), i.e., all characters, scenery, props, backgrounds, shades and colours. It constitutes the 'in-field'; its definition is exclusion or limitation. Moving outside of realism it has two extremes of expression, those of rarefaction and saturation, with the rarefied image tending toward a focusing on a singular object (for example, the container of milk, both in Hitchcock's _Suspicion_ and Tarkovsky's _The Sacrifice_), or its most extreme degree in the sparse or completely empty shot (found in Ozu, Bresson, as well as Antonioni). The saturated image can be found in any image, although most often in the wide screen and shots with a large depth of field where various different scenes or actions are taking place simultaneously (this further relates to a later concept of montage *within* the image) Either way, Deleuze argues that the frame is 'legible as well as visible' (MI:12), which seems crudely the restating of a fact that has been asserted since the late 1960's, i.e., that we do not merely see the image but *read* it.

Deleuze defines shots as 'fixed spatial determinations, slices of space or distances in relation to the camera' (MI:25). Movement expresses a change in the whole while changing the respective positions of shots through cutting. Deleuze argues that the shot, being a mobile section of duration, 'continuously divides duration into subdurations which are themselves heterogeneous, and reunites these into a duration which is immanent to the whole of the universe' (MI:20). Put a little simpler, a shot modifies its set (dividing) and so effects a change in the whole (uniting).

Deleuze asserts that the frame is, and always has been, either geometrical or physical, 'in relation to chosen coordinates or in relation to selected variables' (MI:13), that is, formed with pure shape, light/shadow, and line, or in conjunction with persons, crowds, places, etc. The frame is infinitely 'dividual' (MI:14), such that light and dark dynamics increase and separate the parts of the image to give, literally, more of the image over to us. Deleuze then goes on to classify certain geometrics of the image in relation to their authors: the horizontals and verticals of Dreyer; the diagonals and pyramidials of the Expressionists; the natural lines of earth and sky in Ford. Simple descriptions that are interesting to note, and Deleuze attempts to add some sort of an role to this image in pointing out that its formal focus goes some way towards reaching the 'theme' of the Whole.

This is related to Deleuze's consideration of the out-of-field (hors-champ) which refers 'to what is neither seen nor understood, but is nevertheless perfectly present' (MI:16), and thus still part of the cinematographic image. Thus all that the camera does not frame, and the sound that is not directly linked with an object or subject within the image; both constitute the larger set, or 'plane [plan] of genuinely unlimited content' (MI:16), of which the in-field is a set within (and the Whole is a thread of these sets of sets). Thus sounds (e.g., voices off, walking sounds, crashes) dwell in, and sometimes fill the out-of-field, achieving 'different dimensions' to the movement-image, making the 'communication between image and a whole . . . increasingly rich and complex' (TI:241). The out-of-field is imaginary yet also a possibility that can be actualised and made concrete, such as a panning shot which reveals what before we couldn't see, or the sound source which is suddenly revealed as a record player owned by a character (as in Patrice Leconte's _Monsieur Hire_), and yet this obviously 'gives rise to a new unseen set, on to infinity' (MI:17). Later, Deleuze even notes that the voice may achieve its own 'distance' when the character talks like he is listening to his own voice, 'cutting it off from any direct resonance' (TI:242), resulting in a 'to-ing and fro-ing between speech and image' (TI:247), especially in the modern cinema.

But, also, there is an out-of-field that 'insists' from nowhere in particular that Deleuze elliptically notes, 'a more disturbing presence . . . outside homogeneous space and time' (MI:17), to be perhaps felt when the frame is as closed, and encloses as much, as possible, i.e., when hardly any out-of-field (for it cannot be totally eliminated) can be imagined. A viable example might be found in Paul Schrader's _Patty Hearst_, which, in line with cinematically conveying a sense of claustrophobia in the early scenes, neutralises any likely out-of-field for the frame, and also any psychological comfort for either us or the character, which thus *opens up* the image by channeling our concentration solely on to the in-field. The out-of-field 'disturbing presence' that insists itself can be thought as the Whole into which the frame is integrated, a Whole that we are thus forced to consider in being all that is left as excluded from the image.

Also, Deleuze gives some (small) consideration to those angles of framing whose point of view can be, or seem to be, 'bizarre or paradoxical' (MI:15); i.e., extreme low, high or generally skew-whiff. Surprisingly Deleuze allocates them an 'empty aestheticism' (MI:15), unless they are explained and validated by an initially unseen subject's point of view, and cites Lubitsch's _The Man I Killed_ wherein a mid-height pan behind some spectators drops down to frame a parade through the gap of a man's missing leg. The angle is then *justified* as such by the following shot of a man with no legs peering through the gap. Deleuze here stops too short with his image classification, interpretive or not, and only indicates a further dimension of the image in the concept (if it earns such a term) of 'deframing' [décadrage], usually of subjects, as in Dreyer's close cutting of faces in _The Passion of Joan of Arc_ (MI:15).

Furthermore, with the definition of the shot being a 'slice of space', the tracking shot is, according to Deleuze, most usefully considered as a *sequence* of shots, which forms 'a type of flowing and sliding shot' (MI:27), and has an interior montage. The panning, because it covers a number of possible individual shots, becomes under Deleuze's analysis a sequence of frames, with each successive frame being informed and altered by the preceding one.

Deleuze also finds that the *cinematographic* movement-image's essence lies in reflecting the movement on-screen, of bodies or vehicles, with camera movements; freeing the bodies or things on the screen by extracting their pure movement. Many examples can be thought of where the camera moves with a person, maybe moves away slightly in anticipation of their next move, and then rejoins them (one would be the extensive use of just that sort of tracking shot in Jim Jarmush's _Mystery Train_). What Deleuze seems to be saying is that the mobile camera always works as an equivalent to other forms of physical or mechanical movement, whether by foot, car, bicycle, plane, etc. This classification apparently ignores the possibilities of crane shots whose flow seem to have no general mechanical equivalent, as in the famous opening shot to Welles's _Touch of Evil_.

Deleuze also asserts that the space in the image can not only be a determined one, but possibly an any-space-whatever (espace quelconque), a rich space 'grasped as pure locus of the possible' (MI:109). In expressionism this was achieved with the use of shadows, endowing a depth to the image that metaphysically reaches further than regular depth-of-field in, say, Welles. The use of white and black can effect a lyrical abstraction, connoting power, impotence, and possibly uncertainty in between. Deleuze's any-space-whatever then reaches into 'the colour-space of colourism' (MI:117), in which colours are seen to absorb characters and situations (and thus the spectator). What Deleuze essentially asserts here is that there is not an effect *following* the appearance of colour (in the scene, or filtered over the camera) -- colour *is* the effect.

In Bernardo Bertolucci's _The Sheltering Sky_ the first half seems to be filtered with hot oranges and reds which does not merely *reflect* the male protagonist's passion and confusion, but effects to drive it, to be it. While, after his death, cool blueish tones take over that in a formally similar way move us, directly, to understand the journey of the female protagonist. And in Krzysztof Kieslowski's _A Short Film About Killing_ the lead up to the murder is toned in a sickly yellow-green that creates a resonance for the simple actions of the murderer-to-be, and murderee-to-be. It is this aspect that constitutes the pure potential of the any-space-whatever, its independence from the state of things within the narrative.

Editing and montage is the other way that the shot achieves mobility, but in a totally different way: in the assemblage of movement-images. As Deleuze writes, 'Montage is the operation which bears on the movement-images to release the whole from them' (MI:29), to free an indirect image of time. In Griffith Deleuze finds 'parallel' montage from which the narrative itself flows, striving to create 'a great organic unity' (MI:30), to link differentiated parts (rich and poor, etc.) by rhythmically succeeding one after another, and to alternate pairs of actions that will part and then converge

Deleuze's study of Soviet montage, mainly Sergei Eisenstein, informs as it essentialises. Deleuze finds the key to it in the quantitative, qualitative, intensive, and dynamic dialectical opposition of unequal parts. That the image can not only be juxtaposed, but also can itself be divided -- being both element and cell. He also brings out Eisenstein's activity of selection and co-ordination, of the importance of the interval, and of time flowing from the synthesis of images. That there is not only opposition but transition and upsurge of one image into another, realising that Eisenstein 'gives the dialectic a properly cinematographic meaning' (MI:37).

After considering the achievement of movement and the form of the image, Deleuze then questions the delineation of the two, arguing that movement, in organising 'the perceptive field' (MI:57), is at least a virtual image 'capable of 'drawing close to' the perceived and the perceiver, the world and perception' (MI:57). Thus in the same way the image, being the set of all that appears, can be seen to be identical with cinematic movement, just as light makes up matter. The line that Deleuze is following here is a purely philosophical one that leads him to relate, materially, the cinema to the physical universe, a 'machine assemblage of movement-images . . . the universe as cinema itself, a metacinema' (MI:59), with images being a consciousness that is immanent to matter. And further that the plane of matter has infinite intervals or gaps (e.g., between action and reaction) that define 'living' images or 'blocs [sic] of space time' (MI:61).

Deleuze seems to have derived this thesis of atomic perception from philosophies such as those of Leibniz and Spinoza, the former even saw the universe as made up from indivisible units called 'monads'. It is from this conception of the universe that Deleuze receives and forms the three types of movement-image: the perception-image, on this side of the interval; the action-image, on the other side; and the affection-image, which occupies the interval, without filling it up, and which is the final avatar of the movement-image. As Deleuze notes,

'each of these movement-images is a point of view on the whole of the film, a way of grasping this whole . . . each of these shots ceasing to be spacial in order to become itself a 'reading' of the whole film' (MI:70).

The perception-image is the perception of perception. In the cinema, if we are shown a person going into a room and looking around, and then a cut to a long-shot of the room as it would be from the position of the person, then this is the perception of perception, or the perception-image. It might seem that this is just another term for the familiar point-of-view shot, but Deleuze defines two types of perception-image: the subjectively justified 'point-of-view' shot as described above (direct discourse); and the anonymous viewpoint, where the camera seems to act like a disembodied eye, not quite objective, but semi-subjective (indirect discourse), yet which could always turn out to be truly subjective (e.g., in the horror film, the floating camera in the bushes at the back of the house suddenly spewing a hand that tries the back door). An assemblage of these two types of enunciation is a course of cinematics which Deleuze terms free indirect discourse, 'inasmuch as it testifies to a system which is always heterogeneous, far from equilibrium' (MI:73), towards an autonomous camera-consciousness that reflects its own content, that makes the camera *felt*. (This can be seen in the aesthetic consciousness of Antonioni, and the technicist consciousness of Godard).

What Deleuze is getting at here is the idea that there is a cinematic vision of content: in being able to see a viewpoint (subjective), and then the character and his world from a different point of view, the transition effects a transformation in the viewpoint of the character. This can be most prominent when the subjective image picks out something very specific that was not so obvious in the character/world shot. Additionally, Deleuze notes two states of cinematic perception-images: 'liquid', referring to flowing camera movements or the representation of fluid imagery such as water (the grace of the sea), clouds, or in fact anything that passes 'through or under the frame' (MI:80); and 'gaseous', derived from the free movement of molecules, in which there is a pure perception, and an interaction of images on all levels. Although this last state is virtually confined to experimental cinema: white on white, frames burning within frames, rapid montage, all acting together 'to form the cinema as machine assemblage of matter-images . . . [moving] from a real to a genetic definition of perception' (MI:85).

VII

Deleuze's classification of the affection-image relates to the close-up image, whether of a face crying or angry, or a clock, or indeed any object which is thus immediately 'faceified . . . there is no close-up *of* the face, the face is in itself close-up, the close-up is by itself face and both are affect, affection-image' (MI:88). What Deleuze means by this is that all close-ups give the image as much importance as if it were of a face -- for example, the cut to a clockface *or* a human face will have the same degree of (micro-political) significance if preceded by a large crowd scene -- they both hold not one thought (meaning) but many, by gathering and expressing the effect 'as a complex unity' (MI:103). Also, our sensation of space is virtually abolished, and that which is faceified achieves direct power in becoming an entity -- the human face of a coward *expresses* cowardice, the clock -- time, and so on. Deleuze cites Carl Dreyer's _The Passion of Joan of Arc_ as the perfect affective film, with its basic 'powers' of anger, passion, etc., and its elemental 'qualities' of martyrdom, and the victim. In fact if depth is introduced it 'indicates the obliteration of a character' (MI:107), and thus that a negation of depth can assimilate any length of shot to the plane of the affection-image.

Deleuze then considers a type of image that, he argues, cannot be made felt within the set of the affection-image: the impulse-image. Yet Deleuze does not explain exactly what sort of image we can allocate to the impulse-image set. It is based in naturalism and takes place not in an any-space-whatever but not quite in the determined space of a Hollywood action-image (see below), and, unlike the affection-image, 'embraces time, but only as the destiny of the impulse' (MI:127). Deleuze seems to be referring to those films that inhabit a world just beyond naturalism: possibly those of Spike Lee's _Do the Right Thing_, Pedro Almodovar's _Women on the Verge of a Nervous Breakdown_, and Peter Greenaway's _The Cook, the Thief, His Wife and Her Lover_, with the 'impulse' being the energy that ordinary things take on (like in the television series _Miami Vice_ where even the 'dirty' side of life is still softly and cleanly shot).

VIII

The action-image defines a structure (and is thus more like traditional theories of film) rather than any one particular image, and is the familiar place of 'determinate, geographical, historical and social space-times' (MI:141), of realism, behaviour, and actual milieux whose forces act on a character making him act to 'modify the milieu, or his relation with the milieu, with the situation, [or] with other characters' (MI:141), and so to attain a new situation. This is Deleuze's 'large form' of the action image: SAS', from situation to new situation via an action; and its variables, SAS: reactionary action and no new situation resulting; and SAS': a negative action producing a worse situation. Within the structure of the action-image, for an action to be an action it cannot be itself diffused by total montage, in that at least one single shot must contain and convey the 'duel' that creates the next (new, worse or stable) situation. This is thus behavioural or 'sensory-motor' cinema.

The small form of the action-image is a reversed sensory-motor schema, ASA', which goes from one action to a new course of action via an intermediary situation. (This form includes the reasoning-image, which corresponds to a gap in the narrative and implies something about that gap -- Deleuze cites, the shot of a man wearing a crumpled dinner jacket at a woman's apartment eating breakfast). Deleuze sees the economy of this form used in detective films where 'one moves from blind actions, as indices, to obscure situations' (MI:164), such that a film might take the course of ASA'SA''SA'''. . . etc. But, generally, what Deleuze attempts to get across is the fact that these two forms designate not just actions 'but conceptions, ways of conceiving and seeing a 'subject', a story or a script' (MI:178), and that they both can be played via a 'deforming form'.

We can see that the movement-image delivers an indirect representation or image of time, time in its empirical form, derived from the succession of shots. And it is this montage that relates time back to the progression of shots, and which makes it unable to directly represent it. At the end of the first volume of _Cinema_ Deleuze notes a crisis in the reigning form of the movement-image, the action-image. He finds this crisis, or change, coming about in Italy towards the end of the 1940's, in France at the end of the 1950's, and then in West Germany around the late 1960's. The change concerns the structure of time in film, with a proliferation of fractured voyages and any-space-whatevers, i.e., a general easing of sensory-motor connections. Where the movement-image released an indirect representation of time, flowing from chronological montage, the time-image subverts this smooth passage and so gives a *direct* image of time, shattering the sensory-motor schema. But we must not misunderstand Deleuze, the time-image was there, and in some cases was brought out (Ozu, Soviet cinema), and now the movement-image is certainly not dead.

The mental-image, being at the limit of the movement-image, is able to qualitatively transform the nature and status of the perception, affection, and action images. Deleuze sees in the mental-image the attempt to make the mental,

'the proper object of an image, a specific, explicit image, with its own figures . . . which takes as objects *of* thought, objects which have their existence outside thought, just as the objects of perception have their own existence outside perception. *It is an image which takes as its object, relations*, symbolic acts, intellectual feelings' (MI:198).

This is not to say that this image should try to represent some person's thoughts, but that the image should link directly with thought. (The literal relations of the humour of the Marx Brothers brought a type of mental-image to burlesque). What Deleuze really seems to be saying when he asserts that an image is defined as being a mental-image if it makes *relation* its object, is that it requires of us the act of *interpretation*, to work out what the image relates *to*. As an example Deleuze cites a moment in Hitchcock's _Rear Window_ when the camera explains why the hero has a broken leg -- we are given an image of him and a broken camera, and racing car photos. The mental-image is such that it embodies the two faces of the image, 'one turned towards the characters, the objects and the actions in movement, the other turned towards a whole which changes progressively as the film goes on' (MI:203).

In this second volume of _Cinema_ Keith Reader has found Deleuze working towards the 'reinscription of temporality across every element of film'. [5] According to Deleuze, this new image, this new mutation of cinema, strove to become 'truly thought and thinking, even if it had to become 'difficult' in order to do this' (MI:215). Deleuze starts his second volume off as he means to continue:&endash;

'What is in the present is what the image 'represents', but not the image itself, which . . . is the system of the relationships between its elements, that is, a set of relationships of time from which the variable present only flows' (TI:xii),

by which Deleuze indicates the malleable quality of 'time' in the cinema, where 'sheets of past [can] coexist in a non-chronological order' (TI:xii). Within this volume Deleuze considers a very personal band of 'creative' filmmakers, and their attempts 'to make perceptible, to make visible, relationships of time which cannot be seen in the represented object' (TI:xii).

According to Deleuze the viewer of the time-image 'is prey to a vision, pursued by it or pursuing it, rather than engaged in an action' (TI:3), as with the action-image. Deleuze terms the signs of the time-image *chronosigns*, and, being composed of pure optical and pure sound images (or sequences), these split into *opsigns* and *sonsigns*, which can have poles of objective/subjective, real/imaginary, and physical/mental. Opsigns are either (and Deleuze does not indicate how exclusive these forms are) 'constats': broad and deep images tending toward abstraction; or 'instats': near and flat involving images. But the real importance of Deleuze's new situations (images or sequences) is their power, the power of pure image and sound, to 'replace, obliterate and re-create the object itself . . . to give rise to a seeing function' (TI:12/19). Here Deleuze is at his most incisive in realising that the image can direct our attitude towards the object of the image, *in* the image, in effect subordinating any independent reference to the object. In this new cinema of time we will find, Deleuze writes,

'a camera-consciousness which would no longer be defined by the movements it is able to make, but by the mental connections it is able to enter into. And it becomes questioning, responding, objecting, provoking, theorematizing, hypothesizing, experimenting, in accordance with the open list of logical conjunctions ('or', 'therefore', 'if', 'because', 'actually', 'although . . .')' (TI:23).

What Deleuze is pointing to here is a readable image, a thinking image that is most locatable when the camera moves and works autonomously, striving to express such logical conjunctions as consequence, intention, etc., and no longer just measure, tonality, rhythm and harmony, as in the movement image. Recently, in Martin Scorcese's _Goodfellas_ the camera causes the movement of the characters through the whole of the film, beginning with soft lighting, and classically slow movements and pans in the heady days of the Mafia, and ending with fast and brash camera movements, and harsh editing and lighting -- *leading* the protagonists on the *film's* path. As Deleuze writes, 'it is a movement of world [the film's Whole] which supplements the faltering movement of the character' (TI:59).

Time is given as a perception, it depends on movement itself and belongs to it. Deleuze finds a temporalisation of the image in a variety of situations: the tracking shots of Resnais (especially in _Last Year in Marienbad_) and Visconti, in the depth of field that is seen in Welles's _Citizen Kane_, and yet, confusingly, also in Dreyer's 'crushing of depth and . . . planitude of the image' (TI:39). The shot here contains the force of time, and montage now constructs the relations of forces/time (the 'signaletic' material itself) in the succession of images. A significant situation in the time-image is the recollection-image (an actualisation of a virtuality), a fork in the life of the recollector, stretching as it does the time structure of the film, fragmenting its linearity. It does not give us the past, 'but only represents the former present that the past 'was'' (TI:54), this effect can be seen in the dream/recollection parts of Paul Cox's _Golden Braid_, these images quietly moving the narrative back and forth through virtual time spaces, and near the end of Krzysztof Kieslowski's _A Short Film About Love_ where the images become, almost imperceptibly, the thoughts of the young postal clerk, moving us back, but certainly not in the crude fashion of the flashback. The immediate-now presence of the cinema is perfect for the form of the recollection-image, as, 'it is in the present that we make a memory, in order to make use of it in the future when the present will be past' (TI:52). This image becomes, in Deleuze's eyes, part of a European subjective (read authorial, and to a less important extent reflexive) cinema 'uniting image, thought and camera' (TI:55).

As a development of these images Deleuze realises there to be two possible types of time-image, 'one grounded in the past, the other in the present' (TI:98). Here the past is preserved in time, and should not be seen as un-actualised recollection-images: 'it is the virtual element into which we penetrate to look for the 'pure recollection' which will become actual in a 'recollection-image'' (TI:98). Again Deleuze derives this cinematic feature from his own philosophic attitude, here concerning how we ourselves have a memory: 'Memory is not in us; it is we who move in a Being-memory, a world-memory' (TI:98). Thus via extrapolation Deleuze argues that in the cinema there are circles, or sheets of past, which coexist between past and present, each sheet having its own characteristics, *accents* (peaks of view), and *aspects* (regions, layers): 'its 'shining points' and its 'dominant' themes' (TI:99), e.g., childhood, death of mother, adolescence, adulthood, illness, etc. Cinema can place us in the past, and then choose one of these sheets. Such a conception of time Deleuze finds in _Citizen Kane_, 'the first great film of a cinema of time' (TI:99), where each witness is equal to a slice of Kane's life, bringing in a sheet of past. Thus, within the scheme of time-image films, depth of field becomes a 'figure of temporalization' (TI:110), ushering in (as well as the movement-image conception) a depth of time, by being literally big enough to let time show itself. Here in _Citizen Kane_ high and low angle shots deform time by contracting it, instituting montage *within* the shot, where that near and that far work on each other like shot cells in Eisensteinian montage. (Deleuze also sees Welles's low angle shots as referring to the inescapable contact with the earth -- but this should not be thought of as Deleuze's interpretation of *all* low angle images.) But time-image montage is found elsewhere in,

'the relation between the sequence shots or sheets of past and the short shots of passing presents; the relation of the sheets between themselves, each with the others . . . [and] the relation of the sheets to the contracted actual present which evokes them' (TI:111).

But when we are placed in a certain sheet of time, that time, whether it is past, present, or even future, *becomes* present. Thus Deleuze finds that, in a film such as _Last Year in Marienbad_, the 'past', 'present', and 'future' are no longer in any semblance of succession, but are simultaneously implicated. The three characters occupy separate, and different sheets of time -- which for us become the different *presents* of the film, each sheet can then be 'contradicted, obliterated, substituted, re-created, fork[ed] and return[ed]' (TI:101). Here, in peaks of *present*, the time-image shows its signifying power to push the ineffable, to make these once obviously *actual* places *de-actualized* in being related and aligned with virtual sheets of the past, for example, Resnais's tracking shots, always long and slow, defining and constructing continuums within his memory-ages. What are un-clearly 'flashbacks' in _Hiroshima mon amour_ Deleuze sees become more like recollection-images of an independent memory-world in _Last Year in Marienbad_, which leads Deleuze to surmise that Resnais 'attains a cinema, creates a cinema which has only one single character, Thought' (TI:122).

What Resnais achieves is a utilisation of the concept and the feel of the past without impoverishing it by simply making it the object of flashbacks (even those in _Hiroshima mon amour_). Thus the sketches of sheets of past can just as viably be false recollections or imaginations, or a past that is forgotten but that we have been given access to -- an opening up of 'past', and therefore 'time' in the cinema, using cinematic Thought. The feelings of the characters can then also spread themselves over different sheets of past, forming an interlocking sheet of their own. As Deleuze writes, 'Resnais goes beyond characters towards feelings, and beyond feelings towards the thought of which they are the characters' (TI:125). The players are thus the characters *of* thought, of cinema-thought, subservient to cinema-thought -- just as an actor is subservient to the author's words. In Godard's _Passion_ out of field spoken words early in the film turn up later spoken in person by the two lovers Isabelle (Huppert) and Jerzy (Radziwilowicz), and thus become an echo of an earlier echo of what was to come. From a grounding in space and movement the image has moved to take up topological schemas centering on time and thought, and for Deleuze the screen has become 'the cerebral membrane where immediate and direct confrontations take place between the past and the future . . . independent of any fixed point' (TI:125).

XII

Deleuze moves on to outline two 'description' regimes of the image: organic (kinetic), and crystalline (chronic). The organic regime we have seen in the discussion of the first volume of _Cinema_, and relates to realism. A crystalline description, encompassing de-actualized peaks of the present and virtual sheets of past, 'stands for its object, replaces it, both creates and erases it' (TI:126), and thus, unlike organic 'description', actually 'describes' the setting, being a 'cinema of the seer' (TI:126). We can see this where the camera sketches itself on to the scene, either by use of movement, tracing different possible angles of viewing the scene, or by some other technical device such as filters over the lens to give a different appearance to the setting or object. A crystalline image of people would not just show them in action, but 'carry out a primordial genesis of [them] in terms of a black, or a white, or a grey [or] . . . of colours' (TI:201), perhaps like the early work of Derek Jarman where extremely high-speed film is used that virtually makes faces and bodies alive with the grain of the film.

Deleuze calls these 'crystal-images' perhaps because their set-up, their internal structure is definite, with its symmetrically arranged faces being those of the actual image and the virtual image becoming one whole inseparable form. This form develops into crystalline narration, of pure optical and sound situations, with sequence takes over from sensory-motor montage, and wherein movement is 'zero' or 'incessant', such that 'the anomalies of movement become the essential point instead of being accidental or contingent' (TI:128). This is part of the scheme of the time-image, *from which* movement derives, which is then *false movement* produced by direct time-images, and thus non-chronological time (movement equals time, false time *therefore* false or abnormal movement). Jim Jarmush's _Mystery Train_ is a simple example of false/direct time, where the movements of the characters are made abnormal by interlocking at different times and places in the narrative.

Deleuze reckons this power of the 'false' to be the style of the new cinema, and, watching such films as those of Resnais and Godard, this would seem to be a viable argument. But we must also keep in mind that Deleuze is only referring to a select number of film-makers, and thus the application of these classifications to other cinemas is limited -- though certainly not untenable. We can even find this sort of falsifying narration entering into modern Hollywood, e.g. _Back to the Future_! Are the elements of Zemeckis's film not 'constantly changing with the relations of time into which they enter, and the terms with their connections' (TI:133). We might answer 'yes, but crudely'! But in Deleuze's eyes this new regime of the image works to end the presupposition of reality (which Zemeckis needs), stops narration vainly attempting to refer to a true, correct form (which neo-realism significantly renewed), and then replaces these with the (cinematographic) 'powers of life' (TI:135). Even semiotics cannot explain these neo-narrations of liberated time as there is no narration given by images, there being no underlying linguistic structure to them: 'Perceptible types cannot be replaced by the processes of language' (TI:137), is Deleuze's final word on the matter.

Beyond description and narration lies the story (récit), the development of the relationships of subject/object-images. According to convention, what the camera 'sees' is called objective, and what the character sees (shot of person cut to shot of approximate point of view) is called subjective. Yet in Deleuze's new poetical cinema of opsigns and chronosigns these are collapsed; neither are lost, but the camera's presence and vision enters into 'a relation of *simulation* ('mimesis') with the character's way of seeing' (TI:148). No longer is there the *two* films of traditional cinema, the direct one of the characters as they interact with objects and people, and the indirect one of the camera, surveying scenes and spaces of the characters' environment. Now the camera, whatever pyrotechnics it does, relates to the characters, expressing their vision or their fate. The original story gets enclosed in inverted commas to become something that can no longer refer to any true ideal.

As I understand it, this seems to be partly due to our consciousness of film and the camera, where, for example, the internal montage of the opening crane shot of _Touch of Evil_ *insists* a connection between the characters and the camera, to create one single internal vision. As Deleuze writes,

'Objective and subjective images lose their distinction, but also their identification, in favour of a new circuit where they are wholly replaced, or contaminate each other, or are decomposed and recomposed' (TI:149).

Thus in Godard's _Passion_ the image (op and son, image and sound), in echoing that to come, includes the before and the after, and also the personal and the impersonal in one present. This practice is thus how the direct time-image, breaking with indirect representation and the flow of time, feeds a 'free, indirect discourse, operating in reality' (TI:155).

XIII

It is in the last four chapters that Deleuze moves on from his classification of images, of cinematic movement, to a general consideration of the relationship between thought and film. Deleuze outlines three relationships between cinema and thought, to be found in the whole of the movement-image: cinema and a higher Whole (how we think about the Whole); cinema and thought, through the unfolding of images (that is image by image); and cinema and the relationship between world/nature and man/thought (i.e., from concept to image). Deleuze sees Eisenstein's decomposition of the 'shock' in cinema as 'the very form of communication of movement in images . . . from the image to thought, from the percept to the concept' (TI:157). Montage creates, or leads to a thinking of the Whole, via the effect of images on the cerebral cortex; we 'feel' the images, there is a total physiological sensation. Here Deleuze becomes prescriptive, arguing that the image must 'force thought to think itself as much as thinking the whole' (TI:158), i.e., the sort of cinema that Deleuze loves is just that type of Greenaway/Godard intellectual stimulant that has you unpicking a puzzle as you go along. For all this Deleuze still sees cinema as a primitive '*internal monologue*, a drunken monologue, working through figures, metonymies, synecdoches, metaphors, inversions, attractions . . .' (TI:159), a cinema of resonance but not linguistic palpability or certainty of expression.

In fact Deleuze even goes so far as to identify concept with image, and the image 'is for itself in the concept' (TI:161). Violence, for example, should be 'of the image and its vibrations' (TI:164), rather than that of the represented, while a concept such as 'grandeur' should be brought out by composition, and also not by any attempted representation. Deleuze increasingly attempts to posit thought as *the* object of the image, even though he says it only functions to repeat 'its own birth, secret and profound' (TI:165). This is in line with a film such as Godard's _Passion_, whose visual aesthetics come in at almost a zero degree, the (dis)organization of (abstract) images taking prime position. Thought, for Deleuze, becomes the 'higher purpose' (TI:168) of the essence of cinema, although he states that this does not hold for the majority of films! Our main question here is thus whether the 'thought' that Deleuze indicates is meant in a narrow (intellectual) or wide (intellectual *and* emotional) sense. As I understand Deleuze this question seems resolved when he asserts that cinema mustn't just film *the* world (narrow), but rather 'belief in *this* world' (wide); for, as in philosophy, we should relinquish our attempts to find absolute truths and 'replace the model of knowledge with belief' (TI:172). The attempt to show *belief* in this world is probably achieved by a type of the crystal-image discussed above.

Developing this image/thought conception Deleuze finds depth of field, say in Renoir and Welles, not to be merely metaphorically figurative, but more demanding, almost 'theorematic', the chain of images working more like a theorem, making thought immanent to what is seen. But here Deleuze appears to contradict himself in asking that we see a 'mathematical rigour . . . deductive and automatic' (TI:174) in the cinema, while earlier asserting that the image cannot be semiotically reduced. Yet, in his defence, it is not a rigour that Deleuze asks us to interpretively pick up on, but a rigour that Deleuze asks us to see the *possibility* of in such directors as Pasolini, and that we sometimes *do* actually see in the work of, say, DePalma and Terence Davies. In _Distant Voices, Still Lives_ the camera which tracks outside the terrace houses physically and immanently denotes a movement back in time when they proceed from right to left, with out of field sound pushing this time span in its historical specificity.

XIV

According to Deleuze what matters most now in this new cinema of time is not traditional montage, but 'the *interstice* between images . . . a spacing which means that each image is plucked from the void and falls back into it' (TI:179). This 'interstice' is something like a crevice, space, or chink between images, and therefore an operation of *differentiation of potential* that engenders something new. Deleuze is describing the choosing of images that break any join between them, a shock to get *us* choosing living thinking over dead thought. This is also achieved with adding a 'different' sound to an image, sound being the other frame of the image. What Deleuze is saying is that a film-maker such as Godard attempts to remove the reading of images by association (montage) which makes the succession of images into a chain, each slave in meaning to the next. The replacement is a division of this one-being mode of film for 'the method of *and*, 'this and then that'' (TI:180). This leads to irrational cuts engendering false continuity (in philosophy: phenomenology over structuralism) -- perhaps the cut from a person's face to a shot of their shoes, or any two shots where one 'has no more an end than the other has a beginning' (TI:181). Deleuze has come full circle since his discussion of the movement-image cohesion in montage of the first volume of _Cinema_. Here the image has itself become reflective, living, and where the image reaches its thinking limit thinkers themselves are brought in: e.g., Brice Parain in _Vivre sa vie_, and Jacques Derrida in _Ghost Dance_.

Deleuze then divides up his makers of the new cinema of the brain ('the cerebral process as object and motor of the cinema' (TI:210)) into the intellectual (Eisenstein, Resnais), and the more physical intellectual (Antonioni, Godard), and assigns two poles: brain (citing Stanley Kubrick) and body (noting Cassavetes). All this seems an unnecessary classification after those informative ones of the image. His distinctions are further undercut when he asserts that there is 'as much thought in the body as there is shock and violence in the brain' (TI:205). What Deleuze wants to say is that the mental functions as the true essential element of cinema, but that it is not 'abstract' (yet maybe geometrical, as in Peter Greenaway, especially _Drowning by Numbers_), as passion and feeling *are* part of the 'brain-world' (TI:210). Where Eisensteinian montage attempted to *mirror* the process of thought, this is now replaced by the irrational cut, and independent images, each deframing the previous, thus *creating and inducing* thought immediately in film and viewer: 'The outside or obverse of the images has replaced the whole, at the same time as the interstice or the cut has replaced association' (TI:214).

Thus Deleuze's new prevailing order of modern cinema (he attempts to attach it to neither 'modernism' or 'postmodernism') has irrationally cut sequences with a disjunctive value using normal, anomalous, or abstract (black screen to white screen) images. It has sound and image no longer in unison, nor merely non-attributable (as explicitly made clear in the deviant lip-synching in Godard's _Passion_), but belying and contradicting themselves, forcing there to be almost *two* films, *two* framings, intertwining in the one sound film, and separated by an interstice. As Deleuze so eloquently puts it, 'the visual and the sound are perspectives in a love story, to infinity, the same one and yet different' (TI:257).

Deleuze views cinema as being the highest exercise in thought, brought out from, and in a sense obeying the most basic images and sounds. The movement-image and the time-image are interactable in a linear sense, only that the former can not give us the latter. In conclusion the time-image gives us a direct, if decidedly wonky, representation of time in its pure form. This certainly does not mean the trashing of movement-images, but only subsumes them, making them *false* and subordinate *to* time. Here in the time-image those aspects of the movement-image gain a new sense (montage, frame, sound), take on a appearance, 'not something more beautiful, more profound, or more true, but something different' (TI:40), such that cinema becomes 'an analytic of the image' (TI:22).

In _Cinema_ there are actually two writers, the philosopher realising film in all its intricate workings, and the film lover, puncturing the text with loving synopses of the great and the obscure of film history, each having tentative reference to the argument at hand. At many (undiscussed) points Deleuze over-reaches his basic project, stretching his classifications into links with the society of the film at hand, or engaging in incisive, but old-fashioned thematic auteur studies (as with Welles's relation with Nietzsche in his questioning of truth). Yet these wanderings move us with their humility and student like adoration, and far from being slim in their concrete relevancy to his concepts, they in fact reinforce Deleuze's argument in an indirect way, willing us to attend to his classificatory remarks with a close and reasonable eye.

What do we get from Deleuze's 500 pages of ingenious rhetoric about cinema? We get what a lot of writers on film theory dare not say, that cinema is *not* a language that is in any way determinate, or determinable, but a phenomena, a visual material, an experience. That a truly cinematic film is just one whose sounds, words, and images effectively, and fruitfully interact. _Cinema_ is not a semiology of the cinema, as that replaces an image with an utterance, and so removes movement and time giving it a false appearance, and as Keith Reader says, 'segment-by-segment analyses rob [film's] object of its most essential quality'. [6] In Deleuze's eyes rather, film is 'a plastic mass, an a-signifying and a-syntaxic material . . . not formed linguistically' (TI:29).

_Cinema_ is piecemeal theory, in being derived from only certain types of cinema, mainly from North America, France, Germany, and Italy. Yet Deleuze's many categories for understanding the (whole) film could lead the way for a spawning of possible studies from a Deleuzian point of view. This conclusion is important for Deleuze's work, such that every element of his classifications is alive with potential. It would take a study ten times the size of this one to really bring out the possibilities of each of his film components, and apply them to cinema across the board. His classifications, although virtually neutral, *relight* the image for us, and make us look more closely at those films we thought we knew. As Deleuze writes,

'The soul of the cinema demands increasing thought, even if thought begins by undoing the system . . . on which cinema had fed up to that point' (MI:206).

Deleuze raises questions about what film can *essentially* do. He is not essentialist in the way of saying what cinema as a totality *is*. He virtually reinscribes cinema with its own philosophical apparatus (allowing him to ask 'not 'what is cinema?', but 'what is philosophy?' (TI:280)), which comes to the fore in the cinema of, for example, the recent work of Davies and Greenaway. If he sees a film where the camera moves over a subjectless setting Deleuze does not explicate a general meaning for the shot within the context of a film, but classifies it, in this case a 'pure description' within, it seems, a set of such cinematics called 'the empty observatory' (MI:134). Deleuze also totally ignores ideological forces at work in the making, and in our interpretation of cinema, and further gives no attention to the role of the spectator -- but this can only point out that he is not merely responding to these current film theory preoccupations, but pursuing an intuitive study of film. Deleuze's assigning of works to directors has also provoked criticism from some commentators, charging him with auteurism, but it seems obvious to me that the mere allocation of films to directors is quite different to 'auteur theory', and in any case, they are missing the importance of _Cinema_ in feebly calling our attention to this possibility. His mode of working is not by contextual analysis, but by metaphysical phenomenology -- finding, that is, absolutes of description. As Gaylyn Studlar has written, in one of the few commentaries on _Cinema_, Deleuze is interested, 'in the power of the image in phenomenological terms, in the cinema's ability to provoke wonder, awe, and astonishment'. [7]

What could be the implications for future interpretation if we see _Cinema_ as a true reading of the visual image? When faced with a film that we feel is great and beautiful the work of interpretation should not be to whittle it down to meanings and intention, but it should respond to the film in kind, extending, and carrying on its experience, to say or point to a way to most experience it. Through his deep and open inquisition into what *makes* cinema, this is what I feel Deleuze has given us the tools to do.

Polytechnic of North London, England

July 1991

Footnotes

1. _Cinema 1: The Movement-Image_, trans. Hugh Tomlinson and Barbara Habberjam (Minneapolis: University of Minnesota Press, 1986), hereafter 'MI'; and _Cinema 2: The Time-Image_, trans. Hugh Tomlinson and Robert Galeta (London: Athlone Press, 1989), hereafter 'TI'.

2. Keith Reader, 'The Scene of Action is Different', _Screen_, vol. 28 no. 3, Summer 1987, p.99.

3. Dale Jamieson and Barbara James, 'Review of Gilles Deleuze, _Cinema 1: The Movement-Image_', _Journal of Aesthetics and Art Criticism_, vol. 46 no. 3, Spring 1988, p. 437.

4. Reader, 'The Scene of Action is Different', p. 102.

5. Ibid.

6. Ibid., p. 99.

7. Gaylyn Studlar, 'De-Territorial Imperative', _Quarterly Review of Film and Video_, vol. 12 no. 3, 1990, p. 105.

Daniel Frampton, 'On Deleuze's _Cinema_', _Filmosophy_, July 1991 <http://www.filmosophy.org/articles/deleuze>.

Home

Comments & Questions: author@filmosophy.org

Discuss Filmosophy at Film-Philosophy