Gilles Deleuze’s Bergsonian Film Project by Donato Totaro

Gilles Deleuze’s Bergsonian Film Project by Donato Totaro

Henri Bergson

Both Gilles Deleuze and Henri Bergson were, to extremely varying degrees, philosophers interested in cinema who used cinema to suit their particular intellectual needs. In the case of Bergson, he cultivated his ideas during a zeitgeist that included the invention of cinema (late 19th century). To a large extent, Bergson’s philosophical ideas were shaped by the same cultural, economic, and technological climate that gave rise to narrative cinema. Deleuze on the other hand, erected a two-volume Bergsonian philosophy of cinema toward the end of the century that stands as one of the most stimulating studies of time and cinema. Although a self-professed Bergsonian, Deleuze’s sprawling philosophical style is in stark contrast to Bergson’s precise and systematic philosophical system. Deleuze’s postmodern style is part of its appeal -playful, mercurial, and open to creative interpretation. Terms that are meant to carry critical weight are introduced offhandedly and then left hanging for pages. One neologism gives birth to three others. In a sense, Deleuze’s style, forever Becoming, is more Bergsonian than Bergson.

I’ll begin with a brief, synoptic overview of Deleuze’s cinema project before moving on to a more detailed exposition. The broad sweep of Deleuze’s two cinema books, Cinema 1: The Movement-Image and Cinema 2: The Time-Image , is to chart a fundamental shift from classical pre-WW2 cinema [movement-image] to post-WW2 cinema [time-image] ( Cinema 1: The Movement-Image . Trans. Hugh Tomlinsom and Barbara Habberjam. 1983. London: The Athlone Press, 1986; Cinema 2: The Time-Image . trans. Hugh Tomlinson and Robert Galeta. London: The Athlone Press, 1989). The former cinema, which finds its archetype in the Hollywood genre film, is dependent on movement and action. Characters in the movement-image are placed in narrative positions where they routinely perceive things, react, and take action in a direct fashion to the events around them. The movement-image is a form of spatialized cinema: time determined and measured by movement. In the time-image, which finds its archetype in the European modernist or art film, characters find themselves in situations where they are unable to act and react in a direct, immediate way, leading to what Deleuze calls a breakdown in the sensor-motor system. The image cut off from sensory-motor links becomes “a pure optical and aural image,” and one that “comes into relation with a virtual image, a mental or mirror image” (Gilles Deleuze, “On the Movement-Image,” trans. by Martin Joughin, Negotiations: 1972-1990 (New York: Columbia University Press), 1995, 52).

In the time-image, rational or measurable temporal links between shots, the staple of the movement-image, gives way to “incommensurable,” non-rational links. Because of these non-rational links between shots, vacant and disconnected spaces begin to appear (“any-space-whatevers”). As a consequence, the journey becomes a privileged narrative form, with characters in a more passive role, and themes centered on inner mental imagery, flights of fancy, and emotional and psychic breakdown. The result of this pure optical and sound image is, according to Deleuze, a direct image of time (a time-image or crystal-image).

The Shot and the “Set”

In the opening chapters of Cinema 1: The Movement-Image Deleuze applies Bergsonian philosophy of time, change and movement to filmic construction. An important aspect of Bergsonian philosophy is that movement is distinct from space covered. Regardless of how much you divide space, movement will always occur in a concrete duration (indivisible time). ‘Real movement’ equals concrete duration. False movement occurs when you add abstract time to immobile sections. (In fact Bergson used the ‘cinematographical process’ as an example of false movement: immobile sections [individual still frames] plus abstract time [the projector].) Deleuze notes three Bergsonian aspects to movement and change: “1) sets or closed systems that are defined by discernible objects or distinct parts; 2) the movement of translation which is established between these objects and modifies their respective positions; 3) duration or the whole, a spiritual reality which constantly changes according to its own relations” ( Cinema 1 , 11). Movement is comprised of that which “happens between objects or parts…[and] that which expresses duration or the whole” ( Cinema 1 , 11). After the above exposition Deleuze writes, “Now we are equipped to understand the profound thesis of the first chapter of [Bergson's] Matter and Memory : 1) there are not only instantaneous images, that is, immobile sections of movement; 2) there are movement-images which are mobile sections of duration; 3) there are, finally, time-images, that is, duration-images, change-images…which are beyond movement itself….” (11).

Bergson’s own solutions to Zeno’s paradoxes will add clarity to the above exposition. In one of the paradoxes, Zeno offers the logical deduction that the tortoise, once with a lead, could never be surpassed by the much faster Achilles because each point along the way is infinitely divisible. Each advance Achilles makes is matched by the tortoise’s, with the space remaining between them infinitely divisible, ad infinitum. Bergson claims that this remains a paradox only when the movement, the race, is treated like the space and is divided into an infinite series of movements rather than the single movement that it is. Bergson anticipates how Achilles might explain the paradox: Achilles would simply describe the race as taking one step followed by a second, a third, and so on until he surpasses the slower stepping tortoise. When both movements are treated as indivisible wholes the paradox is removed. (Bergson solves Zeno’s paradox of the arrow in the same manner. You can not treat the object moving with the act of movement itself.) Applied to Deleuze’s breakdown above, we see that each static point along the race is 1 (sets or closed system or immobile section); each point relating to each other becomes 2 (“movement of translation” or “movement-images” or a “mobile section of duration”); and the indivisible race in its whole is understood as 3 (duration or time-image).

Deleuze then applies this to cinema. At the first level is “frame, set or closed system.” This includes all that occurs in the present image: sets, characters and props. “The closed system determined by the frame can be considered in relation to the data that it communicates to the spectators: it is ‘informatic’…” ( Cinema 1 ,18). A sets informatics can vary from “empty” (black or white frame) to “full.” The set is more specific than a shot because it can include sub-sets in the case of moving camera shots that reveal new information, or if changed internally: “…it determines an out-of-field, sometimes in the form of a larger set which extends it, sometimes in the form of a whole into which it is integrated.” At the second level is the shot and movement: “Movement expresses a change of the whole, an aspect of change, a duration or an articulation of duration” ( Cinema 1 ,18). “The shot is the movement-image. In so far as it relates movement to a whole which changes, it is the mobile section of a duration” ( Cinema 1 , 22).

At this point one may wonder what the difference is between Deleuze’s set and the conventional term shot. Deleuze, influenced by Jean Mitry, defines shot as follows: “The word ‘shot’ can be reserved for fixed spatial determinations, slices of space or distances in relation to the camera” ( Cinema 1 , 25). Hence in Deleuze’s terminology, the shot is subsumed by the set, creating more room for theoretical or analytical distinction. However, as with most Deleuzian neologisms, a vagueness clings to the term. For example, when exactly during a moving camera shot does one set become another? Or, how much new information is needed before one set changes to another? If not a precise term, it does add an important component to a psychological consideration of filmic time in the notion of changing ‘informatics’. The idea of continually changing sets raises the question of the relationship between the amount and type of information processed in a shot and the sense of perceived temporality. How is temporality affected by the amount of narrative and/or visual information (less information/stronger sense of time)? With respect to how time is perceived or felt, is there a difference in terms of the type of information given: visual vs. aural, sound vs. dialogue, color vs. black & white? In other words, does the type (and rate) of information given have an affective difference on aspects of time (for example, aesthetic, or psychological time).

In the movement-image, dominant in pre-World War 2 cinema, time and the image are subordinate to movement in all its forms. The movement does not give us a holistic time but a “mobile section of duration” (what Bergson would refer to as spatialized time). The early fascination with pure movement, such that we find in the myths of early film spectatorship (i.e. audiences phenomenally moving for fear of being run over in The Train Arriving at the Station), in pre-cinema optical toys, modern painting (The Futurists), and in communication technology, carried over into various aesthetic cutting designs in D.W. Griffith (parallel montage), the post-Revolution Soviet cinema (dialectical montage), French Impressionism (quantitative-psychic montage), and German Expressionism (intensive-spiritual montage).

This movement-image has two aspects, “one of which is oriented toward sets (frame or closed system) and their parts, the other towards the whole and its changes…” ( Cinema 1 , 55). If with think again of Zeno’s race, the first sense of movement is the changing state of the bodies in their varying positions. The second sense of movement is the relation of these varying positions to the whole race itself understood as a continuous movement from beginning to end. The movement-image fragments into three sub-forms, each dominated by a particular process: perception-image (the perceptual process), action-image (the narrative process), and affection-image (the expressive process). These three types, which open up to many other (less rigorous) sub-forms (limit-image, matter-images, reason-image, etc.) are found, to varying degrees, in all types of pre-WW2 classical cinema.

After Deleuze establishes his broad theoretical and terminological groundwork, a good two-thirds of the book is an at times fascinating, at times infuriating, impressionistic journey through (mainly) pre-WW 2 great “movement-image” auteurs (Griffith, Eisenstein, Gance, Grémillon, Vigo, Murnau, Lang, Renoir, Buñuel, Stroheim, Hawks, Bresson, Nicholas Ray, etc.). Deleuze’s journey is structured around some fascinating dialectical comparative analyses that are founded on varying subtleties within the impregnated movement-image. The best of these being Bresson vs. Dreyer within the context of affection-image (i.e. the close-up); Chaplin vs. Keaton under the action-image; and Kurosawa vs. Mizoguchi in a discussion of physical and metaphysical space.


Connected to the concept of sets is that of the “any-space-whatevers.” Here we have another example of Deleuze appropriating an existing term and transforming it into something entirely his own. Deleuze borrows the term from the French anthropologist Pascal Augé. Augé uses the term to help understand the effects of modern urban planning on the human psyche and interpersonal relations:

“An ‘any space whatsoever’ is a space such as a metro stop, a doctor’s waiting room, or an airport terminal. It is an anonymous space people pass through, … a point of transit between places of ‘importance’, such as the metro, which is merely the space one passes through between home and work. Moreover, in such spaces -and this is what interested the anthropologist Augé- individuals become depersonalized….It is for this reason that Augé argued that the ‘any space whatsoever’is a homogenous, de-singularizing space” (Jeffrey Bell, “Thinking with Cinema: Deleuze and Film Theory,” Film-Philosophy Electronic Salon , available at Online. Accessed 24 September, 1997.).

In the above referenced essay, Jeffrey Bell summarizes author Reda Bensmaia’s analysis of Deleuze’s transformation of Augé’s term. Bensmaia argues that Deleuze uses the term as a form of ‘conceptual persona’, in the way “…philosophers, artists, and scientists each…attempt to establish a sense of order to a fundamentally chaotic and forever changing world….In contrast to Augé, therefore, rather than being an homogenizing and de-singularizing force, Bensmaia shows that for Deleuze the ‘any space whatever’ is a condition for the emergence of uniqueness and singularities”. Deleuze does touch on elements of Augé’s initial usage, such as the train stations in Bresson’s Pickpocket , the airport in Marker’s La Jetée , the empty urban spaces in Antonioni, but then goes on to discuss it in a far more generalized sense:

“Any-space-whatever is not an abstract universal, in all times, in all places. It is a perfectly singular space, which has merely lost its homogeneity, that is, the principle of its metric relations or the connection of its own parts, so that the linkages can be made in an infinite number of ways. It is a space of virtual conjunction, grasped as pure locus of the possible”( Cinema 1 , 109).

Deleuze goes on to discuss any-space-whatever as it is shaped by some primary aspects of mise-en-scène: color, light, dark, white, black, shadows. It becomes an index of personal style and sensibility (‘conceptual persona’), evidenced for example by his comparison of Sternberg (aesthetic and passional), Dreyer (ethical), and Bresson (religious). Deleuze’s claim with space is much the same claim as Andrei Tarkovsky’s, that each director has their own aesthetic and personal sense of time.

Theory Into Practice

The first instance in Cinema 1: The Movement-Image of Deleuze applying theory to praxis comes with a discussion of the famous apartment dolly shot in Frenzy (1972). The conclusions are interesting, but the description is fraught with terrible inaccuracies. Deleuze describes the shot as such, “…the camera follows a man and a woman who climb a staircase and arrive at a door that the man opens; then the camera leaves them, and draws back in a single shot. It runs along the external wall of the apartment, comes back to the staircase that it descends backwards, coming out on to the pavement, and rises up the exterior up to the opaque window of the apartment seen from outside.” Before noting the descriptive errors, here are Deleuze’s observations based on his critical-philosophical terms. The gist of his analysis is that, as the camera moves the set changes (“movement in transition”), but this change is perceived as such only in relation to the whole (duration), which is the murder of a woman:

“This movement, which modifies the relative position of immobile sets, is only necessary if it expresses something in the course of happenings, a change in the whole which is itself transmitted through these modifications: the woman is being murdered…. What counts in these examples is that the shot, of whatever kind, has as it were two poles: in relation to the sets in space where it introduces relative modifications between elements or sub-sets; in relation to a whole whose absolute change in duration it expresses” ( Cinema 1 , 18).

Deleuze’s descriptive errors do not effect the overall theoretical (general) point, but unfortunately render a wholly different meaning to the scene. To begin with, the camera does not “follow” but leads them up the stairs with a dolly back movement. It does not “draw back in a single shot” but in two shots linked with a classic Hitchcockian hidden cut. When the camera dollies back out of the door leading to the pavement, a man carrying a sack of potatoes walks horizontally (right to left) in front of the camera and Hitchcock cuts on the sack of potatoes filling the frame. The most fatal descriptive error comes at the end, with the claim that the shot “rises up to the exterior up to the opaque window of the apartment….” In fact the camera dollies straight back across the street cutting through the busy horizontal street traffic to stop at the opposite pavement framing the entire apartment building and portions of the adjoining buildings. The point of the shot is to move from the very quiet, specific, singular murder (the killer’s apartment) to the noisy, general everyday (the whole apartment complex, street, adjoining buildings, people, etc.). It’s a perfect, succinct summation of the classic Hitchcockian theme of the horror in the everyday. Deleuze’s description infers that the camera brings us back to the specific, the killer’s window, which destroys the meaning of the shot!

Another problem with Deleuze’s description is the use of non-technical camera terminology (the camera, “follows” “runs” “comes back” “descends” “rises”) that does not render the feel and sensibility of the movements (which in many cases is extremely vital to understanding what the camera is attempting to “say”). The already difficult task of visualizing camera movement through the written word is best achieved using proper, accurate terminology (dolly, track, pan, crane, zoom, etc.).

Disregarding the above criticism, Deleuze’s Cinema 1: The Movement-Image can be fruitfully applied to film studies in a variety of productive ways. For example, one can use these Deleuzian terms to base an interesting case study on the different ways in which the long take is used in popular cinema and art cinema. I’ll provide an example by comparing the above shot from Frenzy to the penultimate shot in Michelangelo Antonioni’s The Passenger (1975). The two long takes are similar in several respects. In both cases the set changes dramatically from an inside to an outside space, while a murder occurs off screen. The Antonioni shot is much longer: 6.15 compared to 1.17. (As noted earlier, the Hitchcock long take contains a hidden cut at 57 seconds, but since its purpose is to render the illusion of continuity we can consider it as one shot of 1.17).

The purpose of the long take in Frenzy is to heighten drama and suspense through contrast and irony. With the killer uttering the words, “…you’re my type of woman” the woman’s fate is sealed. The camera pans right slowly along the exterior wall, negotiates a circular pan over the staircase and begins to dolly back down the steps, retracing their walk. Along the corridor we see a small table and hanging coat on the right and red carpet on the floor. As the camera nears closer to the front door the still silence gives way to increasing street noise. The camera continues past the door (the hidden cut noted above occurs here), the sidewalk, into busy midday street traffic (cars, people, fruit carts cutting the frame horizontally) and stops across the street to face the building block in full view. The suspense is not in whether the woman will die -that is certain- but in not seeing how. By leaving the murder to the audience’s imagination, Hitchcock is contrasting the representation of this murder to the Grand Guignol nature of the film’s first murder. The relation of the sets to the whole is altered, but differently than we’ll see in The Passenger . In this case the changed whole is complete, closed and resolved. The function of the long take is to shade and color the murder. The murder, as we’ll see, is inconsequential in The Passenger . What adds to the richness of Hitchcock’s long take is how it registers the film’s theme. The camera’s trajectory from the killer’s quiet, personal inside space to the noisy, communal, indifferent outside space reflects the theme of the horror hiding within the mundane, everyday, just as the serial killer hides behind the façade of petite-bourgeois respectability.

In this thematic reading, the long take is still linked directly to the murder, whereas in The Passenger the long take in itself carries a meaning beyond the murder. The Passenger stars Jack Nicholson as David Locke, a discontent reporter/documentary filmmaker who goes to Africa on a job mission. He comes across a dead man in an adjoining hotel room and inexplicably decides to change identities with him. The corpse is of David Robertson, a British man heavily involved in illegal arms trading with a radical political organization. Armed with the dead man’s diary, plane ticket, passport, and other personal belongings, Locke begins to assume his identity, literally trying to transcend his life by living another. In the famous penultimate long take, Locke decides to respect Robertson’s dangerous appointment with two shady men and waits stoically in a small Spanish village hotel.

To achieve the physically staggering effect of the camera movement Antonioni had a special gyroscopic crane built, named after its Canadian inventor, Wesscom, and took 11 days to film the shot. The shot’s framing begins with Locke, cut off at the waist, lying on the bed and in the middle background a large iron gate window looking out onto a sun-bleached village courtyard. Locke turns onto his side. The camera begins to slowly dolly forward toward the window, leaving Locke’s body below off-frame. On and offscreen sounds emanating from the courtyard are heard: voices, a car, a train. A white Peugot appears in the courtyard. Two men in suits exit the car. Children are seen throwing rocks (all figures are in long extreme long shot). The camera remains static, the window acting as a frame-within-a-frame. One of the men exits screen right. We hear the sound of a door being opened and closed, followed by footsteps. The camera reframes slightly to the right. The second man enters the frame and looks into the room/camera for a signal from the other man, and then turns to walk away. The camera restarts its forward movement toward the window. We hear a door open and shut again as the white car enters the frame and exits right. The camera has now tracked as close as physically possible to the prison-like iron gates of the window. Somehow the camera magically continues its forward movement, the iron gates disappearing from view. A police car enters the shot. The camera is now outside the room and slowly begins to pan around the courtyard, capturing the busy actions of the police and newly established central characters (Locke’s wife and girlfriend). The pan has now completed a 360-degree arch of the courtyard and is facing the window. Inside the room we see Locke’s wife and his girlfriend. The camera zooms in slightly and cranes to the right past an outside wall and a police officer to a second window that gives us a clear view of Locke’s body lying on the bed. The camera cranes right up to the iron gates of the window for a final view of the characters inside the room briefly discussing Locke’s death.

This long take, like the one in Frenzy , has no clear motivation or point of view. The implicit theme of The Passenger is one character’s search for identity or self-hood through some form of transcendence. Many interpretations, from the secular to the profane, have been given to this outstanding shot (room/window-as-body and camera-as-soul, or camera movement as the conclusion of Locke’s secular pilgrimage); but the camera’s spectacular “escape” from the room is itself a symbol of this transcendence. We know that Locke is dead, but do not know with certainty what his death means. Unlike Frenzy , the conclusion of the movement and the whole does not provide closure.

Deleuze’s appropriated term, any-space-whatever, can also be applied to the above discussion. For example, in reference to any-space-whatever Jeffrey Bell states that, “…Chris Marker uses airport terminals, public buildings, etc., as a means of undermining certain presuppositions one might have regarding the identity of character, plot, etc. Antonioni’s use of desert landscapes does much the same thing; in short, the ‘any space whatsoever’ functions in much the same manner that the time-image does: it places the identity of character, plot, etc., into crisis.” In Frenzy we may not see the murder committed but we are certain of its occurrence, as we are of the killer. The murder advances the plot and adds to the loathsomeness of the serial killer (and by extension the audiences emotional relationship to him). In The Passenger the long take, while much longer, is full of new information, but none of it adding to the crisis in character identity. The camera movement itself stands for the character’s final transcendence -he is killed as another person, David Robertson- but the shot gives us no insight into why David Locke felt this urge to transcend.


Leave a Reply