What is a shot? A first definition might proceed as follows: the shot, whether in still photography or cinema, is a record of what happened and is so essentially; there is no shot which is not a record of what happened. The shot is external to the event and only comes to exist in surviving the event, in remaining after it, as a testament to it. The shot is an index, a remainder, a ruin: an impression of what has taken place. Whatever fidelity between shot and event is supposed in this conception, it is always founded on the exteriority of one term to the other. The event is necessarily that which has happened, and the shot is necessarily external to the event.
The account of the shot given in Bazin’s ontology runs counter to this first thesis. It locates the being of the shot elsewhere. Not in its existence as a record of what happened, but in the ontological identity wherein the shot and the event become inseparable. What is thought in this idea of ontological identity is not immediately obvious – Daniel Morgan, in his recent path-breaking study of Bazin, calls it a “murky” idea – and as Morgan demonstrates, it’s an idea the English language commentary on Bazin has steered clear of: “It’s not so much that the idea of ontological identity has been considered and rejected as that it has been ignored.” The task for the following pages is to propose a reading of Bazin’s ontological identity, first in dialogue with Morgan, then via Jean-Louis Comolli’s essay on direct cinema and examples from Miklós Jancsó, Jean Rouch and Maurice Pialat. In doing so we’ll see how Bazin requires us to rethink the relationship between shot and event in cinema.
1. Ontological Identity
Bazin lays out his ontological identity thesis in ‘The Ontology of the Photographic Image’:
The photographic image is the object itself, the object freed from the conditions of time and space that govern it. No matter how fuzzy, distorted, or discoloured, no matter how lacking in documentary value the image may be, it shares, by virtue of the very process of its becoming, the being of the model of which it is the reproduction; it is the model.
It’s neither a theory of representation nor of correspondence. Bazin does not claim any resemblance between film and reality. He allows for the maximum possible distortion and the minimum possible documentary value (e.g. “photography will long remain the inferior of painting in the reproduction of colour” ); he sees the issue of visual resemblance as purely contingent and extrinsic. For Bazin, ontological identity guarantees no truth of representation, no precision of likeness.
Nor is Bazin’s idea of ontological identity explicable in terms of the photographic image’s existence as a physical imprint of the object it depicts. Peter Wollen’s influential reading of Bazin replaces the thought of ontological identity with the notion of an existential bond between image and object which, following Charles Sanders Peirce, Wollen names “indexical”: “an index is a sign by virtue of an existential bond between itself and its object.” Wollen cites some of Peirce’s examples: the rolling gait of the sailor, the weather cock. What these signs indicate – a life lived at sea, the direction of the wind – is something they have been physically touched by, something they are physically continuous with. The photographic image, conceived as a sign, is of the same order: it is the direct physical imprint of that which took place in front of the camera. That this is the case is not in question, but it is not what Bazin argues in his essay. As Morgan puts it: “No one argues that a footprint is a foot or that the barometer is the air pressure.” Wollen retains Bazin’s emphasis on the non-subjective nature of the image, its automatism, its automatic production without the intervention of man, but nowhere does he address the ontology of the photographic image. Wollen subscribes to our first thesis, outlined at the head of this essay, which conceives the shot in its existence and as coming into existence via a constitutive exteriority to the event. However, to address the question of the shot’s existence is not yet to engage the question of (its) being.
So how should we understand ontological identity? Ontological identity, Morgan argues, is the identity between objects in a photograph and objects in the world. He asks us to “recall Bazin’s claim that an object in a photograph is ontologically identical to the object in the world (however murky this idea may be).” Morgan presents this as a simple statement of denoted meaning rather than an interpretative act aimed at penetrating the murkiness of the idea, but in fact it already makes a decision on what might be the most fundamental ambiguity in Bazin’s essay, that is, the question of whether ontological identity is maintained independently between each object in the shot and the corresponding object in the world (which is how Morgan argues it) or whether it is only maintained within the shot as a whole. In other words, when Bazin says “the image is the model” we could wonder if it is the image of the tree in this photograph which is ontologically bonded to its model, or is it the whole ensemble of light, leaves, earth, air and branches, caught at one particular moment one late afternoon? When Bazin says “the image is the model” does he mean “the image of X which appears in this photograph is X”? Or does his reference to “the image” refer to the photographic image as a whole, rather than to the individual elements we could identify as being present in that photograph?
To manage this difference I want to propose a change in terminology, in which we would move from “the image and the model” to “the shot and the event”. When we speak of “the image” it could refer to the image of X in this photograph or the photograph as a whole, whereas “the shot” can only refer to the latter. Likewise, when we speak of the event rather than the model (or the object) it is better suited to naming that which encompasses the whole ensemble of light, leaves, earth, air and branches, all in movement, all in duration. Furthermore, when we conceive the event as the model it quickly overflows any limits or ground which would limit it to being an object of perception and extends out seamlessly to the movement and duration of the shot itself. The event isn’t an object which takes place in front of the camera; it isn’t an object fixed and depicted in the shot. Rather, the event is the event of the shot. It’s the event of the shot which encompasses, and it’s the event of the shot which constitutes the model. Ontological identity is thereby maintained between the shot and its own event, along with those events its event encompasses, or those events of which its event is a part.
Bazin perhaps seems to have understood ontological identity in the same manner as Morgan, or at least this is what is suggested by the passage in his essay in defence of Rossellini where he clearly conceives the image as determined in relation to an object in the world: “There is ontological identity between the object and its photographic image.” Nevertheless, against Morgan, and even against Bazin’s understanding of his own idea, I’m going to argue that the shift from “the image is the model” to “the shot is the event” is necessary if we are going to accomplish the thought of ontological identity.
The model was something which appeared in the world and the photograph; the event overwhelms the difference between the world and the photograph – it cannot be located in either, for insofar as it remains in occurrence it is that which encompasses the whole and therefore has no outside within which it could be located.
Ontological identity is achieved and maintained not between elements in the shot and the corresponding objects in the real world, but rather only in the shot as a whole, which also means the shot as an event in occurrence. This event, the photographic event, shouldn’t be considered as something which took place once, back then, when the shot was taken. Rather, the photographic event is an event maintained in occurrence without ever taking place. The shot maintains the event in occurrence; it is that event. Outside the event, outside the whole, the image and the model exist in the world as free-standing entities, one here, one there, existentially bonded (through the process which brought the photograph into existence) but ontologically distinct.
2. The Detour Through the Direct
In 1969, the year Peter Wollen pointed English language criticism down a different track, Jean-Louis Comolli published a Cahiers du cinéma article, ‘The Detour Through the Direct’, which directly addressed the question of ontological identity. At the core of that article is the following formula:
[The] fragment of reality, as soon as it is designated by the camera for filming, no longer equals itself, but itself plus camera. It is precisely that ‘dislocation’ which ensures that the image is not quite the model.
It might seem at first that in declaring this “not quite” Comolli is seeking to repudiate Bazin’s claim of ontological identity, as if to say: yes, the image is very close to the model but not quite identical, therefore Bazin’s declaration of ontological identity is premature and unfounded. However I think that far from being a rebuttal, Comolli’s “not quite” in fact works to affirm Bazin’s ontological identity thesis. Consider the following formula in Comolli:
Not only is the shooting of the film contemporaneous with the filmable event, but it is itself that event, which therefore films itself.
The second “is” in this sentence is the “is” of ontological identity. The event being filmed is the event of filming. The shot is the event. They are the same; the difference between them is a difference within the same. For Comolli the “is” and the “is not quite” are not in contradiction. The not quite – which is to say, the photographic event – overthrows the principle of identity; it uncouples being from identity. Rather than opposing Bazin’s thesis, Comolli in fact confirms and completes it. With the “not quite” he names the difference within which the ontological identity of shot and event is forged and maintained.
In the sentence prior to the declaration of the “not quite”, Comolli tells us that when we film something it “no longer equals itself”. It’s a remarkable claim: to say the event being filmed (“the fragment of reality”) no longer equals itself is to say the event filmed opens onto the event of filming rather than reflecting back on itself and forming an identity with itself. The event filmed and the event of filming become indivisible from each other. Each prevents the other from identifying with itself. The camera’s intervention undoes the object’s self-identity: together they become something else, an event which does not take place, a whole without identity or exteriority.
It’s true then that the shot is not quite identical to the event, but it’s also true that the event is not quite identical to itself. To film the event is to reach the event prior to suture, prior to the moment when the principle of identity is imposed. To film the event is to reach the event in its difference from the anecdotal identity of “what happened”.
Comolli’s essay is concerned with the “increasingly apparent recourse” in “fiction” films of the late 1960s “to the modes of direct cinema”, and with the tendency of direct cinema documentaries during the same period to “spill into the domain of fiction, with fictions they produce and organise”. He names L’amour fou (France 1968) and Faces (USA 1968) – the films with which Jacques Rivette and John Cassavetes set in motion (after a decade of preparation) the incredible projects they would pursue from film to film throughout the 1970s. He names also the first version of La rosière de pessac (France 1968), La collectionneuse (France 1967), L’enfance nue (France 1968), Silence and Cry (Hungary 1967), The Chronicle of Anna Magdalena Bach (West Germany/Italy 1968), La reprise du travail aux usines wonder (France 1968), and, more generally, “the films of Godard and Garrel”, “all of Rouch’s films” and “certain of Warhol’s”. Films which operate beyond any workable distinction between documentary and fiction. Films which don’t attempt to make the event of the shot disappear into a transparent view onto what happened. Films which play with the act of filmmaking as the instigator of what it films. Direct cinema is important to Comolli because it accepts that film is always the filming of its own event, and it approaches film as if it were an art of the event, much more so than an art of vision. With a direct cinema film like La reprise du travail aux usines wonder or Tourou et bitti (France 1971), it is not the fact that it happened, and that the film documents it, that matters most. Rather it is the occurrence that is the marvel. Not what happens, but its occurrence. These films allow us to recover a sense of occurrence distinct from what happened; they allow us to get around the forgetting of occurrence which takes place in everyday life by giving form to the event in such a way that its occurrence isn’t immediately lost in the image of What Happened. Direct cinema films reach the event in its occurrence; they open up that occurrence and work within it, such that they are no longer operating at the level of representations of what happened, but rather from the other side of the event, from within the occurrence. Comolli’s essay and the direct cinema he discusses in effect announce the end of that forced suture in which cinema ceaselessly abolished occurrence and resurrected it in the image of what happened. Oudart’s essay on suture (published semi-simultaneously with Comolli’s essay in Cahiers du cinéma in the spring of 1969) describes how the history of cinema, in such devices as the shot/reverse-shot, has succeeded in covering up the event of the shot so that it becomes purely transparent to the event in the shot. Suture, in a certain sense, is the operation through which cinema reinforces our first thesis. It insists that the event of the shot be denied and that the shot be conceived solely as a window onto events which have already taken place; it insists that the event be thought only with reference to what happens in it.
If the shot, as our first thesis supposes, was only a record of what happened, then it would be ontologically constrained to the recollection of events which had already taken place, and could not therefore work from within occurrence. It could not enter the inner sanctum of the event, but would instead limit itself to providing images of events already taken place, events seen only from the outside. The possibility which Comolli sees opening up in the films of Cassavetes, Rouch, and Rivette, amongst others, is that, to the contrary, cinema becomes capable of “reaching the event in the course of happening”. These films are defined less by offering images of the event (images which show what really happened) than by operating from within the event, from within that forge or factory in which the event is produced.
One example, Miklós Jancsó’s Elektra My Love (Hungary 1974), is staged in such a way that its long takes don’t present the singularity of the captured moment (the unique never-to-be-repeated event which took place in that moment) so much as reach that which, in the event, is disjunctive with what really happened. The film works in various ways to reframe its event so that it is not taken as a window onto some singular happening. The choreography – the columns of identically-dressed men and women performing uniform actions, the processions, the folk dances, the horsemen circling in the background – makes it clear from the outset that what’s happening here isn’t happening in its singularity. Jancsó stages the action not as if it were the event itself taking place, but rather as a kind of theatrical or festival performance designed to commemorate the event, and then situates the camera in the middle of the action, as one continuously moving part of a continuously moving ensemble, never able to find the limits of the performance space, never able to step over into a place from which it might simply observe.
The event in Jancsó’s film is not something which takes place upon a ground. The space does not precede the movements which take place in it; the movements of the camera and the columns of people construct a space which enters into oblivion as soon as the camera has moved on. There is no constancy of off-screen space – the same space will never appear twice – rather the event is the ceaseless forming and deforming of the space ‘in’ which it occurs, and as such it is not an event which takes place but an occurrence without exterior.
The emphasis on performance – the need that the event be a performance event – is common to most all of the films and filmmakers listed by Comolli, from Nico screaming in the desert (La cicatrice intérieure, 1971, Garrel) to Marceline walking through Les Halles (Chronique d’un été, 1961, Rouch/Morin) to the unknown young woman outside the Wonder factory whose distress we have to suspect is given further occasion by the camera. Jancsó’s Red Psalm (Hungary 1971) is exemplary in this respect. It presents us with a turn-of-the-century peasant strike, from the initial stages of the dispute, through the various arguments and counter-arguments (the interests of the landowners versus the interests of the workers; the workers urged to accept the reality of a market economy which nobody can control; the affirmation that the land belongs to whoever works it), through to the final massacre of the strikers at the end. Yet it dissolves the border between the event and its commemorative re-enactment. The action taking place can no more be identified as the shooting of a film than it can a peasant strike or a folk ritual commemorating and re-enacting a historical event. Which is to say, it is equally as much the shooting of a film, a theatre without audience or stage, and a conflict between workers and landowners. It is all of these and none. The strategy is grounded in Brecht (as per the Short Organum, “there are no illusions that the player is identical with the character and the performance with the actual event” ), yet it is not a matter of seeing through the fiction or illusion to recognise the true identity of the event as act of theatre, but rather instead a disappropriation of the event which suspends all identification.
Nothing allows us to see the event in Elektra as a recording of something exterior to it, or as a recording of something which took place. The event does not unfold for the camera and the camera does not observe the event from a position external to it. Gideon Bachmann’s on-set report indicates how this is of course entirely at odds with what really happened on the scene of production:
Despite the fact that ostensibly it is she, the camera, that observes what surrounds her and moves to do so, in reality every movement is being planned for her and every action exists only for her approval. Thus order is reversed: it is reality which is set in motion by deft manipulation in order to be at the right place at the right time. As soon as she has passed them, actors jump up, throw off a costume or don another, run ahead of her along her planned path, and crouch down again ready for another fleeting close-up.
Bachmann distinguishes between a camera which observes the world and a world produced for the camera, which is the standard distinction between transparency and creation. Instead the difference I want to highlight is between Bachmann’s idea of the camera as witness to an independent reality, which might or might not be created for it, and Comolli’s idea of the camera caught up in the event. Comolli speaks of cinema overcoming representation in a “reciprocal production”:
Through direct cinema the point is reached when the cinema is linked to life according to a system, which is not one of reproduction, but of reciprocal production, so that the film…. is simultaneously produced by and produces the events and situations.
Bachmann imagines a reversal in which the order of things is turned on its head, with the world now rushing around to arrange itself in such a way as to please the camera, as opposed to the camera seeking its objects in a world already sufficient unto itself. Thus whichever way he imagines it, there’s always a gap between the camera and the world it beholds. For Bachmann the camera is always withdrawn from the event, existing only as a witness to what happened. For Comolli there is no gap; the camera is indivisible from the event.
On the one hand, of necessity, Elektra is a record of what happened back then – as any shot must essentially be an imprint of those events which took place during its production: some people came together in a space, briefly, to stage some events, to perform, to make a film – yet on the other hand, it is a film which works from within the event, reaching that which is forever disjunctive with its anecdotalisation in the form of that which takes place. The event of the shot is not abolished in Elektra; it opens onto the event in the shot, reaching that event in its occurrence.
3. The Presence of the Camera
The presence of the camera influences the event that is happening in front of it. (Paul Henley, 350)
Earlier I grouped Tourou et bitti together with La reprise du travail aux usines wonder: two spontaneous films built around a long take, both set in the village square, where people assemble to await an event, or news of an event. If the 1968 film is about an event which hasn’t taken place – or which never finally takes place – an event betrayed, about the return to work, about the end of the exception (a strike settled without addressing the real conditions of the workers’ misery), then Rouch’s film is in a sense the opposite: the people are waiting for an event which hasn’t yet taken place; eventually though, in the absence of the event, the filmmakers decide to make a record of certain material facts – these drums, the tourou and the bitti, which are perhaps being played together for the last time – and when they start to film, then it happens.
The filmmakers had been invited to attend the fourth day of a spirit possession ceremony aimed at persuading the spirits of the bush to protect the harvest, and yet by the end of the afternoon, with the light beginning to fade, no possession had taken place. In the absence of the possession event the filmmakers opted to record the ceremonial event in its visible facticity. But then, during filming, Kuré the Hyena spirit took possession of one of the dancers, quickly followed by the spirit Hadyo possessing another. The coincidence between the filming and the possession leads Rouch to suspect it was the very act of filming that triggered the event, and as such his film offers itself as a test case for arguments around direct cinema, in that it offers an unusually pure distillation of the question of the filmmaker’s influence: did the filming trigger the event it records?
The question of the filmmaker’s influence on what he records has been asked since the beginnings of the direct cinema movement. The new sound and camera technologies allowed an unprecedented engagement with life as it happened. The filmmaker could now hope to be less intrusive, more self-effacing, and to thereby increase the objective quality of the image. At the same time others defined direct cinema not as improved transparency but as the camera’s participation in, and transformation of, whatever it films: Fereydoun Hoveyda, reviewing Chronique d’un été for Cahiers du cinéma in 1961, invokes Niels Bohr and the notion that “the observer disturbs the object being observed”. Noël Burch writing about the function of chance for Cahiers in 1967 (in one of the essays which would later be collected in his Theory of Film Practice) refers to the “Heisenbergian uncertainty principle”: the filmmaker “cannot aim his camera at anything without modifying it”. Rouch himself argues in 1978:
Most people refuse to recognise that any anthropology must destroy what it investigates. Even if you are making a long-distance observation of breast feeding, you disturb the mother and her infant, even if you don’t think so. The fundamental problem in all social science is that the facts are always distorted by the presence of the person who asks questions. You distort the answer simply by posing a question.
This debate – transparency or participation, Leacock or Rouch, category (f) or category (g) in Comolli and Narboni’s ‘Cinema/Ideology/Criticism’ – defines direct cinema equally as much as the technological developments which made it possible. Is the camera able to record life as lived? Or does it necessarily engage in that “reciprocal production” Comolli speaks of, in which the film “is simultaneously produced by and produces the events and situations” ?
Tourou et bitti is filmed (aside from a framing introduction) in a single ten-minute take. The camera approaches the village, comes in close to the action, witnesses the possession take hold, first in one of the dancers, then in another, and then (Rouch recognising that he is almost out of film) begins its retreat, drawing out to the edge of the village again. The possession, which hadn’t taken place after four days of trying, does take place when Rouch approaches with his camera. All the critical accounts of the film seem to agree that the filming participates in the event filmed; that the filming prompts the possession to happen. Rouch himself, writing two years after the film was made, states: “Looking back at this film now, I think that the shooting itself was what unlatched and sped up the possession process.” Paul Stoller agrees, emphasising that it was Rouch’s familiarity amongst the people he was filming which made it possible: “Had someone else been filming the sequence that day in Simiri, I am certain the mediums would not have been possessed. Such is the power of Rouch’s persona in Songhay.” William Rothman, paying close attention to the progression of action in the film as it moves towards the moment of possession, notes how Rouch moves the camera in close just before it happens: “it appears as if it is the very proximity of the camera that provokes the dancer to become entranced, to abandon himself so as to free the invisible spirit of Kuré to enter his body.” Paul Henley disagrees: it’s not the proximity of Rouch to the dancer which triggers the possession event, but rather a causal chain passing from cameraman to violin player to drummer to dancer:
In the middle of the shot, the drummers had given up and Rouch was about to turn off, but then, the godye violin player, seeing that Rouch was still shooting and assuming that this must be because he could see the spirits with his camera, began to play more energetically. This in turn encouraged the drummers to begin again, which then sent the principal medium into trance.
In this way all four writers – Rouch, Stoller, Rothman, Henley – recognise the filmmaker’s presence that afternoon as the catalyst for the possession, and each sees the shot as a record of, or window onto, that event of filming and possession. What happened, happened in the past, happened as a result of the camera’s presence, and the shot exists as a record of it. The shot is therefore external to the mystery; it merely serves as a window onto what took place back then on the afternoon when it was filmed. Rothman, Henley and the others are promoting a theory of the camera’s presence modelled solely on the general scientific notion of the observation disturbing the observed; nothing is said about the photographic event as such. These writers thereby remain within our first thesis outlined above; they efface the event of the shot in favour of the event in the shot; they conceive the shot not as an event but as an inert, inoccurent impression of an event.
We could therefore oppose two different approaches to Tourou et bitti. One which conceives the shot only as providing visible evidence of what took place (wherein what took place was influenced by the presence of the camera), thereby locating the shot outside of occurrence. And one which conceives the shot as the event. For the former, the event is exclusively identified with what happens in it, and the shot can do nothing but bear witness to that event. For the latter, the event is divided – occurrence is not quite identical to what happens in it – and the shot proves capable of working from within that difference, of reaching occurrence in its interiority, of reaching the event in the course of happening.
From within the interior of the event nothing can be confirmed to have taken place. It is only possible to declare that something has taken place by stepping outside of it. Occurrence itself has no outside. It disappears as soon as we stand outside of it. It only takes place through a suturing operation which identifies it with its anecdote: the image of what happened.
Occurrence is not reality. It is not what happened. It reaches something in the event which did not take place then. Thus: the event of the shot is an event which does not take place – it doesn’t take place on the scene of production; it cannot be located between the cries of ‘action’ and ‘cut’. What’s remarkable in Tourou et bitti is the way it maintains both the event in the shot (which happened that afternoon) and the event of the shot (which didn’t), each opening onto the other rather than establishing an identity with itself.
When Bazin (discussing Renoir) and Roud (discussing Godard) speak of a dialectic of abstraction and reality they are both, in effect, pointing to the fact that the event of the shot is not quite identical to the event in the shot. Morgan seizes on this as fundamental to Bazin’s thought and demonstrates how Bazin, no matter what specific film he was writing on, was always thinking in terms of this dialectic. Abstraction and reality are indivisible in the photographic event. A line cannot be drawn separating one from the other – the event of the shot cannot be separated out from the event in the shot because the event of the shot is the occurrence of the event in the shot. It is not possible to produce or even to imagine an abstract shot, wholly distinct from an event which would take place in it. The shot is a force of abstraction, certainly, but it remains categorically impossible to produce (or conceive) the event of the shot in isolation from the event in the shot. Nor is it possible to make a shot which frees itself from all abstraction and goes back purely to reality, not even in some simple, static, unobtrusive long take – this is what Comolli tells us, the reality filmed “no longer equals itself, but itself plus camera”; the shot is not quite identical to the reality which takes place in it. There’s always a force of abstraction at work.
The abstraction which renders the event of the shot – in Rouch’s film – not quite identical to the event which takes place in it should not be set to one side as an abstract view onto the reality of what happens (though this is the option taken by Henley, Rothman et al). Instead the abstraction should be understood as occurrence uncoupled from identity. It’s what we can admire in Tourou et bitti and La reprise du travail aux usines wonder: these permanent records of singular instances, these testaments to the camera being there to record what it also provokes; what we can admire in these films is the way what happens is rendered indivisible from an event which does not take place. It does not take place, it does not happen, and yet it suffuses – haunts – the shot from start to finish: the event of the shot.
Rouch’s opening voice-over is very beautiful and a powerful theory of the cinematic act in its own right: “To enter into a film is to plunge into the real, to be both present and invisible.” To plunge into the real is to plunge into occurrence, and to plunge into occurrence is to open a disjunction in the event wherein it is no longer quite identical to what happens in it. It’s what we’ve discovered already with Comolli: to film an event is to ensure that event is no longer identical to itself. On the one side there’s that which can be verified as having taken place, and then there’s the other side, which doesn’t take place, which can’t be verified, which doesn’t survive its occurrence, and it is into this that the film dives. And yet I’d argue that Rouch is not quite right when he says “present”. “Plunge” is much better: the film is the plunge rather than the point-of-view of the one who plunged. There is no one present.
When Rouch conceives the plunge into the real in terms of being both present and invisible he is trying to give expression to the particular qualities of the mobile handheld sequence shot. But the insistence on presence is a mistake: it reduces the film to a window onto what took place; it requires that the film is not itself an event but a witness to an event. If we’re to hold onto the film’s occurrence it’s essential that we do not appropriate it to the presence of a subject (e.g. camera or filmmaker) whose point of view it would then become. Any appropriation of this kind divides the event against itself and produces the kind of model described in our first thesis. What’s needed instead is a conception of the film event – the plunge into the real – which works without reference to the presence of the camera or the filmmaker. Rouch’s idea of the cine-trance seems to function in precisely this way. It conceives the moment of shooting not on the basis of the presence of the camera but on its absence. In the cine-trance the filmmaker disengages from what can be verified as happening on the scene of production and enters into a condition of absence or abstraction, becoming immersed in the shooting event in such a way that he or she accesses a dimension of the event which will not have taken place. Does the filmmaker remember their cine-trance? Do they later compare the shot with what they experienced in the trance of shooting it? Or is it like Céline and Julie Go Boating (France 1974), where the heroines have no memory of what they experienced in the house and watch the rushes together as if coming to discover these events for the first time?
4. Inside the Event
Life is no longer ‘represented’ by the cinema. The cinema is no longer the image – or the moral – of life. Together they speak to each other and produce each other within that speech. (Comolli, ‘The Detour Through the Direct’, 233)
In the climactic family argument from A nos amour (France 1983), there’s a moment towards the end, directly after the mother (Evelyne Ker) has slapped the father (played by the director Maurice Pialat), where the mother turns to her off-screen daughter, Suzanne (Sandrine Bonnaire), and summons her into the action.
The father is the first to indicate Suzanne. He is arguing with his wife and with his son Robert (Dominique Besnehard), who has intervened in her defence. The father turns to the son: “Go ahead. Defend your mother – who hit me. I hear you practice on your sister,” gesturing towards Suzanne, but without taking his eyes off Robert. With this, the mother then follows the father’s lead and turns to Suzanne, who is sitting outside the frame bottom-right, taking her up as her own concern. “Suzanne” she says, summoning her daughter, asking her to get involved, to stop watching from the sidelines. She directs Suzanne to come around and join with the rest of them, like a kid being called upon to join a family activity.
The mother’s gaze across the frame down towards Suzanne constructs the space in which this brief moment plays out. When she turns from her husband to look at Suzanne, she introduces a division between foreground and background: the argument between father and son is pushed into the background while the foreground is opened up as the empty space between mother and daughter, the space which awaits Suzanne’s response to her mother’s summons. The space is inseparable from the mother’s waiting; it was created by her turning and looking at Suzanne, and is maintained by her holding that gaze – flatly, impatiently – until the moment Suzanne stands up into the frame, which the mother marks by briefly closing her eyes, stubbing out her cigarette and turning back to the physical confrontation with her husband. The division between foreground and background is thus erased. The mother is fighting with the father; Susan has moved around the edge of the table to put herself between father and son. The four members of the family group have been brought together in the same space for the first and only time in the movie.
All this is achieved in a gesture which Ker borrows from Pialat. It’s achieved in the direction she issues – the summons ‘Suzanne’ – then the impatience, the waiting, the suspense, which allows the direction to resonate, which allows it to take possession of the event. ‘Suzanne’ is all she needs to say, or rather in saying it she suggests it’s all she needs to say, for the implication is that Suzanne should know very well what she needs to do. She says it in the manner of a parent reminding a child to do something they’ve been asked to do many times before, so many times they really shouldn’t need to be asked anymore (this confirmed by Suzanne’s reluctance when she stands and makes her way around the table). ‘Suzanne,’ then once she’s said it, she waits, or she puts on a show of restless waiting – hand poised with cigarette, mouth open slightly, then closed, then open and closed again – as she holds Suzanne in her gaze, waiting for her to respond, or suffering what she implies is yet another endless wait for Suzanne to do what she’s told.
That Evelyne Ker issues this direction following on from Pialat’s gesture towards Suzanne, suggests that gesture might have been the director’s cue to the actress. Or is it rather something shared between father and mother? He’ll deal with the boy if she handles the girl. Or could it be Ker acting under her own steam, taking her inspiration from Pialat’s gesture certainly, but deciding for herself that Suzanne should enter into the scene in order that all four of them should be brought together for a family reunion in the midst of this final argument? Is it her momentary inspiration that the picture will be incomplete without Suzanne and so she calls her around, as if the four of them are assembling for a portrait? Or is it the mother who, even now, continues to translate her dissatisfaction in her marriage into her resentment for her daughter? The work of mise en scène and the dynamics of the family are inseparable here. The mother directs; she summons Suzanne into the scene. The mother stages the scene from within the scene. The direction and arrangement of the shot, of the event, are precisely what is in occurrence within the shot. The scene is being directed live. Has Evelyne Ker taken it upon herself to bring the daughter into the scene, just as Pialat took it upon himself to introduce the father when nobody was expecting it? It’s the actress’s complicity with the director, to the extent that she takes over from him when he is otherwise engaged, which is also at the same time the mother and father’s deep-rooted familiarity in handling these family fights, so that even here, in the middle of the scene charting their most violent confrontation, there’s this brief moment when they fall back into the subterranean pattern they’ve been following for years, and we get a portrait of how this family operates, what kind of bond there is between them, born and carried in family quarrels which, for all their violence, also have a certain operational-functional effectiveness. The mother and father directing the family together. It’s how they’ve lived together. It’s how they’ve managed.
Pialat’s shot documents what happened in the presence of the camera on the day the cast and crew assembled to film this scene. It’s an actuality. It returns to that first great triumph of cinema: it grants us the ability to see again, to see for ourselves, what once took place. The workers leaving the factory. The train arriving at the station. The demolition of the wall. Events rendered in their anecdotal simplicity, including those events which derived their fascination from not being named in the official anecdote: “the beauty of moving wind in the trees, the little movement in a beautiful blowing on the blossoms in the trees”. Cinema records the occurrence of events not the presence of objects. When we see the mother impatiently measuring out the time of waiting in the opening and closing of her mouth we are not seized by a That-has-been. We do not have a sense of encountering a past. When watching this family fight are we captured by the film’s ability to maintain their being in a having-been-there? These people were there, in that room, in a unique photographic moment. And yet I do not think this is what impresses us. What the cinema gives us is the occurrence of the event rather than the having-taken-place or having-happened-there of the event. Cinema does not record the presence (having-been-there) of objects; it captures events.
Not every film reaches the event in its occurrence. Many for the most part aim only to work with the image as a fragment of what happened. They locate themselves after the event. They relate what happened. However, what we’ve seen in Jancsó, Rouch and Pialat is a group of filmmakers who work from inside the event. In each case the filmmaker plunges deep into occurrence. In each case we find the event generating itself out of itself and never reaching the limit at which it turns into what happened.
Bazin advocated a cinema which worked with events rather than the images of things which had taken place. He advocated a cinema which plunged the depths of the photographic event rather than providing views onto what happened. Pialat’s shot is exemplary of such a practice. The density of occurrence in this shot escapes description – the closer you look the more the action seems to proliferate – and it’s happening on both sides of the border, both inside and outside the fiction. It’s not just a matter of a lot being packed into a small space. It’s important that the action is also self-contradictory, that it resists being tied down in a unified statement of what happened. Bazin’s concern with the “ambiguity of reality” and with the mystery Rossellini preserves in the child’s face in Germany Year Zero (Italy/France/Germany 1947) is a concern for reaching the event prior to its being decided. It’s the same with Pialat: the kind of proliferation and density he discovers in the event is discovered precisely because he reaches the event prior to its identification in an image of what happened.
 Daniel Morgan, ‘Rethinking Bazin: Ontology and Realist Aesthetics’ Critical Inquiry 32:3, 2006, 471.
 Morgan, ‘Rethinking Bazin’, 450.
 André Bazin, ‘The Ontology of the Photographic Image’, What is Cinema?, vol. 1. Berkeley: University of California Press, 1967, 14.
 Bazin, ‘The Ontology of the Photographic Image’, 12.
 Peter Wollen, Signs and Meanings in the Cinema, 3rd edition, London: Secker and Warburg, 1972, 122.
 Morgan, ‘Rethinking Bazin’, 450.
 The brackets around the ‘its’ warn against conceiving being as if it were a possession of the shot. If the shot has being, then the shot must itself be located outside of being.
 Morgan, ‘Rethinking Bazin’, 471.
 André Bazin, ‘In Defence of Rossellini’, What Is Cinema? vol. 2, Berkeley: University of California Press, 1971, 98; my italics.
 Jean-Louis Comolli, ‘The Detour Through the Direct’ in Williams, Christopher (ed.) Realism and the Cinema, London: BFI, 1980, 242. First published in Cahiers du cinéma 209 and 211, February and April 1969.
 Comolli, ‘The Detour Through the Direct’, 238.
 In an essay on Max Ernst, André Breton asks: ‘Who knows whether we may not thus be preparing to escape one day from the principle of identity?’ (André Breton, ‘Max Ernst’ in Rosemont, Franklin (ed.) What is Surrealism?, London: Pathfinder, 1978, 8. See also Jean Goudal, ‘Surrealism and Cinema’ in Hammond (ed.) The Shadow and Its Shadow: Surrealist Writings on the Cinema, 3rd edition, San Francisco: City Lights, 2000, 93.) Bazin discusses the innate surrealism of cinema in his Ontology essay: ‘For [the surrealist], the logical distinction between what is imaginary and what is real tends to disappear. Every image is to be seen as an object and every object as an image. Hence photography ranks high in the order of surrealist creativity because it produces an image that is a reality of nature, namely a hallucination that is also a fact.’ (16) The surrealist dimension in direct cinema is evident in Rouch (see chapter 2 of Paul Henley, The Adventure of the Real: Jean Rouch and the Craft of Ethnographic Cinema, Chicago: Chicago University Press, 2009) and is clearly signalled in Rivette’s two monolithic films from the period: in the title of L’amour fou (1968) and the shop L’Angle du Hasard in Out 1 (1971).
 Comolli, ‘The Detour Through the Direct’, 225.
 Comolli, ‘The Detour Through the Direct’, 225.
 Jean-Pierre Oudart (1977) ‘Cinema and Suture’, Screen 18: 4, 1977, 35-48. First published in Cahiers du cinéma 211 and 212, April and May 1969.
 The idea of the event developed here – wherein what happens is not quite identical to its occurrence – has its foundation in Heidegger’s analysis of being-towards-death, Being and Time, chapters 46 to 53, and in the ‘Of Redemption’ chapter of Nietzsche’s Thus Spake Zarathustra. It’s Nietzsche and Heidegger who recognise the difference internal to the event, such that the event is not identical to itself: Heidegger in his analysis of the still-outstanding shows how the event is commonly conceived in its having taken place, even if it yet to come; Nietzsche in his account of the eternal return shows how the fragments and ruins of the ‘what happened’ are formally distinct from the affirmative occurrence of the ‘I willed it thus!’
 Deleuze uses this phrase when briefly touching upon Comolli’s article at the end of Cinema 1: ‘the cinema could not transcribe events which had already happened, but necessarily devoted itself to reaching the event in the course of happening’ (Gilles Deleuze, Cinema 1: The Movement-Image, London: Athlone Press, 1986, 206).
 Deleuze and Guattari’s idea of desiring-production (the unconscious imagined as a factory rather than the hidden operations in a theatre) is very much in line with Comolli’s thinking on direct cinema. What matters for Deleuze and Guattari is not product or producer, but ‘the production of production’. ‘Everything is production,’ they say: distribution and consumption are production; pre- and post-production, writing and editing, viewing and criticism are production. Following the same logic, Rivette rejects a ‘cinema partitioned off in tiers’: ‘all the stages,’ he says, ‘should be totally interacting’. The film in direct cinema is not the set of images produced, the film is the production: the production of the production. (References: Gilles Deleuze and Félix Guattari, Anti-Oedipus, London: Athlone Press, 1984, 7, 4. First published 1972. Jacques Rivette in Carlos Clarens and Edgardo Cozarinsky, ‘Jacques Rivette Interviewed’, Sight and Sound 43: 4, 1974, 195-6.)
 Bertolt Brecht, ‘A Short Organum for the Theatre’ in John Willett (ed.) Brecht on Theatre: The Development of an Aesthetic, London: Methuen, 1964, 195.
 Gideon Bachmann, ‘Jancsó Plain’, Sight and Sound 43:4, 1974, 220.
 Comolli, ‘The Detour Through the Direct’, 233.
 Fereydoun Hoveyda, ‘Cinéma vérité, or Fantastic Realism’ in Jim Hillier (ed.) Cahiers du cinéma: the 1960s, London: Routledge, 1986, 249. First published in Cahiers du cinéma 125, November 1961.
 Noël Burch, ‘Chance and its Functions’, Theory of Film Practice, London: Secker & Warburg, 1973, 121. First published as ‘Fonctions de l’alea’, Cahiers du cinéma 194, October 1967.
 Rouch, Jean with Dan Georgakas, Udayan Gupta, and Judy Janda, ‘The Politics of Visual Anthropology’ in Feld, Steven (ed.) Ciné-Ethnography, Minneapolis: University of Minnesota Press, 2003, 219-200. First published in Cineaste 1978, 8:4, 16-24.
 Jean-Louis Comolli and Jean Narboni, ‘Cinema/Ideology/Criticism’ in Bill Nichols (ed.) Movies and Methods, vol. 1, Berkeley: University of California Press, 1976. First published in Cahiers du cinéma 216, October 1969.
 Comolli, ‘The Detour Through the Direct’, 233.
 Jean Rouch, ‘On the Vicissitudes of the Self: The Possessed Dancer, the Magician, the Sorcerer, the Filmmaker, and the Ethnographer’, in Feld, Steven (ed.) Ciné-Ethnography, Minneapolis: University of Minnesota Press, 2003, 101.
 Paul Stoller, The Cinematic Griot: The Ethnography of Jean Rouch, Chicago: University of Chicago Press, 1992, 170.
 William Rothman, ‘Jean Rouch as Film Artist’ in Rothman (ed.) Three Documentary Filmmakers, Albany: State University of New York Press, 2009, 203.
 Paul Henley, The Adventure of the Real, 272.
 André Bazin, Jean Renoir, New York: Simon and Schuster, 1973, 84. Richard Roud, ‘Reality and Abstraction’, Godard, 2nd ed., London: Thames and Hudson, 1970, 71-99. Noël Burch introduces the same dialectic when he speaks of ‘the simultaneously abstract and concrete world of film’ in ‘Chance and its Functions’, 107.
 Though in Werckmeister Harmonies (2001) Béla Tarr does manage to sufficiently isolate the event of the shot from the event in the shot in order that the former occur as the void of the latter – though of course this voiding remains on the side of occurrence and never actually takes place. See Steven Marchant, ‘Nothing Counts: Shot and Event in Werckmeister Harmonies’, New Cinemas 7:2, 2009, 137-54.
 Comolli, ‘The Detour Through the Direct’, 242.
 This voice-over opens the film proper; it come at the very beginning of the long take which makes up the body of the film. It is located just after a brief prologue in which Rouch’s voice-over explains why and what he had come here to film, and after a title card stating ‘Un film de Jean Rouch’. The idea of Rouch as ‘a new type of filmmaker, the “filmmaker-diver,” who “plunges” into real-life situations’ comes from Edgar Morin’s article ‘For a New Cinéma-Vérité’, first published in January 1960, prior to his collaboration with Rouch on Chronique d’un été. In the version of Morin’s article republished in Steven Feld’s edited collection Cine-Ethnography, Rouch appends a footnote to the passage from Morin just quoted: ‘This image of the filmmaker-diver has always pleased (and flattered) me. The filmmaker with his equipment does indeed look like a deep-sea diver or like an interstellar voyager, but one who navigates in a “non-silent” world.’ (Edgar Morin, ‘For a New Cinéma-Vérité’, in Steven Feld (ed.) Ciné-Ethnography, Minneapolis: University of Minnesota Press, 2003, 230 and 264.)
 On the cine-trance see Rouch ‘On the Vicissitudes of the Self’, 99-100, Henley, The Adventure of the Real, 274-77 and Mick Eaton, Anthropology–Reality–Cinema: The Films of Jean Rouch, London: BFI, 1979, 50-52.
 Bazin conceives the camera in terms of the absence of man (‘All the arts are based on the presence of man, only photography derives an advantage from his absence’, ‘The Ontology of the Photographic Image’, 13). Is he then in agreement with Oudart’s ‘Cinema and Suture’, which conceives the shot on the basis of ‘the absent’? It depends whether we read Oudart’s absence as a dislocated presence, a presence elsewhere, namely the presence of the spectator, or as an absolute absence which is only identified with the spectator through a suture. If we read Oudart’s absence as absolute absence then he would be consistent with Bazin.
 Comolli recognises the same kind of operation in Jancsó’s films about captives and captors (such as Silence and Cry), where the captors’ shouted instructions to the captives (‘Come here. Stop. Go over there. Line up against the wall’) constitute an endless on-screen modulation of the mise en scène; they organise the event from inside the event. ‘Orders from the filmmaker to the camera and actors, and orders from a particular character to the rest have the same formulation and the same function – they arrange the shot in terms of action …. We are in fact looking at a film fiction being produced “live”.’ (Comolli, ‘The Detour Through the Direct’, 238-9)
 D.W. Griffith in Ezra Goodman, The Fifty-Year Decline and Fall of Hollywood, New York: Simon and Schuster, 1961, 11.
 Roland Barthes, Camera Lucida, London: Vintage, 1993, 77.
 André Bazin, ‘The Evolution of the Language of Cinema’, What is Cinema?, vol. 1. Berkeley: University of California Press, 1967, 37.