Challenging today’s ascendant digital aesthetic, this essay retraces one powerful line of French theory which treats film as an art which “discovers” significance rather than “constructs” meaning. Champions of today’s technology find that the digital at last permits complete control over image construction and therefore over “cinema effects.” Opposed to this aesthetic which targets the audience, the French aesthetic stemming from Roger Leenhardt and André Bazin concerns itself with the world the filmmaker engages. An interplay of presence and absence, as well as of human agency in the non-human environment, characterizes the French aesthetic at each phase of the filmic process: recording, composing and projecting. This article focuses on the central phase, composing, and on the terminological shift from “image” to “shot” picked up after Bazin by the Nouvelle Vague and passed forward to our own day through Serge Daney. In short, there is a Cahiers du cinéma line of thought, applied to questions of editing, which emphasizes the filtering implied in shots and the ellipses implied in their order. Conventional editors, on the other hand, manipulate or juxtapose images (using processes known as “compositing” today). The Cahiers line of thought developed in symbiosis with neo-realism and with a spate of post-war essay films of the “caméra-stylo” sort (Resnais, Franju) wherein editing works to cut away and filter out the inessential so that a mysterious or abstract subject can be felt as beginning to appear. Rivette, Rohmer and Godard have passed this line of thought on to a later generation represented by Philippe Garrel and a still later one for which Arnaud Desplechin stands as a good example.
Dans le contexte de l’avènement de la nouvelle esthétique du numérique, le présent article retourne à cet important principe de la théorie française voulant que le cinéma en tant qu’art permette de « découvrir » une signification plutôt que de « construire » du sens. Les tenants des nouvelles technologies considèrent que le numérique permet enfin une maîtrise parfaite de la construction de l’image, et conséquemment des « effets de cinéma ». À l’opposé de cette esthétique centrée sur le spectateur, l’esthétique française définie par Roger Leenhardt et André Bazin se préoccupe plutôt de l’univers dans lequel le cinéaste est engagé. Une telle esthétique se fonde d’abord et avant tout sur le jeu qui, à chacune des étapes de la production cinématographique (captation, composition et projection), s’articule entre présence et absence, entre facteur humain et environnement non humain. Le présent article insiste sur l’étape centrale, la composition, et sur le passage de la notion d’« image » à celle de « plan », selon la terminologie reprise de Bazin par les cinéastes de la Nouvelle Vague et relayée jusqu’à nos jours par Serge Daney. Il existerait ainsi, sur le montage, une ligne de pensée nettement identifiable aux Cahiers du cinéma, qui met l’accent sur le filtrage produit par la prise de vues et sur les ellipses produites par l’agencement des plans. Ce que les monteurs traditionnels manipuleraient et juxtaposeraient de leur côté, ce serait simplement des images (en utilisant des procédés comme le « compositing »). La ligne de pensée des Cahiers s’est développée en symbiose avec les productions néoréalistes et la myriade de films d’après-guerre répondant aux principes de la « caméra-stylo » (Resnais, Franju), dans lesquels le montage visait à éliminer l’accessoire, de manière à permettre l’émergence d’un sujet mystérieux ou abstrait. Rivette, Rohmer et Godard transmettront cette ligne de pensée à la génération qui a suivi avec, par exemple, un Philippe Garrel ou, plus récemment encore, un Arnaud Desplechin.
I. The Target of Film Theory
Traditional film scholars are on the defensive, for the “idea of cinema” is changing underneath us. Our students, hedging their bets on the future, compile bibliographies concerning “The Decay and Death of Cinema”; they draw lessons from Siegfried Zielinski’s (1999) sassily titled book Audiovisions: Cinema and Television as Entr’actes in History. Cinema’s acolytes must now be prepared for something completely different, their object of study being constitutionally sensitive to changes in technology and culture, far more so than, say, the novel. Still, I take cinema to be privileged within the spectrum of audio-visual phenomena. This polemical preamble hopes to clear some of the clutter left in the wake of the impact of the digital on film theory. I can then sketch a film aesthetic which owes nothing to the digital, though it can co-exist with and profit from new technologies. In fact, the digital is not really in question in this essay so much as the “discourse of the digital,” much of which would arrogantly de-centre or surpass mere cinema.
As a gesture neither of retreat nor of nostalgia, I keep the feature film in sight as the bull’s-eye of a target made up of a series of concentric rings. The movies that developed a solid shape after the First World War and reigned for seventy years as the world’s most popular and vibrant art form boldly stand out to be viewed and reviewed. The cinema surely exists within, or through, the feature film. Which other candidate might Zielinski identify as an “entr’acte in history” except the broad-shouldered feature movie that, in his view, has stood too long in the doorway, blocking other media? What else do critics have in mind when they say that cinema is in decay, if not feature films as we once knew and studied them?
Of course there have been other types of films exemplifying “ideas of cinema” quite different from that of the dominant feature. A surge in early cinema research over the past two decades has linked the invention of the medium to entertainment, scientific and even spiritual practices, with varied consequences for filmmaking and viewing. “The Cinema of Attractions” is an idea, to be sure, that brings together what we know of the uses of the technology, the practices of filmmakers and exhibitors, the discussions of reporters and cultural commentators and the protocols and laws established to regulate this new phenomenon. For better and for worse, the splendid variety embraced by this particular idea was gradually channelled into a normalized “Classical System.” An integrated industry of entertainment, the “studio system” also names a way of conceiving that industry’s polished product, the feature film, whether produced by one of the eight Hollywood studios or produced anywhere in the world after the fashion of the “movies as usual.” Of course plenty of “unusual” films competed with standard fare, often funded with political or aesthetic ideals in mind by independent producers and entire state systems. How should films be made? How should they look and sound? How should they function in society? From the twenties right up to the Nouvelle Vague, these questions found myriad answers in the brash or secret alternatives which veered away from a norm whose presence is nonetheless felt to be inevitable.
Throughout the heyday of the studio system, perhaps the strongest alternative ideas of cinema survived in non-narrative modes: animation, the documentary, the avant-garde, the short subject, as well as in educational, industrial and amateur film. All these modes, and the expansive ideas concerning cinema’s uses and powers that they put in play, should keep us from a myopic focus on the feature; they stake out territory in concentric circles at varying distance from the bull’s-eye of the feature which has demanded and received primary attention. These alternate modes force us to conceive a more comprehensive view of cinema as a whole. We need only recall that André Bazin, the prophet of Welles, Wyler, Renoir and Rossellini, felt equally compelled to promote animation (McLaren and Whitney), archival compilations (Paris 1900) and the weird scientific shorts of Jean Painlevé. Or take this very article as a case in point: an overriding idea of cinema will be established and traced as much through the experimental shorts of the early 1950s as through features. Still, it is the institutionalized critical legacy surrounding the feature film that has caused the most heated and robust debates in film theory, no doubt because of the social consequences of its ubiquity, its easy cross-over to the aesthetics of the novel and theatre and its ties to industry and to the global entertainment market.
Such debates about the feature fiction film, whether triggered by ideas coming from within that mode or challenged by modes that circle outside it, have made cinema studies among the liveliest sites in the humanities for the past half century. The prospect of the decline of those debates is more worrisome than the putative decay of their topic. For our seasoned ability to understand how the movies have functioned and to question how they came to function this way can guide the study of whatever “audiovisions” demand attention, whether they pre-exist the movies or are being born this new century. The fact is that hordes of amateur and professional scholars have not been able to avoid narrative cinema, because of its sheer quantitative bulk, its psycho-social effects and the ingenious efforts of those who sought to alter its course from within or without. Many of the best minds in the humanities turned from literary, philosophical, socio-cultural or historical pursuits to account for the most imposing medium of the twentieth century. They produced often complex, ingenious and passionate arguments and positions. They produced a way of thinking and they cultivated an instinct of looking and listening. Even if much of what has been written could be discarded without real loss, this discourse—this drive to understand the workings of the fiction film—is precious. To have this subsumed by some larger notion of the history of audio-visions, to have it dissipate into the foggy field of cultural studies, say, or become one testing ground for communication studies, would be to lose something whose value has always derived from the intensity and focus that narrative films invite and sometimes demand.
It was the emergence of the digital that encouraged Zielinski and others to upend the feature film—indeed upend cinema altogether—as the chief target of theory in the audiovisual sphere. Certainly the profession appears upended, at least momentarily, as questions of new media and digital processes sidelined or pre-empted other theoretical topics in film journals and at conferences. A new set of conceptions has arisen at every level, from production to spectatorship. Rather than support or decry millenary proclamations about the complete transformation of the media sphere, let’s use the occasion of cinema’s undeniable digital inflection to rethink the art’s past and its potential.
Today’s audiences imagine that filmmakers can completely structure audio-visual experience, encouraging the idea that movies have always been nothing other than a special effect, “the cinema effect,” as Sean Cubitt (2004) titles his ambitious book. This is certainly the view Lev Manovich (2001) proposes in Language of New Media. He forthrightly proposes that we treat films as instruments which serve two purposes: “To Lie and To Act” (Manovich 1998). Posed this way, cinema articulates perfectly with political and social history: to Manovich films have never been anything other than “machines of the visible” (Comolli 1985) deployed either to structure (or deconstruct) inevitably false representations or to engender direct audience responses such as outrage or submission.
I mean to advance quite a different idea of cinema, one that is in accord with the title of neither Cubitt’s nor Manovich’s text: cinema is not, or has not always been, a primarily special effects medium. The fiction films some of us most care about—and consider central to the enterprise of cinema in toto—have something quite other in mind than lying and agitating: they aim “to discover.” If anything is endangered by digital audio-visual culture, it is a taste for the voyage of discovery. Apparently many today feel that the world has been fully discovered and so now can be only manipulated and controlled to one purpose or another.
II. Film Theory in Three Dimensions
To see if the “discovery channel” is still open for film, I propose a voyage of my own, going through the territory of traditional film theory to look for routes to the present that have been partly abandoned or forgotten. How are we to map that territory? Whereas literary criticism can be bisected into ideas about texts on one side and ideas about reading on the other, film theory has tended to break into three zones of inquiry, corresponding to the three phases of the overall phenomenon: recording, composing, viewing. Each phase can be associated with a separate apparatus: the camera, the editing bench, and the projector. You’ll note that each of these machines has been updated or supplanted by digital technology. “Digital” is a word that connotes ultimate control, perfecting whatever operations its analogue or manual predecessors were designed to perform. The digital enhances, expands and alters those operations. This technological revolution nudges us to return to cinema’s fundamental operations to see if anything has been lost during the sweeping changes of the past two decades.
Each phase should be examined, but for the moment let’s set the camera aside, noting that it was the first machine to have been popularized and commercialized (and thus eyed askance) in today’s new regime. It was not only among theorists that a foreboding, even a panic, could be registered when images began to be generated without an imprint. Courts of law, for instance, suddenly had to reconsider the status of audio-visual evidence.  Yet the documentary film has never been more popular; so too are theoretical questions about the trace, visual memory and authenticity. Philip Rosen (2001) and Thomas Elsaesser (1998), for example, have deflated the apocalyptic rhetoric that accompanied the first digital cameras, arguing that in the main they serve the same “function” as did their analogue predecessors. Films have always provided ancillary cues or guarantees about the sources of their imagery and about how viewers should take them. With the new technology in everyone’s hands, sophisticated directors have learned to cross temporal or reality levels with even greater dexterity, as contemporary films such as Michael Haneke’s Caché and Werner Herzog’s Grizzly Man exemplify so well. Certainly the “ontology of the photographic image” has been altered, but the relevance of considering this ontology—including the return to centre stage of André Bazin—suggests that the recording phase of cinema, far from having been obviated, has re-emerged as a fecund site of thought and argument.
As they have with cameras, so have consumers bought into the value of digital projectors and monitors. The least studied phase of the film phenomenon, “viewing” goes in the opposite direction from the trace, aiming outward from the composed film towards its future in the world of audience perception and interpretation. Films are constructed with the capabilities and predilections of audiences in mind but, despite the Pavlovian will of producers, no one can fully dictate how these will operate. Digital projection devices, designed to improve quality and repeatability, have in fact put greater control of reception into the hands of consumers who handle it as they see fit, watching films when and how they like, on big screens or small. They are no longer tethered like Plato’s slaves staring straight ahead “On Screen, in Frame.”  I would argue that it has always been so; that interpretation always wanders beyond directorial control and into the larger world, the real historical world, where our feelings about films (or novels or other artworks) find their place or application. Digital screens allow audiences to enact this freedom at the primary site of viewing.
Thus the phases associated with camera and projector traffic with what lies beyond artistic or rhetorical control. In the first case, mechanical recording is open to the contingencies of the “captured moment” (Lastra 1997). While most of these can be kept at bay (especially in a studio situation), the very act of filtering implies a teeming, uncontrolled reality ready to clutter the record. The partly inhuman transfer (I’m tempted to say inscription or even communication) of available visual information onto celluloid and of audio onto some medium like wax or magnetic tape has goaded theorists from the very beginning. Such imprinting preserves traces of a past which filmmakers then fill with signification through narrative, editing and other compositional measures which Cubitt groups under the term “the cut.” But the cinema effect is ultimately psychological, occurring within the spectator, not on the cinematic medium, whatever it may be. And since projection is the name for a psychoanalytic syndrome as well as for a filmic process that involves physiology, the digital runs up against the contingencies of the future. Thus the film phenomenon begins and ends in zones that are constitutively indeterminate.
III. Composing and Editing an Essay on Film
It is the middle phase, that of “composition,” that is by definition highly determined, and it is this phase that has largely absorbed film theory. More remote and opaque to the general public than the camera or projector, the “editing bench” has in its turn nevertheless entered the common market, available now as software with names like “Final Cut Pro.” Composing a film today invariably means transferring all audiovisual inputs into digital information which is then manipulated up to the point of the final cut. Only a few in the industry complain about this undeniable technological progress and the speed and convenience it brings, not to mention the limitless options for correcting and enhancing the raw material. There remain intractable artisans who, thinking of editing as a form of sculpture, need to touch celluloid and measure the length of shots in meters rather than in time code. Many actors and directors of the old school rue the decomposition of scenes into discrete elements as the debasement of their profession, forgetting perhaps that “découpage” of one sort or another has always been involved. Speaking for those who share this concern, Jean-Pierre Geuens (2002) laments the shift of emphasis from shooting to post-production. When a director in classical filmmaking yelled “Quiet on the set,” he isolated the sacred place and holy moment of creativity in this art form. Today, however, noise on the set is filtered out; moreover, scores of soundtracks and extensive foley work build up an audio experience quite independent of what occurred on set. As for the set itself, shooting actors against green-screens can replace their face-to-face interaction and their bodily response to the physical layout of the scene.
Thus the careful “composition” of mise en scène has given way to “compositing,” which manipulates and layers a number of visual elements, only one of which is the actor’s performance in real time. It has been suggested that Martin Scorsese insisted on shooting Gangs of New York at the Cinecittà studios in Rome, knowing that this might be the last big-budget film shot entirely on set with all the actors present for their scenes.  He hoped to capture (or discover) the nuances of significance in the gestures of Daniel Day-Lewis, Leonardo DiCaprio and the rest of the cast as they played against each other in real space. Of course an artfully dressed stage makes up this “real space,” and the actors were able to find their gestures only after the scene had been worked over, often through scores of retakes. But the composition occurred right then and there, not later on at the computer. Actual sunlight, enhanced by reflectors of course, bounced off faces in real time. This meant a lot to Scorsese in his attempt to dig into his script and his actors. Things would be different in Lord of the Rings or King Kong, which are so densely composited they might be classified as animation.
Cubitt insists that long before the digital—indeed from the beginning—films have been built from elements cut together on a bench or in a lab. He argues that “The Cut” makes units and perspectival wholes out of mere visual energy; that it constitutes the control of the producer over the cinema effect and therefore is at the heart of a phenomenon which audiences have popularly conceived of as a “special effect,” that is, as marvellous.  Cubitt may not think so, but hasn’t “the cut” lost its edge in the new era? Originally an editor sliced a strip of celluloid at the frame line, whereas today that film frame, 35mm square, has gone fuzzy on the monitor. The editor stops the continual flow of lines at a given moment not quite a given place. Surely the work of such professionals has been transformed under this new dispensation. As for the general public, we are more baffled than ever by what occurs between the time of shooting and that of screening a movie. This has hardly changed across the century. In 1908, the popular weekly magazine L’illustration published an explication of various shooting and editing tricks to demystify cinematic magic for their huge readership (Anonymous 1908, pp. 203, 212 and passim). Almost thirty years later, in 1936, Roger Leenhardt (1988, p. 201) devoted an instalment of his impressive series “La petite école du spectateur” to explaining what happens at the “editing bench” to a more elite group who followed the journal Esprit:
If you pass from being a spectator to being a creator, from the screen to the editing table, you will find that a filmstrip is composed of a series of pieces spliced together in sequences, each of which has an exact length, which is suited to both its own expressiveness and its effect on the others [before and after]. There you see, in a precise sense, how there really is a cinematic metrics.
This sentence opens a discourse about the nature of filmic construction which depends on an idea of cinema different from the one that reigns today. Despite his implication that cinematic “metrics” is universal, and thus in principle predictable, Leenhardt insists that filmmakers achieve rhythm only by probing the temporality in their material as it runs up against a developing whole to which it contributes. In the mutual interaction of part and whole, they discover rhythm, rather than impose it. The aptness of such discovery is validated by every viewer who intuits the propriety of a cut. The editor organizes a single experience out of fragments by letting each moment stay on screen its proper length before giving way to its neighbour. “In English,” Leenhardt (1988, p. 203) reminds us, “découpage is called continuity,” and continuity is experienced as rhythm, “the control exercised by the mind over the material which has been filmed or is to be filmed.” The achievement of rhythm, as Augustine had understood and Bergson proclaimed, links matter to memory, letting the mind parse the world so as to bring out its significance.
Curiously, and crucially, because continuity derives from the space between shots, as rhythm derives from the interval between notes, Leenhardt (1988, p. 203) goes on to declare that “the essence of cinema [is] ellipsis.” Film editors (we call them “cutters” in English) create the illusion of significance and presence through the emptiness between and around their material. Of all cinema’s effects, this is surely the most special and the most specific. Ellipsis may be an optional trope for authors of prose fiction, but it “acts as the armature in the construction of a film.” The filmmaker operates with chunks of recorded material which he or she doesn’t so much sculpt (the Flaherty model) as organize in relation to an idea, phenomenon or event that arises in the emptiness between and around what is shown. Whereas the French Impressionists of the silent era had luxuriated in the fullness of the ethereal image (the image that radiated photogénie), and whereas Eisenstein asserted the primacy of muscular metaphor reached by imaginative leaps across the stark opposition of images or sounds and images, Leenhardt modestly points to the everyday workings of cinema through metonymy and ellipsis.
In flaunting the magnificence of the quotidian, Leenhardt launched what would become the “Cahiers du cinéma line of thought,” which Bazin solidified and the Nouvelle Vague took up, passing it to our own day through a critic such as Serge Daney and a director such as Arnaud Desplechin.  Jean Rouch, using an affectionate but accurate African expression, referred to Leenhardt as “l’ancêtre, l’ancêtre” (Leenhardt 1979, p. 76). And Bazin (1983, pp. 149-50) once claimed the “petite école du spectateur” to be the only worthy set of ideas he had read about sound cinema. For both men the existence of sound changed the essence of cinema. Leenhardt would never have come up with his views during the silent era. Sound de-sublimates the image, for it secures the picture to a definite spatio-temporal source (Leenhardt 1988, p. 201). Sound calls filmmakers back to the physicality more than the poetry of the audio-visual material on their editing benches, reminding them that they are cutting “shots” not “images,” a crucial substitution of terms.
Opposed to the “photo-effect” which calls attention to itself, I want [the spectator] to be sensitive to the qualities of truly good cinema photography, a bit neutral in appearance, a discreet servant that understands the spirit of the film.Leenhardt 1986, p. 45
Leenhardt (1979, p. 80) recounts his first meeting with his future cinematographer Philippe Agostini in a café, where he complained about the pretentious trend in film credits to read not “Cinématographie par…” but “Image de …” The cinematographer was not offended at all: “‘Images de Philippe Agostini.’ It’s ridiculous! If you ever make a film I’d be glad to work for you without trying to create ‘my’ image.”
This understated aesthetic was adopted by many of the most important post-First World War novelists and filmmakers who, perhaps because of the Great War and the ascendant role of journalism, learned the value of presenting situations as directly as possible. Facts are not to be interpreted so much as paired down and then assembled in a rhythm that gives them resonance. Leenhardt insists that the medium has matured, now that it includes ambient noise and dialogue to fill out scenes that formerly might have been enhanced through artistic angles, lighting, special effects and other recherché narrative devices. He argues that even without the addition of sound, 1930s films were more palpable than their predecessors because of the switch from orthochromatic to panchromatic stock, together with the more natural lighting this permitted (Leenhardt 1979, p. 80). The object had come into clear focus in its “density,” he says, and the basic unit of a film was unquestionably the shot cut from the volume of the world, not the “diaphanous image.”
For Leenhardt this was more than a matter of taste for hard-hitting style (a fad for Hemingway, Hammett, Dos Passos and James Cain); it was also a matter of profession, for as he details in his autobiography, he was working at the time as a newsreel editor for Éclair-journal. Daily he cut stories out of the hundreds of metres of film dumped on his editing bench. Daily he needed to find ways to present or suggest topics and events that were too large or too amorphous for an overview (Leenhardt 1979, pp. 68-69). Thus he took ellipsis not as an occasional rhetorical figure that accelerates or underscores a point; it stands as the key technique necessary for the very operation of the documentaire, that genre of film he was proud to practise. It was to screen his documentaries that Bazin invited him to his ciné-club at the Sorbonne in 1943 where they initiated discussions that would go on for years. You can sense Bazin in Leenhardt’s evocation of cinema’s “primordial realism.” Leenhardt (1988, p. 204) then goes further than Bazin might in highlighting his own work as an editor: “It is not in the cinematic material that art resides . . . but only in assemblage, rapprochement, ellipsis.”
This last term he latched onto well before Bazin, realizing how much film and prose fiction used it in their parallel projects to grasp the everyday character of modern life. And while American authors and cinema may have led the way, Leenhardt, like so many others in the 1930s, was a devotee of André Malraux, France’s most modern, journalistic author. Using the prestige gained from taking the 1934 Prix Goncourt with La condition humaine, a novel embodying this emerging aesthetics of speed and precision, Malraux backed other politically engaged writers, such as the controversial Andrée Viollis. In his preface to her anti-colonial reports from the field, S.O.S. Indochine, Malraux (1934) came up with a vocabulary that Leenhardt (1988, p. 204) latched onto in his fight against the aestheticism that still lingered from the silent age:
Malraux defined a new literary aesthetic which would rely on ellipsis in opposition to the ancient art of metaphor. This aesthetic is the aesthetic of cinema. It corresponds to the stage of precision which human information has achieved—photography is only one of its forms—and to a taste for the matter-of-fact, the document, which characterizes modern times. . . . Ultimately, it reveals a new method in the interpretation and expression of the world. . . . Not the studied search for a “meaning” through acting or décor but a simple work of “rendering.” Not an artistic exercise in expression but a technical effort of description.
This “effort of description” occurs when ellipses suggest the contour of a subject by omitting moments and aspects in a spatio-temporal volume. Through reduction, whatever is given on the screen must be bolstered by all that is absent; this can carry abundant associations (rapprochement). Bazin would come up with a substantially different view of ellipsis,  but he stands alongside Leendhardt in recognizing the primacy of what is not given on the screen. For despite what casual commentators have so often claimed, Bazin does not extol the cinema for rendering the visual world in its plenitude; rather he believes that it puts us into relation with a reality that we only partly glimpse on the screen. Unlike the photograph (the subject of the “ontology” essay) which adds something to reality, the cinema works largely by subtraction, forming reality’s negative imprint, as it were. Godard is after the same point when in Histoire(s) du cinéma he includes the anecdote about Ernst Opik, the astronomer who in 1932 calculated that over half the mass of the solar system has gone missing. From the movement of what could be seen by his instruments, Opik hypothesized the existence of a vast cloud (the Ort Cloud) . . . the unseen back side of the planets, so to speak. Godard and Bazin are haunted by this missing half, and by the structure of things that such virtual existence portends. Evidently cinema, this vaunted medium of the visible, is in fact an art of absence.
Leenhardt and Malraux led the way for Bazin’s more systematic ideas about film. But those ideas were forged even more by the appearance of striking documentaries responding to Second World War and its aftermath. Bazin may have been chiefly concerned with narrative cinema like most critics, but he could sense the entire institution challenged and altered by documentary. Hervé Joubert-Laurencin calculates this importance by noting that eight of the eighteen essays collected in the first volume of his collected writings (Ontologie et langage) concern documentary, whereas in his overall oeuvre they make up but four percent. 
Bazin was the first to intuit that Second World War had brought modern cinema into existence.  The most important films aimed to be responsive and responsible to a descriptive mission that the customary (classical) style was incapable of fulfilling. The world to be represented had become too vast, too rapid, too complex for standard cinematic representation. Thrust outside the studio, cinema struggled to grasp a confusing reality through portable cameras, special lenses, tape-recorders, fast stock, colour, even infrared. These inventions and improvements opened the way to techniques of aerial photography, night for night shooting, improvised mobility and location sound whereby cinema learned to command huge expanses of space and to represent phenomena and events previously out of range.
Less obvious, but I believe more significant, is the way documentaries manipulate the screen’s temporal dimension. After the war, thanks to experiences provided by documentary, feature films could attempt to shape fictions according to temporalities quite different from the ones laid out by the standard découpage system. Bazin dared filmmakers and audiences to open themselves to worlds of experience whose temporal coordinates are disturbingly and revealingly variable. The war had recruited cinema to capture the fundamental urgency of existence that was beyond the capacity of classical dramatic filmmaking. In his magisterial essay on neo-realism, Bazin (1971, p. 33) attributed the immediate power of “resistance films” to a cinematography exhibiting the qualities of “a Bell and Howell newsreel camera, a projection of hand and eye, almost a living part of the operator, instantly in tune with his awareness.” The very limitations of image quality produce a gain in realism in on-the-spot situations, rather more like an artist’s sketch than an oil painting. Bazin’s analysis of Malraux’ elliptical style complements the arguments he was already formulating against the supremacy of classical découpage, intimating that its nearly “incomprehensible” disjunctures force the spectator into an active mode of vigilance as in an emergency. Espoir, La bataille du rail and Paisà deprogrammed standard dramatic tempo, forcing the spectator to catch up with events that unfold or come into existence at an unpredictable rate. Each of these films takes place in the intermittent time of guerrilla action and reaction on a contested landscape. Their timescales can be reduced to the shortest imaginable: a man noticing a spider on the wall just as he is executed, a stray bullet hitting a partisan in Florence, the sudden recognition of social conditions by a black soldier in Rome. Such moments force the spectator to be alert to happenings whose causes are invisible, or too minute or oblique to be noticed. Such films must be grasped on the run.
Jarring ellipses bring out the speed and violent pace of life as lived and imagined after the Second World War, although they surrender visual context (including panoramic space), psychology and character interplay. All of this puts the viewer into a relation with the topic that can be called psychologically realistic. As for “total realism,” or rather its impossibility, Bazin (1971, p. 29) likened the situation to physiological limitation: the cones and rods of the optical system are sensitive to different visual domains; animals endowed with keen night vision see only in black and white, missing the information provided by colour. Bazin wanted to identify each film’s chosen place on the “realism spectrum” so as to watch it on its own terms, blind spots included.
At the other end of that spectrum was Georges Rouquier’s Farrebique (1945), a prize winner at the very first Cannes film festival. Ellipsis functions quite differently here. No contingency or happenstance deflects the fastidious depiction of a farmhouse in central France. The opposite of the combat film, its timescale is not that of the urgent instant but rather of the inexorable calendar year, and by implication, the time of the earth itself, the time that industrialization had lost track of. Rouquier wanted nothing to “date” his film (Weiss 1981, pp. 55-57), subtitling it “Les quatre saisons,” as opposed to such Rossellini titles as “Germania anno zero,” or “Europa 51”. Farrebique remains historical for all that (Sorlin 1988, p. 193).
Slowing things down, Rouquier enabled the land to be recognized in the literal sense of the term, including mud puddles, homely men and women, animals and their manure . . . all that standard cinema would tastefully skip over or leave out in its attention to drama and art. “Rouquier,” Bazin (1998, p. 106) said,
had understood that verisimilitude had slowly taken the place of truth, that reality had slowly dissolved into realism. So he painfully undertook to rediscover reality, to return it to the light of day, to retrieve it naked from the drowning pool of art.
Bazin is not naive. Even if Farrebique were to have jettisoned its cloying metaphors and “parasitical aestheticism,” it would not have directly captured the real. In one of his trademark formulations he says:
Thus, the most realistic of the arts shares the common lot. It cannot make reality entirely its own because reality must inevitably elude it at some point. An improved technique, skilfully applied, may narrow the holes of the net but one is compelled to choose between one kind of reality and another.Bazin 1971, pp. 24-25
In his essays on neo-realism, the net or filter would describe the relation of filmmaker to subject matter.
In every film some portion and type of information reaches us from a superabundance of the visible stream; whatever is on the screen has made it through the filter of the lens, so to speak. As in infrared photography, reduction can bring out the structure of the subject, or details otherwise difficult to discern. 
The filter may be considered a mechanism of ellipsis, for in its most basic cinematic application ellipses let our minds grasp something too extensive in space and time to be conveniently presented completely to our eyes. When Vigo relied on inventive ellipses for À propos de Nice, it was first of all because the city and its festival necessarily escape a single take. Although he shows us nothing beyond the city limits, Vigo’s selection and sequence of shots amount to a personal point of view: Nice filtered through his consciousness. The filter operates optically at the recording stage when a certain lens, film stock or literal filter regulates the kind and amount of light recorded by the camera, as well as when the filmmaker chooses particular angles, focal lengths and distances. Another kind of filter is at work on the editing table when one subject is pursued while others fall through the net. So when a series of already filtered shots are chosen to be joined into sequences, often 90% of the raw material and 50% of the selected shots pass through the moviola and onto the floor to be swept into the trash bin. The finished film, Bazin says again and again, puts us into contact with reality through what our eyes see concentrated on the screen. But don’t be fooled by Bazin, just as he tells us not to be fooled by appearances: what is on screen is not reality but its precipitate, its remains which, like the mummy he mentions on page one of “The Ontology of the Photographic Image,” conjures the presence of something richer; the image is the phantom of the fully real that hovers around, behind, or before the screen, the mummy who comes alive (Joubert-Laurencin 1998, p. 103).
IV. Tracking a Line of Thought
This “aesthetic of discovery” stands at the antipodes of a cinema of pure digital creation; it asks us to accommodate our vision to the conditions of visibility given by the world rather than, as in the aesthetics of new media, reworking the world until it conforms not only to our conditions of viewing but to our convenience and pleasure. The “cinema effect” aims straight at the spectator’s neurological makeup. It forgets the ethical stance Bazin applauded in neo-realism, an ethic that he believed made it modern by giving priority to what is outside the human. Hervé Joubert-Laurencin has convinced me that the shift to modern theory (i.e., a theoretical stance adequate to the new cinema of Renoir, Welles, Malraux, Rossellini and so on) can be located in a single sentence that Bazin composed in the 1945 version of the “Ontology” essay, later amending and softening it for the version collected in What is Cinema? Bazin’s first and radical formula is this: “Le cinéma apparaît comme l’achèvement dans le temps de l’objet, étroite photographie.”  The inescapable otherness of its photographic basis led Bazin to derive this key reversal that turns attention from human time to “the time of the object,” and so from the image to the shot. The entire line of thought from Leenhardt through Daney depends on this redefinition of cinema’s elemental makeup. Daney says right out: “. . . cinema is not made of images but of shots, and the shot is the indivisible bloc of image and time.”  In 1945 Bazin somehow grasped this radical idea. While the photographic document stands fixed, the filmed documentary presents its objects quivering in its own time, and suspended within a field of multiple determinations. As the film unrolls, any object’s integrity or identity can come into question. Ambiguity thickens the referent as a plethora of the object’s relations to its situation keep it from being completely “fixed.” This is how cinematic “composition” fulfills, indeed surpasses photography, taken as an art of the image…
Bazin wasn’t yet finished. He went further or, rather, he believed cinema was going further. In a well-known sentence, he said that Rossellini’s cinema is composed not of images nor of shots but of “facts” (Bazin 1971, p. 100).  Imagine a musical score comprising a series of facts instead of notes. This is how Jacques Rivette (1985, pp. 192-93) trumpets Viaggio in Italia, declaring it to be the first truly modernist film. In its most emblematic sequence, George Sanders and Ingrid Bergman are forced to look into themselves as they watch an archaeological dig at Pompeii. The bodies of two humans, surprised by death in an instant two millennia ago, gradually emerge into view, shocking the observing couple into self-recognition. They have stared at the facts and have trouble looking away. Rossellini would pass on to the Nouvelle Vague this ethos of digging underneath the stereotypes of plot, character, and action. “Things are there; why manipulate them?” he famously said (Hoveyda and Rivette 1985, p. 212), but getting at those things, getting at the facts, is another matter. In this case editing hollows out a space for the gradual or sudden appearance of a truth.
Two years later, Alain Resnais eliminated character and plot altogether in Nuit et brouillard, his effort to represent not just the holocaust, but our obligation to confront it. Shots taken during the operation of the camps amount to unavoidable facts which exist beyond human comprehension. The colour sequences signal a need to see and not to see, to uncover the truth and to go forward into the future. Resnais’ signature tracking shots run into the black hole of abomination. No one since Eisenstein has made so much of montage; but whereas the Russian juxtaposed image to image, Resnais opposes images taken in colour to facts given in black and white. The black-and-white footage is made up of “non-Images,” as Daney (1994, p. 256) was to call them.  The “givenness” of an unimaginable event derails the smoothly tracking camera, sinking it beneath the verdant landscape to what remains (the camp’s “remains”). What do we now do with them? “Indeed, what do we do?” asks the film in its conclusion. Bazin’s (1971, p. 35) remark about Viaggio in Italia applies more aptly to Nuit et brouillard: “Facts are facts, our imagination makes use of them, but they do not exist inherently for this purpose.”
Toute la mémoire du monde, made a few months later, elaborates the existential conundrum that doubles the shock that Nuit et brouillard’s pictures deliver: how can the present partake of the past; yet how can it avoid the past? The Bibliothèque nationale stands as a monument to this paradox. Here the hard facts of human history do not lie beneath an anonymous field; they are concentrated in a fortress of memory. Like the fussy employees of the Bibliothèque, Resnais treats each book or artefact—from anonymous periodical to renowned manuscript—as a fact in an insoluble mystery. The facts overwhelm us, even when fastidiously tended, organized, catalogued and guarded. As Roger Odin has put it, the film is composed of an opposition between actual and metaphorical space, between things and imagination. The accumulated weight of the past, sublimely immobile, is put into play by the incessant movement of human need and desire, figured in the camera that tracks the labyrinthine stacks (Odin 2005, pp. 78-79). As in Nuit et brouillard, two temporal orders are shoved together: fact and desire. In the sole dramatic sequence of an otherwise utterly descriptive work, a book is called for, retrieved, and brought out of its mausoleum to the guarded public reading room. Resnais signals the transit from one temporality to another as a chariot wheels up to the customs house of the circulation desk. Then, as it passes into the reading room, the monologue lifts itself in lyricism:
And now the book travels towards an ideal line, an equator more decisive for its existence than passing through a mirror. It is no longer the same book. A moment ago, it was part of a universal memory, abstract, indifferent . . . but now this one has been singled out, indispensable to its reader.
The book is grasped by the waiting hands of a human being (an intellectual worker) whose interest allows it to respire for an hour before being returned to the eternal memory bank of which it forms one unit or cell. With magisterial distance Resnais looks down on the reading room, stands above that search for “happiness” which keeps the scores of human bookworms (figured also as bees) busy as they burrow into their separate texts. Their seriousness and absorption—as they call for books, silently scribble notes, scratch their heads—makes them appear to be as unthinking as the books they devour in their quest.
Resnais’ two haunting essay films fulfill the hopes that, nearly a decade earlier, had driven Alexandre Astruc’s manifesto “Pour une nouvelle avant-garde: la caméra-stylo.” Astruc (1968) challenged filmmakers to address philosophical subjects through cinematic discourse. Resnais took up this challenge. His films can be taken as “compositions” in the literary sense; intriguingly they are composed of the camera’s confrontation with both a recalcitrant physical world and with an equally unalterable literary reality (texts by established authors such as Jean Cayrol, Remo Forlani, and soon Marguerite Duras). Even when commissioned, the freedom of the short film allowed Resnais, Marker, Franju, Varda and others to experiment with bold compositional strategies.
When given the chance to make features at the end of the 1950s, these same people helped usher the modern cinema onto the screen: Cléo de 5 à 7, Les yeux sans visage, Hiroshima mon amour.  This latter film by Resnais shatters the traditional homogeneity of the field of representation (the diegetic field) by layering distinct temporal regimes atop one another: as in Nuit et brouillard, footage drawn from the realm of facts contests the desires of humans to deal with both the past and the present (“You saw nothing in Hiroshima, nothing.”). Neither “the document” nor the “documentary” contains or can deliver the trauma of what is essentially the absence hollowed out of human history by the atomic blast. Resnais must have seen the Peace Museum which had just opened in the city and which occasioned the film in the first place as a monument to emptiness. In order to represent such emptiness, human beings with profound desires were needed, just as Resnais needed to picture those fanatical readers at the Bibliothèque nationale. Marguerite Duras supplied what Resnais required. She created not characters so much as embodied desires, two nameless people probing each other with the language of desire, probing two separate pasts onto which each opens like a doorway. Resnais’ camera tracks through these doors, discovering absence in the heart of passion, death in the midst of life.
V. Pursuits of Happiness; Pursuits of Theory
I have dwelt on the post-war documentary because it responded to and then greatly expanded a certain “idea of cinema” that grew into the modernism of the French Nouvelle Vague and influenced feature filmmaking the world over during the 1960s. With Leenhardt’s editing bench and Astruc’s caméra-stylo in mind, Godard was tempted to think of filmmaking as writing (“standing before the set as before a blank page” [in Milne 1972, p. 76]). Only later did he come to consider that he worked like a painter.  But even here his notion of painting would seem to answer the aesthetic challenge laid out in the title of Merleau-Ponty’s posthumous book, Le visible et l’invisible. It has nothing to do with cultivating the image for itself. The “Cahiers Line” that passed from Leenhardt to Bazin to the Nouvelle Vague was taken up zealously after the 1960s by Serge Daney, who towards the end of his too-brief life countered “le cinéma du look” at every turn. Daney championed probing filmmakers such as Philippe Garrel, who traffics in the unknown, the unforeseen. Today this “idea of cinema as discovery” is espoused most eloquently by Arnaud Desplechin, who speaks of “digging scenes,” by which he means getting beneath the ordinary to open up what is latent or hidden within a script, within a situation, and within the actors who enter into this voyage with him.  As Desplechin shows when he is at his best, a cinema of discovery and depth only begins with the shooting. Composing (not compositing) a film in the editing stage has ever been a “beau souci” (Godard 1956) whereby the interplay of seen and unseen makes a subject present. The final line of Toute la mémoire du monde seems spoken into the ear of the filmmaker busy in front of his bench or computer, where he laboriously “piece[s] together the fragments of that one secret that has, perhaps, a very beautiful name . . . and is called Happiness.”
I have tried to show that the search for such a “secret” was formulated as an aesthetic of discovery in France after the coming of sound, an idea of cinema that has persisted and developed through the Nouvelle Vague into our new century, even if it has been eclipsed by the idea of cinema as animation. Those filmmakers and spectators who take to the road of discovery must be prepared for the discomfort and the longueurs that every genuine journey entails. This includes journeys of thought. While it may be tempting in the age of the digital to step back to more general questions of media, communications, or culture, the arduous pursuit of ideas of cinema—digging into its multiform existence, its missions, accomplishments, and possibilities—rewards the effort. When disciplined, this exploration retains a very beautiful name. It is called film theory.
-  For a full discussion of this issue, see Schwartz 1999.
-  This is the title of an important essay by Stephen Heath that sums up much apparatus theory. It is collected in Heath 1981.
-  See for instance the review by Peter Travers in Rolling Stone (January 2, 2003).
-  Chapter three of Cubitt’s book is titled “Magical Film: the Cut” (Cubitt 2004). See especially p. 66.
-  See Joubert-Laurencin 2001.
-  In his long review of Malraux’ Espoir, Bazin (1981, pp. 146-54) questions the use of ellipsis in film.
-  Information presented at the conference “Les écrans documentaires” (Université Paris 7, November 1997).
-  Robert Kolker in The Altering Eye, Giorgio di Vicente in Cinema e modernità and Gilles Deleuze in Cinéma 2. L’image-temps would make the same claim in the era of academic film criticism.
-  A good example is in Bazin’s article “Defense of Rossellini” (1971, p. 98).
-  Joubert-Laurencin measures the difference between the 1944 and 1958 formulations in his “Document de synthèse,” prepared for his habilitation at the Université Paris I (December 2004, p. 10). Bazin’s softer 1958 version, with which we are all familiar, reads as follows: “Le cinéma apparaît comme l’achèvement dans le temps de l’objectivité photographique.”
-  Quoted in Hervé Joubert-Laurencin, “Document de synthèse” (p. 48), prepared for his habilitation at the Université Paris 1 (December 2004).
-  Bazin uses the term “facts” frequently in discussing neo-realism. See, for example, Bazin 1971 (pp. 35-77).
-  Hubert Damisch (2005, p. 76) cites this phrase in his dispute with Didi-Hubermann: “Montage of Disaster.” This essay has been translated by Sally Shafto and is available on the Cahiers web site.
-  See the “Avant-propos” and many of the essays in Blüher and Thomas 2005.
-  Sally Shafto has written a doctoral dissertation elaborating this; see Shafto 2000.
-  Arnaud Desplechin uses this expression in many interviews, for example with Jeff Reichart in Reverse Shot (www.reverseshot.com/legacy/summer05/desplechin.html).
Note sur le collaborateur
Il est professeur d’études cinématographiques et de littérature comparée à Yale University. Il a débuté sa carrière en publiant des articles sur la théorie du cinéma. Il est notamment l’auteur d’un ouvrage consacré à André Bazin. Il s’est aussi penché sur des films difficiles, tout particulièrement ceux de Mizoguchi, dont certaines analyses sont regroupées dans Film in the Aura of Art (1984). Il s’est ensuite intéressé au cinéma français et à la culture française, ce dont témoignent Mists of Regret (1995) et Popular Front Paris (2005). Il rédige actuellement Approaches to Problems in World Cinema ainsi qu’un texte destiné à What Cinema Is !, ouvrage collectif qui paraîtra dans la collection « Manifesto » des éditions Blackwell.
- Anonymous 1908: anonymous, L’illustration, 28 March 1908.
- Astruc 1968: Alexandre Astruc “Pour une nouvelle avant-garde: la caméra-stylo,” in Peter Graham, The New Wave: Critical Landmarks, Garden City/New York, Doubleday, 1968.
- Bazin 1971: André Bazin, What is Cinema?, Vol. 2, Berkeley, University of California Press, 1971.
- Bazin 1981: André Bazin, Cinema of the Occupation and Resistance, New York, Ungar Press, 1981.
- Bazin 1983: André Bazin, “Roger Leenhardt: les dernières vacances” , in Le cinéma français de la Libération à la Nouvelle Vague, Paris, Cahiers du cinéma, 1983.
- Bazin 1998: André Bazin, “Farrebique or the Paradox of Realism” , in Bert Cardullo (ed.), Bazin at Work, London, Routledge, 1998.
- Bluher and Thomas 2005: Dominique Bluher and François Thomas (eds.), Le court métrage français de 1945 à 1968, Rennes, Presses universitaires de Rennes, 2005.
- Comolli 1985: Jean-Louis Comolli, “Machines of the Visible,” in Gerald Mast and Marshall Cohen (eds.), Film Theory and Criticism, Oxford, Oxford University Press, 1985, pp. 741-760.
- Cubitt 2004: Sean Cubitt, The Cinema Effect, Cambridge, MIT Press, 2004.
- Damisch 2005: Hubert Damisch, “Montage of Disaster,” Cahiers du cinéma, no. 599, 2005.
- Daney 1994: Serge Daney, Persévérance, Paris, P.O.L., 1994.
- Elsaesser 1998: Thomas Elsaesser, “Digital Cinema: Delivery, Event, Time,” in Thomas Elsaesser and Kay Hoffmann (eds.), Cinema Futures: Cain, Abel or cable? The Screen Arts in the Digital Age, Amsterdam, Amsterdam University Press, 1998, pp. 201-22.
- Geuens 2002: Jean-Pierre Geuens, “The Digital World Picture,” Film Quarterly, Vol. 55, no. 4, 2002, pp 19-30.
- Godard 1956: Jean-Luc Godard, “Montage, mon beau souci,” Cahiers du cinéma, no. 65, 1956.
- Heath 1981: Stephen Heath, Questions of Cinema, Bloomington, Indiana University Press, 1981.
- Hoveyda and Rivette 1985: Fereydoun Hoveyda and Jacques Rivette, “Entretien avec Roberto Rossellini” , in Jim Hillier (ed.), Cahiers du cinéma: The 1950s, Cambridge, Harvard University Press, 1985.
- Joubert-Laurencin 1998: Joubert-Laurencin, “Tombeau d’André Bazin,” Imagens, no. 8, 1998.
- Joubert-Laurencin 2001: Hervé Joubert-Laurencin, “AB/SD, ou les frères passeurs,” Trafic, no. 37, 2001, pp. 107-16.
- Lastra 1997: James Lastra, “From the Captured Moment to the Cinematic Image: A Transformation in Pictorial Order,” in Dudley Andrew (ed.), The Image in Dispute: Art and Cinema in the Age of Photography, Austin, University of Texas Press, 1997, pp. 263-91.
- Leenhardt 1979: Roger Leenhardt, Les yeux ouverts, Paris, Seuil, 1979.
- Leenhardt 1986: Roger Leenhardt, “La Photo,” Chroniques du cinéma, Paris, Éditions de l’étoile, 1986.
- Leenhardt 1988: Roger Leenhardt, “Cinematic Rhythm” , in Richard Abel, French Film Theory and Criticism, vol. 2, Princeton, Princeton University Press, 1988.
- Malraux 1934: André Malraux, “Préface,” in Andrée Viollis, S.O.S. Indochine, Paris, Seuil, 1934.
- Manovich 1998: Lev Manovich, “To Lie and to Act: Cinema and Telepresence,” in Thomas Elsaesser and Kay Hoffmann (eds.), Cinema Futures: Cain, Abel or Cable? The Screen Arts in the Digital Age, Amsterdam, Amsterdam University Press, 1998, pp. 189-99.
- Manovich 2001: Lev Manovich, Language of New Media, Cambridge, MIT Press, 2001.
- Milne 1972: Tom Milne (ed.), Godard on Godard, New York, Viking, 1972.
- Odin 2005: Roger Odin, “L’entrée du spectateur dans le documentaire,” in Dominique Bluher and François Thomas (eds.), Le court métrage français de 1945 à 1968, Rennes, Presses universitaires de Rennes, 2005.
- Rivette 1985: Jacques Rivette, “Letter on Rossellini”  in Jim Hillier (ed.), Cahiers du cinéma: The 1950s, Cambridge, Harvard University Press, 1985.
- Rosen 2001: Philip Rosen, “Old and New: Image, Indexicality, and Historicity in the Digital Utopia,” in Change Mummified: Cinema, Historicity, Theory, Minneapolis, Minnesota University Press, 2001.
- Schwartz 1999: Louis-George Schwartz, “Mechanical Witness, on the Use of Evidentiary Film and Video in United States Courts,” Ph.D. dissertation, University of Iowa, 1999.
- Shafto 2000: Sally Shafto, “Ut Pictura Cinema: The Aesthetic Adventure of Jean-Luc Godard,” Ph.D. dissertation, University of Iowa, 2000.
- Sorlin 1988: Pierre Sorlin, “Stop the Rural Exodus,” Historical Journal of Film, Radio, and TV, Vol. 18, no. 2, 1988.
- Weiss 1981: John H. Weiss, “An Innocent Eye? The Career and Documentary Vision of Georges Rouquier up to 1945,” Cinema Journal, Vol. 20, no. 2, 1981.
- Zielinski 1999: Siegfried Zielinski, Audiovisions: Cinema and Television as Entr’actes in History, Amsterdam, Amsterdam University Press, 1999.