Stephen Heath, "Narrative Space"

Published in Screen (Autumn 1976), 17:Tg-75, and included with minor revisions in the author's Questions of Cinema (London: Macmillan, T98T, and Bloomington: Indiana University Press, 1981). From: Rosen 1986, 379ff.

"It is precise that 'events take place' "—Michael Snow

At a climactic point in Hitchcock's Suspicion, Lina (Joan Fontaine) receives a visit from two police inspectors come to inform her of the death of a friend in circumstances which cannot but increase her fears concerning the probity—the rectitude—of her husband Johnnie (Cary Grant). The scene finds its center in a painting: the massive portrait of Lina's father which bears with all its Oedipal weight on the whole action of the film [image]—this woman held under the eye of the father (the name as crushing as the image: General MacLaidlaw), sexuality in place as transgression ("Lina will never marry, she's not the marrying sort. . . Lina has intellect and a fine solid character," declares the General early on in the film), as radically "impossible" (leaving her father for Johnnie, Lina is henceforth racked by doubt, a suspicion that is irresolvable, for her and the film)—and before which she now positions herself to read the newspaper report of the friend's death and to gather strength enough to face the scrutiny of the law, the look relayed from portrait to police and to portrait again (stills 1, 2, 3, 4).





Thus centered, the scene is set out according to that unity so characteristic of classical cinema in its narrative spectacle: the new arrives—the visit, the death, the doubt augmented—and the action is continued, pushed forward, but within a movement of rhyme and balance, of sustained coherence: on either side of what Lina is here given to see (the insert "Stop Press" report for which she puts on her /380/ glasses, catching up one of the basic figures of the film, and which we share from her reading, as previously we share the photo of Johnnie in the society magazine [image] or his telegram on the eve of the Hunt Ball), from the entry of the two inspectors back to their departure at the end of the scene, a perfectly symmetrical patterning builds up and pieces together the space in which the action can take place, the space which is itself part of that action in its economy, its intelligibility, its own legality.

Consider simply in this respect, across the scene, the shots at the start and close of the visit (still 5, 6).



The coherence is clear—the end comes round to the beginning, one shot echoing the other in the resolution of rhyme—at the same time that the distance traveled forward in the scene is registered, space redefined in the light of the dramatization effected—alone, diminished by the high angle, Lina is helplessly entangled in the network of shadows, enmeshed in "the spider's web" of her doubt (the image is common in critical discussions of the film). Moreover, the first shot itself is immediately and dramatically exhaustible in its situation in the film: the maid, Ethel, announces the visitors and functions globally as a comic turn—"Oh! Mr Aysgarth! What will my young man think!"—in what is, after all, a Hollywood version of England in the 1930s; the dog, another turn, is an impetuous present from Johnnie to Lina; the house is an example of Johnnie's profuse irresponsibility ("Johnnie, you're a baby," comments Lina, dumbfounded, when he shows her the house after the honeymoon). Everything is placed, there is nothing out of line. And yet, something does jar, already, in this first shot. The composition is faultless, the framing describes the theatricality of the inspectors' entry (the ring at the door, the interruption, the unknown), with the columns, steps, and walls providing a stage effect, the characters are centered, perspective is sharp: the image is in every sense clearly directed. But not quite. Out of the action, breaking the clarity of direction, obstinately turned away, one of the inspectors is pulling to the left, gazing abruptly at something hidden from us, without reason in this scene.

If a painting stands straight at the center of the scene, the look that holds Lina's reception of the news, that organizes the scene itself, it goes askew at the edges of the beginning and end, instants indeed of another painting. What occupies Benson, the gazing inspector, lost in a kind of fascinated panic, is precisely this other painting, hung on the side wall behind the column by the front door and with a little—repeated—scene of its own within the larger scene in which it is somehow included (stills 7, 8, 9, 10 and 11, 12, 13).










At the beginning, just after Lina asks the maid to show the inspectors in, there is a shot of the latter still waiting by the front door but from an angle that now reveals the post-cubist, Picasso-like painting[1] that is the object of Benson's gaze (still 7); the next shot cuts in closer to give the painting in detail while Benson cranes forward to see it (still 8), a brief piano phrase totally different from the expressive orchestrations elsewhere dominant emerging on the sound track; cut back to the angle and distance of the previous shot as the maid comes to take the /382/ inspectors in to her mistress, Benson turning round with a look of shock on his face (still 9) and then back to the painting again before following the maid and his colleague, continuing nevertheless to throw backward glances at the painting (still 10). Similarly at the end of the scene: Lina accompanies the inspectors to the door and, while she and the other inspector, Hodgson, are exchanging a few words, Benson once more pulls to the very edge of the frame, toward the disconcerting painting (still 11); cut to a shot of him craning, with the brief piano phrase, exactly parallel to the one at the beginning (still 12, cf. still 8); cut back to the medium three-shot, Benson totally disframed, Hodgson having literally to order him back into the scene, into the action (still 13).

The play here is complex: this other painting has no reason, is "useless" (isolated, without resonance over the film, marked off by the piano phrase and /383/ by the fact of its link with Benson who remains more or less apart in the main substance of the scene, out of frame and with only one line of any significance), beyond the limits of the film; and yet it arrives in the film, set into the rhyming balance of the scene, serving to demonstrate the rectitude of the portrait, the true painting at the center of the scene, utterly in frame in the film's action. A "Hitchcock joke"? Perhaps. But a joke that tells in a film that hesitates so finely in its enclosure of space, the terms of its points of view. Organized from Lina's point of view (insofar as we have the scenes that she has with respect to her husband, never seeing him separately in a way that might decide the sense of his actions, break the doubt) but under the inspection of an eye (the portrait its mirror) that gives the theater of the suspicion, the setting of Lina's career, the film as story is easy in its ambiguity: no matter if Johnnie is crooked or not, the picture—from portrait to film—is straight, receivable, readable, psychologically and dramatically; Lina's character, her doubt, our experience of that are in place and it is this place that is important, that is the film's reason. Hence, however, a problem of ending (it is contingently interesting in this connection that Hitchcock had an alternative ending, that an attempt was made by cutting to produce a version that would eliminate any equivocation as to Cary Grant-as-Johnnie's honesty, that there were difficulties). Lina and Johnnie struggle in the car, Johnnie explains, Lina's doubt is resolved, the car U-turns to take them back together. The Unity of the place—containing transgression and sexuality and doubt and guilt, the whole family romance—splits, the perspective now lost, the picture of ambiguity broken in the absolute-since-here-arbitrary "banality" of the enforced happy ending (the constraint of"Cary Grant") which brings back, as its contradiction, the memory of the "original" struggle outside the church when, in an abrupt moment of violence, Lina is suddenly somewhere else, fighting off Johnnie in the distance of a shot and a space (a windswept empty wasteland in the middle of an English village hitherto and thereafter presented with all the cosy sporting bucolicness one might expect) that is never finally recaptured in—for—the film, remains left over within it, a kind of missing spectacle. Benson's painting too—"his" insofar as it catches him out in his gaze—has its effect as missing spectacle: problem of point of view, different framing, disturbance of the law and its inspectoring eye, interruption of the homogeneity of the narrative economy, it is somewhere else again, another scene, another story, another space.

Snow's stress: events takc place. What, then, is this "taking place" in film? Suspicion suggests such a question, its action so tightly dependent on the construction and holding of place, its references to painting in the course of that construction and holding, its points of joke or difficulty, excess or otherness. A question that is today posed with insistence, practically and critically, in filmmaking and film theory. Annette Michelson, for instance, describing the achievement of the work of Snow himself, writes that he "has redefined filmic space as that of action," has refound "the tension of narrative" in "the tracing of spatio- /384/ temporal donnees."[2] Snow's example, which is indeed crucial in this respect, can serve here as a simple reminder of the importance of a whole number of differing explorations in independent cinema of space and time, narrative and place. Equally, attention has been directed in film theory to "spatial and temporal articulations," to "kinds of space" and their narrative determinations or disruptions. The basic text of such an attention is Noel Burch's Theory of Film Practice,3 and something of its implications—its positions—can be seen in the work on Ozu Yasujiro by Edward Branigan, David Bordwell, and Kristin Thompson; work which hinges on the demonstration of a certain "foregrounding" of space in Ozu's films and on the argument that this foregrounding confirms Ozu as a "modernist" filmmaker: "the modernity of Ozu's work involves the use of specific spatial devices which challenge the supremacy of narrative causality"; "space, constructed alongside and sometimes against the cause/effect sequence, becomes 'foregrounded' to a degree that renders it at times the primary structural level of the flm"; "it is this foregrounding of the spatial code in Ozu's films that justifies us in classifying Ozu as a 'modernist' film-maker,''4 work which in its example in the field of theory again underlines the insistent actuality of the question of space in film and the "taking place" of events.

If that same question were posed to the start of cinema's history, the answer would come easily enough, without problem: the space of film is the space of reality, film's ambition and triumph is "to reproduce life" (Louis Lumiere); "nature caught in the act," as a spectator put it after one of the first Cinematographe showings in the Grand Cafe, while another extolled the finding at last of the "universal language," "la langue universelle est trouvee!"s As its source and authority (its very "author"), the universal language has no less than the universe itself, the world embraced by the eye of the camera and delivered over on screen, the world in views (films are listed as vues in early French catalogues). The long-shot is there in classical narrative cinema as it subsequently develops as the constant figure of this embracing and authoritative vision, providing the conventional close to a film, the final word of its reality.

That reality, the match of film and world, is a matter of representation, and representation is in turn a matter of discourse, of the organization of the images, the definition of the "views," their construction. It is the discursive operations that decide the work of a film and ultimately determine the scope of the analogical incidence of the images; in this sense at least, film is a series of languages, a history of codes. The universalist temptation, of course, is exactly the grounding in analogy: film works with photographs and, in the technological, economic, and ideological conjuncture of the birth and exploitation of cinema, the photograph is given as the very standard of the reproduction of the real ("photographic realism"). Scientifically, the addition of movement to the photograph to give a picture of life as we see it in the hustle and bustle of the arrival of the train at La Ciotat could be regarded as without interest; illusion is not analysis: Marey, the chronophotographer, has no hme for the inema in the developm nt of which /385/

_ Narrative Space 385

he nevertheless plays a part. Ideologically, the addition of movement (as later the addition of sound to the moving picture) is the possibility of the investment of the photograph as currency of the real in systems of representation that can engage that reality and the guarantee of its vision in a constant—industrial— production of meanings and entertainment within the terms of those meanings.

Meaning, entertainment, vision: film produced as the realization of a coherent and positioned space, and as that realization in movement, positioning, cohering, binding in. The passage from views to the process of vision is essentially that of the coding of relations of mobility and continuity. Early film space tends simply to the tableauesque, the set of fixed-camera frontal scenes linked as a story ("The Original Comedy Chase/The Most Familiar and Laughable Incident in the Whole List of Childhood Tales/Shown in Eight Snappy Scenes').6 Evidently, the tableau has its structure of representation, but that structure misses the subject in the very moment of the movement it now offers: the spectator is placed in respect of the scene but the movement is potentially and perpetually excessive. To link scenes as story is not yet to contain that excess in the achievement of a homogeneously continuous space, the spectator cut in as subject precisely to a process of vision, a positioning and positioned movement. It is here that we touch on the history of cinema in its development of codes and systems: beneath that,on the fact of cinema as order of space and time: "film is not a sum of images but a temporal form"; "movement is not just perceived in itself but localized in space. . . the spectator is not just responsive to what is moving but also to what stays in place and the perception of movement supposes fixed frames."7 Such phenomenological descriptions insist on the interlocking spatiotemporality of film and suggest in their turn something of the general area of the problems of fllm in this connection, those problems that are currently and rightly important. Bearing in mind the particular points of the emergence of that current importance, the aim in what follows will be to provide a descriptive and theoretical context for understanding the debate and to indicate, in so doing, certain critical conclusions with regard to film as "narrative space."

Photography and cinema share the camera. Photography is a mode of projecting and fixing solids on a plane surface, of producing images; cinema uses the images produced by photography to reproduce movement, the motion of the flow of the images playing on various optical phenomena (phi-effect, retinal persistence) to create the illusion of a single movement in the images, an image of movement. Phenomenologically, the result is characterized as "neither absolutely two-dimensional nor absolutely three-dimensional, but something between."[1] The "something between" is the habitual response to the famous "impression of reality" in cinema and it is this impression, this reality that are of concern here in their implications for a consideration of space in film.

Stress has been laid in recent work on the situation of cinema in terms of a development of codes of figuration inherited from the Quattrocento, notably /386/codes of perspective. The focus of attention thus defined is, exactly, the camera: "a camera productive of a perspective code directly constructed on the model of the scientific perspective of Quattrocento" (Marcelin Pleynet);[2] the stress, in other words, is on the camera as machine for the reproduction of objects (of solids) in the form of images realized according to the laws of the rectilinear propagation of light rays, which laws constitute the perspective effect. In this connection, there are already a number of remarks and clarifications to be made, remarks that will bear on Quattrocento perspective, the photograph, and cinema, and in that order.

The perspective system introduced in the early years of the fifteenth century in Italy (developing above all from Florence) is that of central projection: "It is the art of depicting three-dimensional objects upon a plane surface in such a manner that the picture may affect the eye of an observer in the same way as the natural objects themselves.... A perfectly deceptive illusion can be obtained only on two conditions: (a) the spectator shall use only one eye, (b) this eye has to be placed in the central point of perspective (or, at least, quite near to this point)."[3] The component elements of that account should be noted: the possible exact match for the eye of picture and object, the deceptive illusion; the center of the illusion, the eye in place. What is fundamental is the idea of the spectator at a window, an aperta finestra that gives a view on the world—framed, centered, harmonious (the istoria). Alberti, in his treatise Della Pittura written circa 1435, talks of the picture plane as of a pane of glass on which the world in view can be traced: "Painters should only seek to present the form of things seen on this plane as if it were of transparent glass. Thus the visual pyramid could pass through it, placed at a definite distance with definite lights and a definite position of center in space and a definite place in respect to the observer."[4] The cost of such fixed centrality is the marginal distortion which ensues when the observer's eye is not correctly in position in the center of the perspective projection but pulls to the edge (like Benson's gaze in Suspicion, which then receives the shock of another—confusing—painting). Anamorphosis is the recognition and exploitation of the possibilities of this distortion; playing between "appearance" and "reality," it situates the center of the projection of the painting (or of a single element, as in Holbein's "The Ambassadors" in the National Gallery) obliquely to the side, the sense of the painting—its representation—only falling into place (exactly) once the position has been found. Galileo abhorred these perversions of the "normal" view into a turmoil of lines and colors ("una confusa e inordinata mescolanza di linee e di colori"[5]) but, developed in the course of the sixteenth century and particularly appreciated in the following two centuries, they can be seen as a constant triumph of central perspective, a kind of playful liberation from its constraints that remains nevertheless entirely dependent on its system, a ceaseless confirmation of the importance of center and position. What must be more crucially emphasized is that the ideal of a steady position, of a unique embracing center, to which Galileo refers and to which anamorphosis pays its /387/ peculiar homage, is precisely that: a powerful ideal. To say this is not simply to acknowledge that the practice of painting from the Quattrocento on is far from a strict adherence to the perspective system but demonstrates a whole variety of "accommodations" (in certain paintings, for example, buildings will be drawn with one center according to central perspective while a separate center will then be chosen for each set of human figures); it is also to suggest that there is a real utopianism at work, the construction of a code—in every sense a vision— projected onto a reality to be gained in all its hoped-for clarity much more than onto some naturally given reality; a suggestion that merely repeats the conclusions of Francastel in his study of the birth of Quattrocento space: "It was a question for a society in process of total transformation of a space in accordance with its actions and its dreams.... It is men who create the space in which they move and express themselves. Spaces are born and die like societies; they live, they have a history. In the fifteenth century, the human societies of Western Europe organized, in the material and intellectual senses of the term, a space completely different from that of the preceding generations; with their technical superiority, they progressively imposed that space over the planet."[6] For five centuries men and women exist at ease in that space; the Quattrocento system provides a practical representation of the world which in time appears so natural as to offer its real representation, the immediate translation of reality in itself.

The conception of the Quattrocento system is that of a scenographic space, space set out as spectacle for the eye of a spectator. Eye and knowledge come together; subject, object and the distance of the steady observation that allows the one to master the other; the scene with its strength of geometry and optics. Of that projected utopia, the camera is the culminating realization (the camera obscura, described by Giambattista della Porta in 1589 in a treatise on optics, commands attention in the wake of the spread of the Quattrocento system); the images it furnishes become, precisely, the currency of that vision, that space: "Strong as the mathematical convention of perspective had become in picture making before the pervasion of photography, that event definitely clamped it on our vision and our beliefs about 'real' shapes, etc. The public has come to believe that geometrical perspective, so long as it does not involve unfamiliar points of view, is 'true,' just as a long time ago it believed that the old geometry of Euclid was 'the truth' "; "Every day we see photographs which are central perspective images. If another system were applied to the art of painting one could believe that one was living in a bilingual country."[7] Insofar as it is grounded in the photograph, cinema will contribute to the circulation of this currency, will bring with it monocular perspective, the positioning of the spectator/subject in an identification with the camera as the point of a sure and centrally embracing view (Metz draws further conclusions from this identification in his essay "The Imaginary Signifier").[8]

"Our field of vision is full of solid objects but our eye (like the camera) sees this field from only one station point at a given moment. . . "[9] The compar- /388/ ison of eye and camera in the interests of showing their similarity has come to seem irresistible: our eye like the camera, with its stationary point, its lens, its surface on which the image is captured, and so on. In fact, of course, any modern scientific description of the eye will go on to indicate the limits of the comparison. Our eye is never seized by some static spectacle, is never some motionless recorder; not only is our vision anyway binocular, but one eye alone sees in time: constant scanning movements to bring the different parts of whatever is observed to the fovea, movements necessary in order that the receptive cells produce fresh neuroelectric impulses, immediate activity of memory inasmuch as there is no brute vision to be isolated from the visual experience of the individual inevitably engaged in a specific sociohistorical situation. In a real sense, the ideological force of the photograph has been to "ignore" this in its presentation as a coherent image of vision, an image that then carries over into a suggestion of the world as a kind of sum total of possible photographs, a spectacle to be recorded in its essence in an instantaneous objectification for the eye (it would be worth considering the ideological determinations and resonances of the development and commercialization of polaroid photography); a world, that is, conceived outside process and practice, empirical scene of the confirmed and central master spectator, serenely "present" in tranquil rectilinearity (a curvilinear perspective, for which arguments of "optical realism" can be adduced if need be, comfortably rejected as out of true, as "wrong").

Cinema is involved with photograph and camera, its principal matter of expression that of moving photographic images ("principal" as we know it in its history), its prime achievement that of the creation of the "impression of reality"—"neither absolutely two-dimensional nor absolutely three-dimensional, but something between." The latter description reads in many ways like an account of the effect of depth of field which gives very much the possibility of a cued construction of space in accordance with the Quattrocento system. Yet cinema can also use in one and the same film quite other projections (lenses with long focal length, for example), projections which approximate more or less, but differently, to the perspective model; simply, angles and distances change, the center shifts its points. It may well be that classically cinema acquires "the mobility of the eye" while preserving the contained and delimited visual field on which "correct" perspectives depend, but the mobility is nevertheless difficult: movement of figures "in" film, camera movement, movement from shot to shot; the first gives at once a means of creating perspective (the movements of the figures in a shot can "bring out" the space, show relative positions, suggest depth) and a problem of "composition" (film is said to destroy the "ordinary laws" of pictorial organization because of its moving figures which capture attention against all else); the second equally produces problems of composition and, though often motivated in the manuals by some extension of the eye-camera comparison (the camera executes the same movements as the head; horizontal panning is turning the head, etc.), is strictly regulated in the /389/ interests of the maintenance of scenographic space (the version of space, indeed, which determines the justifying comparison); the third, again apt to receive the comparative motivation ("Insofar as the film is photographic and reproduces movement, it can give us a life-like semblance of what we see, insofar as it employs editing, it can exactly reproduce the manner in which we normally see it"[10]), effectively indicates the filmic nature of film space, film as constantly the construction of a space (thus Branigan will conclude that "that space exists only at twenty-four frames per second"[11]). The ideal of space remains that of photographic vision, which brings with it the concern to sustain the camera as eye; in the sense of the detached, untroubled eye discussed earlier, an eye free from the body, outside process, purely looking (no matter, finally, if the falsity of eye-camera comparison be admitted since it can be retrieved with a confirming twist: the eye in cinema is the perfect eye, the steady and ubiquitous control of the scene passed from director to spectator by virtue of the cinematic apparatus: "The director's aim is to give an ideal picture of the scene, in each case placing his camera in such a position that it records most effectively the particular piece of action or detail which is dramatically significant. He becomes, as it were, a ubiquitous observer, giving the audience at each moment of the action the best possible viewpoint.")[12] The ideal, however, is a construction, the mobility acquired is still not easy, the shifting center needs to be settled along the film in its making scenes, its taking place; space will be difficult.

To put it another way: mobility is exactly what is possible in film, complicit— the possibility of holding film within a certain vision, thereby "perfected"— and radical—the possibility of film disturbing that vision, with which nonetheless it is immediately involved, historically, industrially, ideologically. Cinema is not simply and specifically ideological "in itself"; but it is developed in the context of concrete and specific ideological determinations which inform as well the "technical" as the "commercial" or "artistic" sides of that development. For Marey, cinema did nothing "to rid the eye of any of its illusions" sinceit is set up precisely to play on the illusions of a conventional vision, to "reproduce life" as Lumiere put it; for Vertov, cinema could be made to challenge that vision by constructions of dissociations in time and space that would produce the contradictions of the alignment of camera-eye and human eye in order to displace the subject-eye of the sociohistorical individual into an operative—transforming— relation to reality. Film is dominantly articulated in the interests of the "theatrical cinema" Vertov sought to shatter, the world of the scene and the stasis of its relations of vision; but Brecht, and Benjamin with him, will see in the very fact of the succession of film images a certain contradiction to be exploited against that theater, for a different vision, a different space. In its developments and possibilities, its constraints and disruptions, it is the whole question of space in film that must now be examined further.


The examination of space in film may be divided for the moment into two: the /390/ examination of space "in frame," of the space determined by the frame, held within its limits; the examination of space "out of frame," the space beyond the limits of the frame, there in its absence and given back, as it were, in the editing of shot with shot or in camera movement with its reframings. The division can be maintained long enough to allow an order for the remarks that follow, remarks which will finally suggest more clearly its inadequacy.

Screen, frame: Notions of screen and frame are fundamental in the elaboration of the perspective system. Leonardo da Vinci writes: "Perspective is nothing else than seeing a place (or objects) behind a pane of glass, quite transparent, on the surface of which the objects behind that glass are drawn. These can be traced in pyramids to the point in the eye, and these pyramids are intersected on the glass pane."[13] The pane is at once a frame, the frame of a window, and a screen, the area of projection on which what is seen can be traced and fixed; from the Quattrocento on, the "pane" delimits and holds a view, the painter's canvas as a screen situated between eye and object, point of interception of the light rays.

It is worth noting, indeed, in Renaissance (and post-Renaissance) painting the powerful attraction of the window as theme, the fascination with the rectangle of tamed light, the luminously defined space of vision. In Ghirlandaio's Vecchio e bambino (Louvre, Paris), Titian's Isabel di Portogallo (Prado, Madrid) or Dürer's Selbstbildnis (Prado), for example, a window opens to the right, behind the figure portrayed, on to the perspective of a distant horizon; the figure placed almost as by a cinema screen, the sudden illumination of another view, a frame of light to which we are invited to attend. More important, however, is to grasp the very idea of the frame as fully historical in the developments it is given. Before the fifteenth century, frames hardly exist, other than as the specific architectural setting that is to be decorated (wall, altarpiece, or whatever); it is during that century that frames begin to have an independent reality, this concomitant with the growth of the notion itself of "a painting" (the first instance of the use of the word "frame" in an artistic sense recorded by the Oxford English Dictionary is c. 1600). The new frame is symmetrical (the centered rectangle, clearly "composable") and inevitable (the Quattrocento system cannot be realized without it, it becomes a reflex of "natural" composition). Significantly, it brings with it the easel (first recorded instance c. 1634—"a frame or easel called by artists"), "significantly" because the easel is precisely dependent on the idea of "a painting" as single, central view. The painter stands as spectator in /391/ front of his easel (in this history it is men who are the professionals of painting, the authoritative gaze), capturing on the canvas screen the scene behind onto which it gives and which it sets as such: no longer englobed in the area of the painting (dome or arch or ceiling), the painter is definitely upright, an eye on the world, an eye that stations itself, with the easel carried from place to place, much like a tripod. Easel painting, that is, established along with perspective system and camera obscura (the latter itself rapidly becomes a portable apparatus for the mobile painter), is a step in the direction of the camera, a camera that will provide screen and frame and the image reflected, fixed, painted with light: a camera that will culminate this whole vision.

"Frame" describes the material unit of film ("the single transparent photograph in a series of such photographs printed on a length of cinematographic film," "twenty-four frames a second") and, equally, the film image in its setting, the delimitation of the image on screen (in Arnheim's Film as Art, for example, "frame" and "delimitation" are assumed as synonymous). Framing, determining and laying out the frame, is quickly seen as a fundamental cinematic act, the moment of the very "rightness" of the image: "framing, that is to say, bringing the image to the place it must occupy," a definition taken from a manual for teachers written in the 1920s.[14] Quickly too, and in consequence, it becomes the object of an aesthetic attention concerned to pose decisively the problems of the composition of the frame, of what Eisenstein calls "mise en cadre."

"There it is, our 1.33 to 1 rectangle, it will tolerate precious little tampering with at all" (Hollis Frampton).[15] The compositional rectangle is there, carried through into cinema; space is structured within its frame, areas are assigned position in relation to its edges. In a sense, moreover, the constraint of the rectangle is even greater in cinema than in painting: in the latter, its proportions are relatively free; in the former, they are limited to a standard aspect ratio (Frampton's 1.33 to I rectangle, the aptly named "academy frame") or, as now, to a very small number of ratios,[16] with techniques such as masking the sides of the frame to change the size of the rectangle in general disfavor. Hence the rectangle must be mastered—"Maitriser le rectangle," the title of one of the key sections in a modern manual for young people. Hence the rules for mastery, rules which come straight from the Quattrocento system, its balanced vision and the composition of the clarity thus decided; so, from the same modern manual: "To consider the rectangle as a surface crossed by lines of force. . . and with strong points (the points of intersection of those lines) is to guarantee it a solid base structure and to refuse the notion of it as a sort of visual hold-all" "If, therefore, we have to place an actor in this rectangle, one of the best places will be that which follows one of the lines of force in question. And the face, 'strong point' of the human person, will be placed at one of the strong points of the rectangle"; "A second character will naturally be placed at one of the strong points. . ."; "Let us quickly note when we come to 'landscapes' how inharmonious is a division of the surface which does not correspond to the /392/ famous 'third' and how placing the horizon midway in the frame is only apparently logical."[17] In cinema, however, these rules also have their "excess," there is always a further court of appeal—life itself, the very aim of cinema: "But cinema is life, is movement. The cineaste must not fall into the traps of a plastic aesthetic. Failure to remember the rules of framing will often bring agreeable surprises, for it is not without truth that the world is already, in itself, harmonious.''[18]

If life enters cinema as movement, that movement brings with it nevertheless its problems of composition in frame, as was mentioned earlier in the discussion of perspective. In fact, composition will organize the frame in function of the human figures in their actions; what enters cinema is a logic of movement and it is this logic that centers the frame. Frame space, in other words, is constructed as narrative space. It is narrative significance that at any moment sets the space of the frame to be followed and "read," and that determines the development of the filmic cues in their contributions to the definition of space in frame (focus pull, for example, or backlighting). Narrative contains the mobility that could threaten the clarity of vision in a constant renewal of perspective; space becomes place—narrrative as the taking place of film—in a movement which is no more than the fulfillment of the Renaissance impetus, an impetus that a De Kooning can describe as follows: "It was up to the artist to measure out the exact space for a person to die in or to be dead already. The exactness of the space was determined or, rather inspired by whatever reason the person was dying or being killed for. The space thus measured out on the original plane of the canvas surface became a 'place' somewhere on the floor."[19] What is crucial is the conversion of seen into scene, the holding of signifier on signified: the frame, composed, centered, narrated, is the point of that conversion.

Cinema as "life in its truth as scene," the frame as the instance of such a vision. Metz talks here of the regime "of the primal scene and the keyhole": "the rectangular screen permits every type of fetishism, all the effects of 'just-before,' since it places at exactly the height it wants the sharp vibrant bar which stops the seen..."[20] The fascination of the scene is there, and from the beginnings of cinema with its tableaux, its dramatic masks (including the keyhole-shaped matte; as in A Search for Evidence, AM & B 1903), its occasional thematic directness (in Gay Shoe Clerk, Edison Co. 1903, which involves a flirtatious shoe clerk, an attractive young lady, and her chaperone armed with an umbrella, a cut-in closeup shows the young lady's ankle with the clerk's hand gripping her foot into the shoe);[21] the fetishism is there, with the edge, the limit, the setting, the careful place, and from Alberti on—witness that whole series of machines and devices for the production of a certain distance of image, a sure illusion of scrutiny. Simply, the "just-before" in film is spatially moving, the itinerary of a fixity perpetually gained, and the frame stands—acts—in relation to that.

As for the screen, it receives and gives the frame; its flatness halts the image and lays the base of that triangle for which thc spectator's eye provides the apex. /393/ Doubtless there is a sheer pleasure for the position of the eye in the very fact of the projection of the frames onto and from the screen, in their "hitting the screen";[22] a space is established with no "behind" (it is important that the Lumiere brothers should set the screen as they do in the Grand Cafe and not with the audience on either side of a translucent screen, that cinema architecture should take its forms in consequence, that there should be no feeling of machinery to the side of or beyond the screen, that the screen should be one of the most stable clements in cinema's history), a pure expanse that can be invested with depth. The screen, that is, is at once ground, the surface that supports the projected images, and background, its surface caught up in the cone of light to give the frame of the image. Ground and background are one in the alignment of frame and screen, the "on screen in frame" that is the basis of the spatial articulations a film will make, the start of its composition.[23]

Psychoanalysis, it may be briefly added, has come to stress the dream as itself projected on a screen: the dream screen, blank surface present in dreams though mostly "unseen," covered over by the manifest content of the projected dream; a screen that represents the breast (infinitely extensive center of the baby's visual space) and then also sleep (the desire for sleep) as an original ground of pleasure "before" difference, "before" identity, "before" symbolization.[24] In cinema, the images pass (twenty-four per second), the screen remains; covered but there specified—the images of this or that film—but the same—the satisfying projection of a basic oneness. The force of this relation, however, must be understood: it is the passing of the images that produces the constancy of the screen; without those images the screen is "empty," with them it is an impression, a surface/ground that the film and the spectator find as the frames hit the screen that they find intact, safely in the background (revealing and disturbing moment when a character in a flm throws something, as is said, "at the screen").[25]

Movement, transitions. From the very first, as though of right, human figure enter film, spilling out of the train, leaving the factory or the photographic congress, moving—this is the movies, these are moving pictures. The figures move in the frame, they come and go, and there is then need to change the frame, reframing with a camera movement or moving to another shot. The transitions thus effected pose acutely the problem of the filmic construction of space, of achieving a coherence of place and positioning the spectator as the unified and unifying subject of its vision. It is this process of construction, indeed, which is often regarded as the power of cinema and as defining the overall reality of film as that of a kind of generalized "trick effect": "if several successive images represent a space under different angles, the spectator, victim of the 'trick effect,' spontaneously perceives the space as unitary...."[26]

Early films are typically organized as a series of fixed scenes, with a strict unity of time and place. The example was cited above of Tom, Tom, the Piper's Son which tells the well-known story in "eight snappy scenes," simply joined /394/ the one after the other as so many tableaux. The actions of the characters in frame, as though on a stage, make out the sense of the image, center the eye in paths of reading, but within the limits of the distance of the fixed frontal view which creates difficulties of effectively maintaining such a centered perception given the continual wealth of movements and details potentially offered by the photographic image (Ken Jacobs in his film of the same title minutely explores the surface of Tom, Tom, the Piper's Son, refilming from the screen and finding in so doing not just "other" actions but also "central" actions not easily grasped or possibly even missed in the original—as, for instance, the handkerchief stealing in the opening shot). Those difficulties, in the context of its commercial exploitation, are fundamental for cinema's development. The center is the movement, not movements but the logic of a consequent and temporally coherent action. The vision of the image is its narrative clarity and that clarity hangs on the negation of space for place, the constant realization of center in function of narrative purpose, narrative movement: "Negatively, the space is presented so as not to distract attention from the dominant actions: positively, the space is 'used up' by the presentation of narratively important settings, character traits ('psychology'), or other causal agents."[27] Specific spatial cues—importantly, among others, those depending on camera movement and editing—will be established and used accordingly, centering the flow of the images, taking place.

Which is to say, of course, that the tableaux space of the early films is intolerable in its particular fixity, must be broken up in the interests of the unity of action and place and subject view as that unity is conceived from the narrative models of the novelistic that cinema is dominantly exploited to relay and extend. Burch puts it well: "It was necessary to be able to film objects or people close up—to isolate a face, a hand, an accessory (as the discourse of the novel does)— but avoiding any disorientation of the spectator in respect of his or her own _' 'reasoned' analysis of the spatial continuum...."[28] The need is to cut up and then join together in a kind of spatial Aufhebung [sublation] that decides a superior unity, the binding of the spectator in the space of the film, the space it realizes. In the late 1930s and early 1940s, the average shot length of a full-length Hollywood film has been estimated at about 9-10 seconds,[29] but that fragmentation is the condition of a fundamental continuity.

"There are no jerks in time or space in real life. Time and space are continuous. Not so in film. The period of time that is being photographed may be interrupted at any point. One scene may be immediately followed by another that takes place at a totally different time. And the continuity of space may be broken in the same manner."[30] Why is it, Arnheim goes on to ask, that the "juggling with space" possible in film (and including the breaking of a single "real life" space into "several successive images. . . under different angles") does not cause discomfort? The answer refers back to the "something between" status of film previously mentioned: "Film gives simultaneously the effect of an actual happening and of a picture. A result of the 'pictureness' of film is, then, that a /395/ sequence of scenes that are diverse in time and space is not felt as arbitrary.... If film photographs gave a very strong spatial impression, montage probably would be impossible. It is the partial unreality of the film picture that makes it possible.''[31] The emphasis on the "pictureness" of the image is crucial here (there would be problems of cutting for spatial unity with holography): the space constructed in film is exactly a filmic construction. Thus Mitry, for example, will write that shots are like "cells," "distinct spaces the succession of which, however, reconstitutes a homogeneous space, but a space unlike that from which these elements were subtracted."[32]

The conception at work in such descriptions can be seen (even if in this or that writer that conception may be inflected "aesthetically," turned in the direction of film as "art"). The filmic construction of space is recognized in its difference but that difference is the term of an ultimate similarity (indeed, a final "illusion"); the space is "unlike" but at the same time "reconstitutes," using elements lifted from real space. In fact, we are back in the realm of "composition," where composition is now the laying out of a succession of images in order to give the picture, to produce the implication of a coherent ("real") space; in short, to create continuity.

The compositional rules for spatial clarity and continuity are sufficiently well known not to need extended discussion at this stage; it will be enough merely to stress one or two of their determinations. Firstly, the establishment of fixed patterns of clarity for the variation of scale of shot in a scene: there are "normal ways" of organizing dialogue scenes, action scenes, and so on;[33] these systems allowing for a certain free play—"exceptions"—within their overall structure in the interests of "dramatic effect" ("In the normal way, it is almost certainly better to cut the scencs as we have indicated, but. . . there may be exceptions when the rules need to be modified to convey certain dramatic effects").[34] What may be remembered above all in this context is the extreme importance attached to providing an overall view, literally the "master shot" that will allow the scene to be dominated in the course of its reconstitution narratively as dramatic unity ("Even where a sequence starts on a detail, it is important that the whole setting should be shown at some stage").[35] Take the beginning of Jaws: a beach party with the camera tracking slowly right along the line of the faces of the participants until it stops on a young man looking off; eyeline cut to a young woman who is thus revealed as the object of his gaze; cut to a high-angle shot onto the party that shows its general space, its situation, before the start of the action with the run down to the ocean and the first shark attack—the shot serves, that is, as a kind of master fold in the sequence, setting it correctly in place. Second, the establishment of the 180-degree and 30-degree rules. The former matches screen space and narrative space (the space represented in the articulation of the images), ground and background; with its help, "one will always find the same characters in the same parts of the screen."[36] The 180-degree line that the camera is forbidden to cross answers exactly to the 180-degree line of the screen behind /396/ which the spectator cannot and must not go, in front of which he or she is placed within the triangle of representation, the space of the image projected, that is repeated in the very terms of the fiction of the imaged space. As for the latter, a "quick, simple rule that issues directly from the necessities of cinematic fragmentation" and that avoids the "disagreeable sensation" of a "jump in space,"[37] it is finally nothing other than a specific perspective rule for a smooth line of direction in film, for the achievement of a smooth line in from shot to shot. Third and last, following on from those more particular remarks, the establishment generally as a powerful evidence, as a natural basis, of the idea of continuity as smoothness in transitions: the rules of the filmic construction of space on screen (master shot, 180-degree and 30-degree rules, matching on action, eyeline matching, field/reverse field, etc.) background the image flow into a unified subject-space, immediately and fully continuous, reconstitutive: "Making a smooth cut means joining two shots in such a way that the transition does not create a noticeable jerk and the spectator's illusion of seeing a continuous piece of action is not interrupted."[38]

Continuity in these terms is also decisive with regard to transitions and changes of frame effected by camera movement. "Imperceptible" reframing movements, more definite pans, and tracking shots are developed in the interests of the narrative composition of space in relation to the actions of the characters;[39] here, too, rules are elaborated accordingly, the camera having, for instance, to impregnate space with the anticipation of action: "if the actor is accompanied by a movement of the camera, more 'room' must be left in front of him or her than behind, so as to figure sensorially the space to be crossed."[40] In this respect, it is worth bearing in mind the extent to which the sequence-shot-with-deep focus long take valued by Bazin in his account of "the evolution of cinematic language" can stay within such a conception of space. The narrative of a Welles or a Wyler in Bazin's account is carried through in a manner that retains the particular effects to be derived from "the unity of the image in time and space," a manner that refinds and draws out the essential "realism" of cinema; a realism in which space is all important: "the cinematographic image can be emptied of all reality save one—the reality of space."[41] The space of Citizen Kane or The Best Years of Our Lives is still entirely dramatic, however; heightened indeed in its drama: as was suggested earlier, deep focus allows composition for a high degree of perspective ("depth of field" exactly), and this can be increased over the long take with its potential definition of a complex action in a single shot, its filling out of movements and positions in a temporally visible demonstration of space as narrative place.[42] It should anyway be noted that the average shot length overall of Citizen Kane is 12 seconds, "about average for its period,"[43] and it remains true that classically continuity is built on fragmentation rather than the long take—on a segmentation for recomposition that can bind the spectator in the strong articulations of the unity it seeks to create. Elsewhere, Bazin was to refer to the version of the spatial realism he ontologically cherished provided /397/ by Italian Neorealism; a version that might show the possibilities of the long take away from an absorbed dramatic space; and so, by contrast, the force of the classical continuity in that dependence on segmentation/articulation and its effective inclusion of the longer take within its terms of spatial construction.

Those terms, as they have been described here, are the terms of a constant welding together: screen and frame, ground and background, surface and depth, the whole setting of movements and transitions, the implication of space and spectator in the taking place of film as narrative. The classical economy of film is its organization thus as organic unity, and the form of that economy is narrative, the narrativization of film. Narrative, as it were, determines the film which is contained in its process in that determination, this "bind" being itself a process— precisely the narrativization. The narration is to be held on the narrated, the enunciation on the enounced; filmic procedures are to be held as narrative instances (very much as "cues"), exhaustively, without gap or contradiction. What is sometimes vaguely referred to as "transparency" has its meaning in this narrativization: the proposal of a discourse that disavows its operations and positions in the name of a signified that it proposes as its preexistent justification. "Transparency," moreover, is entirely misleading insofar as it implies that narrativization has necessarily to do with some simple "invisibility" (anyway impossible—no one has yet seen a signified without a signifier). The narration may well be given as visible in its filmic procedures; what is crucial is that it be given as visible for the narrated and that the spectator be caught up in the play of that process, that the address of the film be clear. (Does anyone who has watched, say, The Big Sleep seriously believe that a central part of Hollywood films, differently defined from genre to genre, was not the address of a process with a movement of play and that that was not a central part of their pleasure?)

Within this narrativization of film, the role of the character-look has been fundamental for the welding of a spatial unity of narrative implication. In so many senses, every film is a veritable drama of vision, and this drama has thematically and symptomatically "returned" in film since the very beginning: from the fascination of the magnifying glass in Grandma's Reading Glass to Lina's shortsightedness in Suspicion to the windscreen and rearview mirror of Taxi Driver, from the keyhole of A Search for Evidence to the images that flicker reflected over Brody's glasses in Jaws as he turns the pages of the book on sharks, finding the images of the film to come and which he will close as he closes the book; not to mention the extended dramatizations such as Rear Window or Peeping Tom. How to make sense in film if not through vision, film with its founding ideology of vision as truth? The drama of vision in the film returns the drama of vision of the film: the spectator will be bound to the film as spectacle as the world of the film is itself revealed as spectacle on the basis of a narrative organization of look and point of view that moves space into place through the image flow; the character, figure of the look, is a kind of perspective within the /398/ perspective system, regulating the world, orientating space, providing directions—and for the spectator.

Film works at a loss, the loss of the divisions, the discontinuities, the absences that structure it—as, for example, the "outside" of the frame, offscreen space, the hors-champ. Such absence is the final tragedy of a Bazin, who wants to believe in cinema as a global consciousness of reality, an illimitation of picture frame and theater scene—"The screen is not a frame like that of a picture, but a mask which allows us to see a part of the event only. When a person leaves the field of the camera, we recognize that he or she is out of the field of vision, though continuing to exist identically in another part of the scene which is hidden from us. The screen has no wings. . ."[44]—but who can only inspect the damage of "camera angles or prejudices,"[45] acknowledge nonetheless the frame, the scene, the mask, the hidden, the absent. The sequence-shot-with-deep-focus long take functions as a utopia in this context—the ideal of a kind of "full angle," without prejudices, but hence too without cinema; the ideal recognized in Bicycle Thieves, "plus de cinéma."[46]

Burch writes that "off-screen space has only an intermittent or, rather, fluctuating existence during any film, and structuring this fluctuation can become a powerful tool in a filmmaker's hands."[47] The term "fluctuation" is excellent, yet it must be seen that the work of classical continuity is not to hide or ignore offscreen space but, on the contrary, to contain it, to regularize its fluctuation in a constant movement of reappropriation. It is this movement that defines the rules of continuity and the fiction of space they serve to construct, the whole functioning according to a kind of metonymic lock in which offscreen space becomes onscreen space and is replaced in turn by the space it holds off, each joining over the next. The join is conventional and ruthlessly selective (it generally leaves out of account, for example, the space that might be supposed to be masked at the top and bottom of the frame, concentrating much more on the space at the sides of the frame or on that "in front," "behind the camera," as in variations of field/reverse field), and demands that the offscreen space recaptured must be "called for," must be "logically consequential," must arrive as "answer," "fulfillment of promise," or whatever (and not as difference or contradiction)—must be narrativized. Classical continuity, in other words, is an order of the pregnancy of space in frame; one of the narrative acts of a film is the creation of space,[48] but what gives the moving space its coherence in time, decides the metonymy as a "taking place," is here "the narrative itself," and above all as it crystallizes round character as look and point of view. The fundamental role of these is exactly their pivotal use as a mode of organization and organicization, the joining of a film's constructions, the stitching together of the overlaying metonymies.

"If in the left of the frame an actor in close-up is looking off right, he has an empty space in front of him; if the following shot shows an empty space to the left and an object situated to the right, then the actor's look appears to cross an /399/ orientated, rectilinear, thus logical space: it seems to bear with preclsion on the object. One has an eye-line match."[49] The look, that is, joins form of expression— the composition of the images and their disposition in relation to one another— and form of content—the definition of the action of the film in the movement of looks, exchanges, objects seen, and so on. Point of view develops on the basis of this joining operation of the look, the camera taking the position of a character in order to show the spectator what he or she sees.[50] Playing on the assumption of point of view, a film has an evident means of placing its space, of giving it immediate and holding significance; Burch talks of the establishment of an organization founded on the "traditional dichotomy between the 'subjective camera' (which 'places the spectator in the position of a character') and the 'objective camera' (which makes the spectator the ideal, immaterial 'voyeur' of a profilmic pseudo-reality)."[51]

This account, however, requires clarification. The point-of-view shot is "subjective" in that it assumes the position of a subject-character, but to refer to that assumption in terms of "subjective camera" or "subjective image" can lead to misunderstanding with regard to the functioning of point of view. Subjective images can be many things; Mitry, for example, classifies them into five major categories: "the purely mental image (more or less impracticable in the cinema); the truly subjective or analytical image (i.e. what is looked at without the person looking), which is practicable in small doses; the semi-subjective or associated image (i.e. the person looking + what is looked at, which is in fact looked at from the viewpoint of the person looking), the most generalizable formula; the complete sequence given over to the imaginary, which does not raise special problems; and finally the memory image, which is in principle simply a variety of the mental image but, when presented in the form of a flash-back with commentary, allows for a specific filmic treatment which is far more successful than in the case of other mental images."[52] The point-of-view shot includes "the semi-subjective or associated image" (its general mode) and "the truly subjective or analytical image" (its pure mode, as it were) in that classification but not necessarily any of the other categories (a memory sequence, for instance, need not contain any point-of-view shots); what is "subjective" in the point-of-view shot is its spatial positioning (its place), not the image or the camera.

To stress this is to stress a crucial factor in the exploitation of the film image and its relation to point-of-view organization. Within the terms of that organization, a true subjective image would effectively need to mark its subjectivity in the image itself: Examples are common: the blurred image of Gutman in The Maltese Falcon is the subjective image of the drugged Spade; the blurring of focus marks the subjectivity of the image, exclusively Spade's, and the spectator is set not simply with Spade but as Spade. They are also limited, since they depend exactly on some recognizable—marking—distortion of the "normal" image, a narratively motivated aberration of vision of some kind or another (the character is drugged, intoxicated, shortsighted, terrified ... down to he or she /400/ running, with hand-held effects of the image "jogging," or even walking, with regular speed of camera movement forward matched on a shot that effectively establishes the character as in the process of walking; the latter represents the lowest limit on the scale, since the camera movement is there a weak subjective marking of the image which itself remains more or less "normal"—except, of course, and hence this limit position of the banal action of walking, that the normal image is precisely static, that movement in a central perspective system can quickly become a problem of vision). The implication of this, of course, is then the strength of the unmarked image as a constant third person—the vision of picture and scene, the Quattrocento view, Burch's "voyeur" position—which is generally continued within point-of-view shots themselves; the point-of-view shot is marked as subjective in its emplacement but the resulting image is still finally (or rather firstly) objective, the objective sight of what is seen from the subject position assumed. Indicatively enough, the general mode of the point-of-view shot is the shot which shows both what is looked at and the person looking. Instances of the pure shot, showing what is looked at without the person looking, however, are equally conclusive. Take the shot in Suspicion of the telegram that Lina receives from Johnnie to tell her of his intention to attend the Hunt Ball: the telegram is clearly shown from Lina's reading position and the end of the shot—the end of the reading—is marked by her putting down her glasses onto the telegram lying on a table, the glasses thus coming down into frame; the position of the shot is marked as subjective with Lina but the image nevertheless continues to be objective, "the real case" for the narrative.[53]

Point of view, that is, depends on an overlaying of first and third person modes. There is no radical dichotomy between subjective point-of-view shots and objective non-point-of-view shots; the latter mode is the continual basis over which the former can run in its particular organization of space, its disposition of the images. The structure of the photographic image—with its vision, its scene, its distance, its normality—is to the film somewhat as language is to the novel: the grounds of its representations, which representations can include the creation of an acknowledged movement of point of view. This is the sense of the spectator identification with the camera that is so often remarked upon (Benjamin: "the audience's identification with the actor is really an identification with the camera"; Metz: "the spectator can do no other than identify with the camera").[54] The spectator must see, and this structuring vision is the condition of the possibility of the disposition of the images via the relay of character look and viewpoint which pulls together vision and narrative. Emphasis was laid earlier on the structures of the structuring vision that found cinema; what is emphasized now is the dependence of our very notion of point of view on those structures; dependence at once insofar as the whole Quattrocento system is built on the establishment of point of view, the central position of the eye, and insofar as the mode of representation thus defined brings with it fixity and movement in a systematic complicity of interaction—brings with /401/ it, that is, the "objective" and the "subjective," the "third person" and the "first person," the view and its partial points, and finds this drama of vision as the resolving action of its narratives.

Identification with the camera, seeing, the "ideal picture" of the scene: "the usual scene in a classical film is narrated as if from the point of view of an observer capable of moving about the room."[55] Such movement may be given in editing or by camera movement within a shot, and the importance accruing to some master view that will define the space of the mobility has been noted. Movement, in fact, will be treated as a supplement to produce precisely the "ideal picture" (going to the movies is going to the pictures): on the basis of the vision of the photographic image, that is, it will provide the "total" point of view of an observer capable of moving about the room without changing anything of the terms of the vision, the scene laid out for the central observer (and spectator); every shot or reframing adds a difference, but that difference is always the same image, with the organization—the continuity, the rules, the matches, the pyramid structures—constantly doing the sum of the scene.

That said, it remains no less true, as has again been noted and as will become important later on, that movement represents a potentially radical disturbance of the smooth stability of the scenographic vision (hence the need for a systematic organization to contain it). Such a disturbance, however, is not as simple as is sometimes suggested, and it is necessary briefly to consider at this stage two instances of disturbance as they are conventionally described; both bear on the mobility of the camera.

The first is that of what Branigan characterizes as the impossible place: "To the extent that the camera is located in an 'impossible' place, the narration questions its own origin, that is, suggests a shift in narration."[56] "Impossible," of course, is here decided in respect of the "possible" positions of the observer moving about, the disturbance involved seen as a disjunction of the unity of narration and narrated, enunciation and enounced. Thus defined, impossible places are certainly utilized in classical narrative cinema, with examples ranging from the relatively weak to the relatively strong. At one end of the range, the weak examples would be any high or low angles that are not motivated as the point of view of a character; or, indeed, any high or low angles that, while so motivated, are nevertheless sufficiently divergent from the assumed normal upright observing position as to be experienced as in some sense "impossible" in their peculiarity (the most celebrated—and complex—example is the dead-man-in-the-coffin point of view in Vampyr).[57] At the other end, the strong examples—those intended by Branigan—can be illustrated by a description of two shots from Killer's Kiss: (I) as Davey, the boxer-hero, is seen stooping to feed his goldfish, there is a cut to a shot through the bowl, from the other side, of his face peering in as the feed drops down; since the bowl is on a table against a wall, the place taken by the camera is not possible; (2) Rappello, the dance-hall owner, furious at being left by the heroine, is drinking in a back room, its /402/ walls covered with posters and prints; a closeup of a print showing two men leering from a window is followed by a shot of Rappello who throws his drink at the camera ("at the screen"!); a crack appears as the drink runs down a plate of glass; impossibly, the shot was from "in" the print. The second—and related— instance of disturbance is that of the development of camera movement as a kind of autonomous figure; what Burch calls "the camera designated as an 'omnipotent and omniscient' (i.e. manipulative and pre-cognitive) presence."[58] This presence too is utilized in classical narrative cinema, and weak and strong examples can once more be indicated. In Taxi Driver, Travis Bickle is seen phoning Betsy after the porno-film fiasco; as he stands talking into the pay phone, fixed on a wall inside a building, the camera tracks right and stops to frame a long empty corridor leading out to the street; when Travis finishes his call, he walks into frame and exits via the corridor. The tracking movement designates the camera with a certain autonomy—there is an effect of a casual decision to go somewhere else, off to the side of the narrative—but the example is ultimately weak: the corridor is eventually brought into the action with Travis' exit and, more important, it has its rhyming and thematic resonances—the corridors in the rooming house used by Iris, the marked existential atmosphere of isolation, nothingness, etc. Stronger examples are provided in the work of an Ophuls or a Welles—the spectacular tracking shot at the start of Touch of Evil or the intense mobility in many of the shots at the end of that same film.

These two instances of disturbance have been characterized here in their existence in established cinema simply to make one or two points directly in the context of the present account. Thus, the examples given of autonomy of camera movement are all clearly operating in terms of "style" (Welles, Ophuls, the tics of a new American commercial cinema that has learnt a consciousness of style). The crucial factor is not the valuation of camera movement, autonomous or not, but the point at which a certain work on the camera in movement produces the normality of the third person objective basis as itself a construction, gives it as role or fiction, and breaks the balance of the point-of-view system. Similarly, the examples of the impossible place from Killer's Kiss, which also have their reality as stylistic marking in the film, are without critically disruptive extension in themselves, are simply tricks (in the sense of spatial prestidigitations): the impossible place is entirely possible if held within a system that defines it as such, that it confirms in its signified exceptionality. The felt element of trick, moreover, raises the general point of the realization of film as process. It is too readily assumed that the operation—the determination, the effect, the pleasure— of classical cinema lies in the attempt at an invisibility of process, the intended transparency of a kind of absolute "realism" from which all signs of production have been effaced. The actual case is much more complex and subtle, and much more telling. Classical cinema does not efface the signs of production, it contains them, according to the narrativization described above. It is that process that is the action of the film for the spectator—what counts is as much the represen- /403/tation as the represented, is as much the production as the product. Nor is there anything surprising in this: film is not a static and isolated object but a series of relations with the spectator it imagines, plays, and sets as subject in its movement. The process of film is then perfectly available to certain terms of excess—those of that movement in its subject openings, its energetic controls. "Style" is one area of such controlled excess, as again, more powerfully, are genres in their specific version of process. The musical is an obvious and extreme example with its systematic "freedom" of space—crane choreography—and its shifting balances of narrative and spectacle; but an example that should not be allowed to mask the fundamcntal importance of the experience of process in other genres and in the basic order of classical cinema on which the various genres are grounded. Which is to say, finally, the radical disturbance is not to be linked to the mere autonomization of a formal element such as camera movement; on the contrary, it can only be effectively grasped as a work that operates at the expense of the classical suppositions of "form" and "content" in cinema, posing not autonomies but contradictions in the process of film and its narrative-subject binding.

The construction of space as a term of that binding in classical cinema is its implication for the spectator in the taking place of film as narrative; implication-process of a constant refinding—space regulated, orientated, continued, reconstituted. The use of look and point-of-view structures—exemplarily, the field/ reverse field figure (not necessarily dependent, of course, on point-of-view shots)[59]—is fundamental to this process that has been described in terms of suture, a stitching or tying as in the surgical joining of the lips of a wound.[60] In its movement, its framings, its cuts, its intermittences, the film ceaselessly poses an absence, a lack, which is ceaselessly recaptured for—one needs to be able to say "forin"—the film, that process binding the spectator as subject in the realization of the film's space.

In psychoanalysis, "suture" refers to the relation of the individual as subject to the chain of its discourse where it figures missing in the guise of a stand-in; the subject is an effect of the signifier in which it is represented, stood in for, taken place (the signifier is the narration of the subject).[61] Ideological representation turns on—supports itself from—this "initial" production of the subject in the symbolic order (hence the crucial role of psychoanalysis, as potential science of the construction of the subject, with historical materialism), directs it as a set of images and fixed positions, metonymy stopped into fictions of coherence. What must be emphasized, however, is that stopping—the functioning of suture in image, frame, narrative, etc.—is exactly a process: it counters a productivity, an excess, that it states and restates in the very moment of containing in the interests of coherence—thus the film frame, for example, exceeded from within by the outside it delimits and poses and has ceaselessly to recapture (with post-Quattrocento painting itself, images are multiplied and the conditions /404/ are laid for a certain mechanical reproduction that the photograph will fulfill, the multiplication now massive, with image machines a normal appendage of the subject). The process never ends, is always going on; the construction-reconstruction has always to be renewed; machines, cinema included, are there for that—and their ideological operation is not only in the images but in the suture.

The film poses an image, not immediate or neutral,[62] but posed, framed and centered. Perspective-system images bind the spectator in place, the suturing central position that is the sense of the image, that sets its scene (in place, the spectator completes the image as its subject). Film too, but it also moves in all sort of ways and directions, flows with energies, is potentially a veritable festival of affects. Placed, that movement is all the value of film in its development and exploitation: reproduction of life and the engagement of the spectator in the process of that reproduction as articulation of coherence. What moves in film, finally, is the spectator, immobile in front of the screen. Film is the regulation of that movement, the individual as subject held in a shifting and placing of desire, energy, contradiction, in a perpetual retotalization of the imaginary (the set scene of image and subject). This is the investment of film in narrativization; and crucially for a coherent space, the unity of place for vision.

Once again, however, the investment is in the process. Space comes in place through procedures such as look and point-of-view structures, and the spectator with it as subject in its realization. A reverse shot folds over the shot it joins and is joined in turn by the reverse it positions; a shot of a person looking is succeeded by a shot of the object looked at which is succeeded in turn by a shot of the person looking to confirm the object as seen; and so on, in a number of multiple imbrications. Fields are made, moving fields, and the process includes not just the completions but the definitions of absence for completion. The suturing operation is in the process, the give and take of absence and presence,the play of negativity and negation, flow and bind. Narrativization, with its continuity, closes, and is that movement of closure that shifts the spectator as subject in its terms: the spectator is the point of the film's spatial relations—the turn, say, of shot to reverse shot—their subject-passage (point-of-view organization, moreover, doubles over that passage in its third/first person layerings). Narrativization is scene and movement, movement and scene, the reconstruction of the subject in the pleasure of that balance (with genres as specific instances of equilibrium)—for homogeneity, containment. What is foreclosed in the process is not its production—often signified as such, from genre instances down to this or that "impossible" shot—but the terms of the unity of that production (narration on narrated, enunciation on enounced), the other scene of its vision of the subject, the outside—heterogeneity, contradiction, history—of its coherent address.

Return to Glossary


[1] R. Arnheim, Film as Art (London: Faber, I969), p.20.

[2] See M. Pleynet, interview (with Gerard Leblanc), Cinethique (I969), no. 3.

[3] G. Ten Doesschate, Perspective: Fundamentals, Controversials, History (Nieuwkoop: B. de Graaf, 1964), pp. 6-7.

[4] Leon Battista Alberti, On Painting (New Haven and London: Yale University Press, I966), p.5I.

[5] Galileo, Opere, ed. A. Favaro (Florence: Edizione nazionale, I 890- I 909),9: I 29.

[A recent book by Ernest B. Gilman, The Curious Perspective: Literary and Pictorial Wit in the Seventeenth Century (New Haven and Londo~l: Yale University Press, 1978), stresscs—as in effect does Galileo—a parodic, almost subversive implication of the use of anamorphosis: "Although the curious perspective system would have been impossible without the achievement of a systematic linear perspective in the earlier Renaissance, its effect was to parody, question, and even undermine the central cognitive assumption behind perspective representatior~" (p. 233). It remains, however, that the "wit" of anamorphosis is constantly a reference to a rational and stable system that it assumes in the very moment it parodies or questions and is thus al~vays available as a final image of order; as witness the idea of anamorphosis in a passage from Leibuiz on universal harmony quoted by Gilman (p. 97). "It is as in the inventions of perspective, where certain lovely drawings appear only as confusion, until one finds their true point of view or sees them by means of a certain glass or mirror.. . . Thus the apparent deformities of our little world come together as beauties in the greater woHd, and there is nothing opposed to the unity of a universally perfect principle." What is clear and important is that the Renaissance perspective system opens the way to an assurance and a trap for the look, the vision of the subject, to an illusion of realitY in the play of which two terms a whole problematic of representation is established—a problematic in which cinema is engaged, moves.—S.H.]

[6] R Francastel, Etudes de sociologie de l'art (Paris: Denoel, I970), pp. I36-37.

[7] W. M. Ivins, Art and Ceometry (New York: Dover, 1964), p. I08; Ten Doesschate, Perspective, p. I57.

[8] C. Metz, "Le Signifiant imaginaire," Communications (1975), 23:35-I37; translation, "The Imaginary Signifier," Screen (Summer I97S) I6(2):52-54

[9] Arnheim, Film as Art, p. I8.

[10] E. Lindgren, The Art ofthe Film (London: Allen and Unwin, I948), p.54. _~

[11] Branigan, "TheSpaceofEquinoxFlou~er,~p.

[12] K. Reisz and G. Millar, The Technique of Film Editing (New York and London: Hastings House, 1968), p. 215. _l

[13] J. R Richter, ed., The Literary Works of Leonardo da Vinci (London: Oxford University Press, 1939), I:I50. The figure is Leonardo's own, ibid. [Leonardo was much exercised by difficulties in the match between the Albertian perspective system and visual appearances, exploring elsewhere the possibility of an alternative system based on a spherical optics; see J. White, The Birth and Rebirth of Pictorial Space (Boston: Boston Book and Art Shop, I967) pp. 207-15.—S.H.]

[14] E. Reboul, Le Cinema scolaire et educateur (Paris: Presses Universitaires de France, I926).

[15] Hollis Frampton, interview with Simon Field and Peter Sainsbury, Afterimage (Autumn 1972), no. 4, p. 65

[16] Frampton writes elsewhere: "The film frame is a rectangle, rather anonymous in its proportions, that has been fiddled with recently in the interests of publicising, so far as I can see, nothing much more interesting than the notions of an unbroken and boundless horizon. The wide screen glorifies, it would seem, frontiers long gone: the landscapes of the American corn-flats and the Soviet steppes; it is accommodating to the human body only when that body is Iying in state. Eisenstein once proposed that the frame be condensed into a 'dynamic' square, Nawative Space 417vEich is as close to a circle as a rectangle can get, but his arguments failed to prosper." Hollis Frampton, "The Withering Away of the State of Art," Artforum (L)ecember 1974), p. 53.

[17]Apprendre le cinema, speeial issue of Image et son (May 1966), no. I94 bis, pp. I 19, 121.

[18] ibid., p. 123. , i

[19] Quoted by Rosalind Krauss in "A View of Modernism," Artforum (September 1972), p. 50 Krauss comments: "Perspective is the visual correlate of causality that one thing follows the next in space according to rule. . . perspective space carried with it the meaning of narrative: a suecession of events leading up to and away from this moment; and within that temporal succession—given as a spatial analogue—was secreted the 'meaning' of both that space and those events."

[20] C. Metz, "Histoire/Discours," in Langue, discours, societe', ed. J. Kristeva, J.-C. Milner, and N. Ruwet (Paris: Seuil, T975), p. 304; translatioll, "History/Discourse," Edinburgh '76 Magazine (~tJ7(~), p. 23

[21] A still from the shot can be found in Niver, First Tu~enty Years, p. 36.

[22] "There must be a lot of essential pleasure just in the films when they hit the screen—I heard this expression yesterday, 'to hit the screen,' that s fantastic in English. Hit the screen— this is really what the frames do. The projected frames hit the screen." Peter Kubelka, interview with lonas Mekas in Gidal, Structural Film Anthology, p. 102.

[23] It can be noted that much independent film work has been concerned with experiencing dislocations of screen and frame; Sharits, for example, writes: "When a film 'loses its loop' it allows us to see a blurred strip of jerking frames; this is quite natural and quite compelling _; subject ~l~aterial. When this non-framed condition is intentionally iTIduced, a procedure I am _' currently exploring, it could be thought of as anti-framing.' " Paul Sharits, "W~rds per Page," ~ Afterimage (Autumn 1972), no. 4, p. 40. For an attempt by a filmmaker to provide a theoretical _j formulation of such dislocation using the notion of a "second screen" (in fact, the frame on _~ screen in a narrative coherence of ground/background) that independent cinema will destroy _j ("in independent cinema, there is no second screen"), see Claudine Eizykman, La Jouissance cinema (Paris: Bourgois, 1976) esp. pp. 147-51. _~

[24] B. D Lewin, "Sleep, the Mouth, and the Dream Screen," Psychoanalytic Quarterly (1946), _~ S:419-34

[25] Discussion of screen and dream screen is suggested at the close of a recent article by Guy Rosolato, "Souvenir-ecran," Communications (1975), no. 23, pp. 86-87. See also my "Screen ~¤Images, Film Memory," Edinhurgh '76 Magaine (1976), pp. 33-42. ~

[26] C. Metz, Essais sur la signifTcation au cinema (Paris: Klincksieck, 1972), 2:189. ~ .

[27] Thompson and Bordwell, "Space and Narrative in the Films of Ozu," p. 42. For an initial discussion of proccdures of image ccntering ("specification procedures"), see my "Film and System: Terms of Analysis" part 11, Screen (Summer 1975), 16(2):gg-loo

[28] Noel Burch, "De Mabuse d M: le travail de Fritz Lang," in Cinema: Theorie, Lectures, special issue of the Revue e'Esthetique (Paris: Klincksieck, 1973), p. 229. ~

[29] See Barry Salt, "Statistical Style Analysis of Motion Pictn!-~ I l~`, (~)f~i.'n ;i,' (1-.~!1 ~n~4 _; pp. 13-22.

[30] Arnheim, Film as Art, p.

[31] 16id., p. 32.

[32] J. Mitry, Esthetique et psychologie de cinema (Paris: Editions Universitaires, 1965) 2: lo. _'

[33] Branigan, "The Space of Equinox Flou~er," p. 75 gives the schema of the inverted pyramid structure characteristic of classical Hollywood film: ''I. Establishing Shot (a major variant: we see a detail of the scene, then pull back or cut to the establishing shot) 2. Long Shot (master _shot) 3. Medium Two-Shot 4. Reverse Angles (over-the-shoulder shots) 5. alternating Medium Close-ups 6. Cut-away (or Insert) 7. alternating Medium Close-ups 8. Re-establishing Shot _: (usually a reverse angle or tw~-shot)." _

[34] Reisz and Millar, Technique of Film Editing, pp. 224-25.

[35] 16id., p. 225-26. _~

[36] Apprendre le cinema, p. T42.

[37] Ibid., p. 151.

[38] Reisz and Millar, Technique of Film Editing, p. 216. To emphasize the reality of this smoothness as construction rather than " reflection," it can be noted that the Navajo Indians studied by Worth and Adair, though capable of producing the 'torrect" continuity (for example, by matching on action), were very far from the "rules" in their films, articulating another system of space as an area of action (in which "jumps" from the standpoint of the vision of the rules becomes essential continuities); see Sol Worth andJohn Adair, Through Navajo Eyes (Bloomington: Indiana University Prcss, 1972), p. 174 and stills 22-35, 35-40

[39] Barry Salt has pointed to the importance of thc outdoor-action subject film (notably the western) historically in this development; "The Early Development of Film Form," Film Form (Spring T976), no. I, pp 97 - 98. 1

[40] Apprendre le cinema, p. 125 ("an orientated empty space is a promise"). 1

[41] Andre Bazin, What Is Cinema? (Los Angeles and Berkeley: University of California Press, lT967), I:To8. 1

[42] Which is not, of course, to say that deep focus must ~lecessarily be used in this way; for analysis of "a refusal of pcrspcctive within depth of field," see Cl. Bailble, M. Marie, and M.- C. Ropars, Muriel (Paris: Galilee, 1974), pp. 128-36.

[43] Salt, "Statistical Style Analysis," p. 20.

[44] Andre Bazin, Qu'est-ce que le cinema? (Paris: Cerf, 1959), 2:100.

[45]Bazin, Qu'est-ce que le cinema? (Paris: Cerf, 1962), 4:57. S3

[46] ibid., p. 59. For discussion of Bazin on Neorealism, see Christopher William's article of that title in Screen (Winter 1973 - 74), 14(4):61 - 68.

[47] Burch, Theory of Film Practice, p. 21.

[48] Branigan, "The Space of Equinox Flou~er," p. 103.

[49] Apprendre le cinema, p. 148.

[50] For a detailed analysis of the point-of-view shot, see Edward Branigan, "Formal Permutations of the Point-of-View Shot," Screen (Autumn 1975), 16(3):54-64. _~

[51] Noel Burch andJorge Dana, "Propositions," Afferimage (Spring 1974), no. i, p. 45.

[52] As summarized by Metz in his "Current Problems in Film Theory," Screen (SpringSummer T973), /4(~—2):4~9

[53] In fact, and not suprisingly, the less narratively "metonymical" and the more "metaphorical" what is looked at in the pure point-of-view shot is (without the marking of image distortion), the nearer such a shot will come to subjectivizing the image. Released from prison at the beginning of High Sierra, Roy Earle is shown walking through a park, breathing the air of freedom; shots of him looking up are followed by shots of treetops against the sky, with a certain effect of subjectivization insofar as the treetops against the sky are outside the immediate scope of the movement of the narrative and, objectively useless (unlike Lina's telegram in Suspicion), belong only for Roy's character (he was born of a modest farming family and is not the hardened criminal his reputation would have him be). _

[54] Walter Benjamin, Illuminations (London: Fontana, 1970), p. 230; C. Metz, "Le Signifiant imaginaire," p. 35; translation, p. 52.

[55] Edward Branigan, "Narration and Subjectivity in Cinema," mimeographed (University of Wisconsin at Madison, 1975), p. 24. _

[56] ibid.

[57] Discussed by R. Barthes, "Diderot, Brecht, Eisenstein," Screen (Summer 1974), 15(z):38 [included in this anthology—ED]; Branigan, "Formal Permutations," p. 57; and M. Nash, "Vampyr and the Fantastic," Screen (Autumn 1976) 17(3):32-33, 54-60. ~

[58] Burch and Dana, "Propositions," p. 45. _

[59] Salt distinguishes three varieties of field/reverse field and assigns an order and approximate dates for their respective appearances: "It is necessary to distinguish between differnet varieties of angle—reverse-angle cuts; the cut from the watcher to his point of view was the first to appear; the cut from one long shot of a scene to another more or less oppositely angled long shot, which must have happened somewhat later—the first example that can be quoted is in R0verens Brud (Viggo Larsen, 1907); and the cut between just-off-the-eye-line _;angle—reverse-angle shots of two people interacting—the earliest example that can be quoted occurs in The Loafer (Essanay, 19l l). "The EaHy Development of Film Form," p. 98.

[60] For details of the introduction and various accounts of suture, see "On Suture," in Stephen Heath, Questions of Cinema (London: Macmillan, and Bloomington: Indiana University Press, 1981), pp. 86 - 101.

[61] See J.-A. Miller, "La Suture," Cahiers pour l'analyse (1966), no. I, pp. 37-49; translation, "Suturc," Screcn (Winter 1977-78), 18(4):24-34.

[62] "Another characteristic of the film image is its neutrality"; Encyclopaedia Britannica (Macropaedia, 1974), 12:498.