Spatial and Temporal Articulations

The terminology a film-maker or film theoretician chooses to employ is a significant reflection of what he takes a film to be. The French term découpage technique or simply découpage with its several related meanings is a case in point. In everyday practice, découpage refers to the final form of a script, incorporating whatever technical information the director feels it necessary to set down on paper to enable a production crew to understand his intention and find the technical means with which to fulfill it, to help them plan their work in terms of his. By extension, but still on the same practical workaday level, découpage also refers to the more or less precise breakdown of a narrative action into separate shots and sequences before filming. French film-makers, of course, are not the only ones to have a term for this procedure. Both English- and Italian-speaking film-makers have a similar term for this final version of the script — called a "shooting script" in English and a copione in Italian — though they always speak of "writing" it or "establishing" it, thereby indicating that the operation the word describes is no more important in their minds than any other in the making of a film. A third French meaning of découpage, however, has no English equivalent. Although obviously derived from the second meaning of a shot breakdown, it is quite distinct from it, no longer referring to a process taking place before filming or to a particular technical operation but, rather, to the underlying structure of the finished film. Formally, a film consists of a succession of fragments excerpted from a spatial and temporal continuum. Découpage in its third French meaning refers to what results when the spatial fragments, or, more accurately, the succession of spatial fragments excerpted in the shooting process, converge with the temporal fragments whose duration may be roughly determined during the shooting, but whose final duration is established only on the editing table. The dialectical notion inherent in the term découpage enables us to determine, and therefore to analyze, the specific form of a film, its essential unfolding in time and space. Découpage as a structural concept involving a synthesis is strictly a French notion. An American film-maker (or film critic, in so far as American film critics are interested in film technique at all) conceives of a film as involving two successive and separate operations, the selection of a camera setup and then the cutting of the filmed images. It may never occur to English-speaking film-makers or English-speaking critics that these two operations stem from a single underlying concept, simply because they have at their disposal no single word for this concept. If many of the most important formal break-throughs in film in the last fifteen years have occurred in France, it may be in part a matter of vocabulary.

An examination of the actual manner in which the two partial découpages, one temporal and the other spatial, join together to create a single articulated formal texture enables us to classify the possible ways of joining together the spaces depicted by two succeeding camera setups and the different ways of joining together two temporal situations. Such classification of the possible forms of temporal and spatial articulations between two shots might seem to be a rather academic endeavor, but to my knowledge no one has previously attempted such a classification, and I believe that it may well open up some important new perspectives.

Setting aside such "punctuation marks" as dissolves and wipes, which may be regarded as mere variations on the straight cut, five distinct types of temporal articulation between any two shots are possible.

The two shots, first of all, may be absolutely continuous. In a certain sense, the clearest example of this sort of temporal continuity is a cut from a shot of someone speaking to a shot of someone listening, with the dialogue continuing without a break in voice-over. This is, of course, precisely what happens whenever a shot is followed by a reverse-angle shot. Although the term "straight match-cut," as is made clear later on in this chapter, refers more specifically to spatial continuity, it is also another example of absolute temporal continuity. If shot A shows someone coming up to a door, putting his hand on the doorknob, turning it, then starting to open the door, shot B, perhaps taken from the other side of the door, can pick up the action at the precise point where the previous shot left off and show the rest of the action as it would have "actually' occurred, with the person coming through the door and so on. This action could even conceivably be filmed by two cameras simultaneously, resulting in two shots that, taken together, preserve an absolute continuity of action seen from two different angles. To obtain as complete a continuity in the edited film, all we would have to do is cut the tail of shot A into the head of shot B on the editing table.

A second possible type of temporal relationship between two shots involves the presence of a gap between them, constituting what might be called a temporal ellipsis or time abridgement. Referring again to the example of someone opening a door filmed by two cameras (or by the same camera from two different angles), a part of the action might be omitted when these two shots are joined together (in shot A someone puts his hand on the doorknob and turns it; in shot B he closes the door behind him). Even the most conventional films frequently use this technique as a means of tightening the action, of eliminating the superfluous. In shot A someone might perhaps start up a flight of stairs, and in shot B he might already be on the second or even the fifth floor. Particularly when a simple action such as opening a door and walking through it is involved, it might be emphasized that the ellipsis or abridgement can occur in any one of a large number of possible variations; the "real" action might span some five or six seconds, and the time ellipsis might involve the omission of anything from a twenty-fourth of a second to several seconds, and might occur at any point in the action. This is equally true in the case of absolute temporal continuity; the transition between shots may occur anywhere. A film editor might maintain that in both cases there is only one "right" point at which to make a straight match-cut or abridge the action, but what he really means is that there is only one place where the shot transition will not be consciously noticed by the viewer. This may well be. But if we are seeking a film style that is less "smooth," that actually stresses the structures that it is based upon, a whole range of possibilities remains open.

This first type of temporal ellipsis involves, then, an omission of a time-span that is not only perceptible but measurable as well. The occurrence and the extent of the omission are necessarily always indicated by a more or less noticeable break in either a visual or an auditory action that is potentially capable of being completely continuous. (A continuous temporal-auditory action, verbal or otherwise, occurring in conjunction with a discontinuous temporal-visual action, as in Jean-Luc Godard's Breathless and Louis Malle's Zazie dans le Métro, is, of course, not at all precluded.) In the previous examples of going through a door or going up a flight of stairs we become aware of the existence of a temporal discontinuity or gap as a result of the spatial continuity having been forcefully enough maintained to allow the viewer to determine mentally that some portion of a continuous action has been omitted and even enable him to "measure" the actual extent of the omission. (Temporal continuity can likewise only be measured relative to some other uninterrupted visual or auditory continuity.) Thus, if a shot transition takes us from one location to another, more distant one without there being any way of relating the two distinct spaces (such as a telephone or some other means of communication), the temporal continuity between them will remain indefinite unless it is preserved through the use of such clumsy devices as successive close-ups of a clock-dial or some convention such as cross-cutting, an emphatic alternation between two actions occurring in two distinct spaces.

A third type of temporal articulation and a second type of abridgement are possible, the "indefinite ellipsis." It may cover an hour or a year, the exact extent of the temporal omission being measurable only through the aid of something "external" — a line of dialogue, a title, a clock, a calendar, a change in dress style, or the like. It is closely related to the scenario, to the actual narrative and visual content, but it nonetheless performs a genuine temporal function, for, even though the time of the narrative obviously is not the same as the time of the film, the two time spans can nevertheless be related in a rigorously dialectical manner. The reader may object that the boundary between the "measurable" ellipsis and the "indefinite" ellipsis is not clear. Admittedly a segment of time abridged in the process of splicing together two shots showing someone walking through a door can be measured rather accurately — namely, as that part of the action that we know must be gone through but do not see, whereas we are less capable of measuring "the time it takes to climb five flights of stairs." However, "the time it takes to climb five flights of stairs" still constitutes a unit of measurement, much as "one candle power" is the amount of light furnished by one candle; this is not at all the case, on the other hand, when we realize that something is occurring "a few days later," as in an indefinite ellipsis.

A time reversal constitutes another type of possible temporal articulation. In the example of someone walking through a door, shot A might have included the entire action up to the moment of going through the door, with shot B going back to the moment when the door was opened, repeating part of the action in a deliberately artificial manner. This procedure constitutes what might be called a short time reversal, or an overlapping cut, such as Sergei Eisenstein used so often and to such striking advantage — as in the bridge sequence in October (Ten Days That Shook the World) — and such as certain avant-garde film-makers have used (see also Francois Truffaut's La Peau douce and Luis Buñuel's The Exterminating Angel. At this point, however, it is worth noting that time reversals, like time ellipses, are commonly used on a very small scale, involving the omission or repetition of only a few frames, as a means of preserving apparent continuity. The preservation of an appearance of continuity is, of course, what is always involved in any conventional use of time abridgement. What we are referring to now, however, no longer involves simple mental deception — that is to say, making an action that is not visually continuous convey a "spirit" of continuity — but the actual physical deception of the eye. When it comes to "match-cutting" two shots showing someone walking through a door, for perceptual reasons which are quite beyond the scope of this book), a few frames of the action may be omitted or repeated in order that the filmed action may seem more smoothly continuous than would have been the case had the shot been picked up precisely where the previous one left off.

The flashback is a more usual form of time reversal. Just as a time ellipsis can span either just a few seconds or several years, so too can a time reversal. The fifth and last type of temporal articulation thus is the indefinite time reversal, which is analogous to the indefinite time ellipsis (the exact extent of a flashback is as difficult to measure without outside clues as is the extent of a flashforward) and the opposite of a measurable time reversal. The reason why the flashback so often seems such a dated and essentially uncinematic technique today is that, aside from its use by Alain Resnais and in a few isolated films such as Marcel Carné's Le Jour se lève and Marcel Hanoun's Une Simple histoire, the formal function of the flashback and its precise relationship to other forms of temporal articulation have never been understood. Like the voice-over, the flashback has remained little more than a convenient narrative device borrowed from the novel, although both have recently begun to assume other functions.

But might not this inability to measure the exact temporal duration spanned by either flashback or flashforward point to some basic and previously overlooked truth? Are not jumps forward and backward in time really identical on the formal organic level of a film? Are there not ultimately, then, only four kinds of temporal relationships, the fourth consisting of a great jump in time, either forward or backward? Alain Robbe-Grillet obviously believes this is so, and in that sense, his and Resnais's Last Year at Marienbad perhaps comes closer to the organic essence of film than it is currently fashionable to believe.

* * *

Three types of articulation between the spaces depicted in two successive shots are possible — apart from, and independent of, temporal articulations, even though they have obvious analogies to them.

A first kind of possible spatial relationship between two shots involves the preservation of spatial continuity in a manner similar to that in which temporal continuity is preserved, although this spatial continuity may or may not be accompanied by temporal continuity. The door example in all three variations is an instance of spatial continuity; in each case, the same fragment of space fully or partially seen in shot A is also visible in shot B. Any change in angle or scale (matching shots, that is, taken from the same angle but closer or farther away) with relation to the same camera subject or within the same location or the same circumscribed space generally establishes a spatial continuity between two shots. That much is obvious. It seems to follow that there is only one other form of possible spatial articulation between two shots: spatial discontinuity — in other words, anything not falling into the first category. This discontinuity, however, can be divided into two distinct subtypes bearing a rather curious resemblance to the two distinct subtypes of time ellipses and reversals. While showing a space different in every way from the space visible in shot A, shot B can show a space that is obviously in close proximity to the spatial fragment previously seen (it may, for instance, be within the same room or other closed or circumscribed space). This type of spatial discontinuity has given rise to a whole vocabulary dealing with spatial orientation, and the fact that such a vocabulary should be necessary serves to emphasize how essentially different this type is from an obvious third possibility, complete and radical spatial discontinuity.

This vocabulary dealing with spatial orientation brings us to a key term, one of some concern to us here: the "match" or "match-cut." "Match" refers to any element having to do with the preservation of continuity between two or more shots. Props, for instance, can be "match" or "not match." On a sound stage one can often hear remarks such as "these glasses are not match," meaning that the actor was not wearing the same glasses or was not wearing glasses at all in a shot that has already been filmed and is supposed to "match" with the shot at hand. "Match" can also refer to space, as in eye-line matches, matches in screen direction, and matches in the position of people or objects on screen. There are also spatiotemporal matches, as in the door example, where the speed of movement in the two shots must "match," that is, must appear to be the same. To clarify this notion of "match" or "match-cutting," a brief history of how it developed is in order.