In Plato, art is mystification because there is a heaven of ideas; but in the earthly domain all glorification of the
earth is true as soon as it is realised.
Simone de Beauvoir, The Ethics of Ambiguity
[trad. Bernard Frechtman. Citadel Press, New York, 1962]

 
 
Recently I cut a film about two women and their suburban London ménage. After the completion of the editing, I came to know these people personally, and visited them in their home. When the film was eventually transmitted on television, I found myself perceiving it in an unnervingly bifocal manner. To the extent that I fed into the images my subsequent knowledge of the characters and location, the film broke down into incoherence. To the extent that it did cohere, it projected a world that repudiated any connection with the people and place as I now knew them.

 
Such an experience is the subjective correlative of the dual nature of film, which exists both as record and as language; and this du­ality, a source of paradox, has generated much confusion in the debates surrounding “observational cinema.”
 

 
The problems we face are not new in principle; but the terms in which they must be resolved have changed radically within the past fifteen years. Beneath the many styles in which documentary has historically manifested itself may be discerned a common purpose: to enable the character of film as record to survive, so far as is possible, its metamorphosis into language. But the development of lightweight cameras and tape recorders, mobile and relatively inconspicuous, has brought documentary to a cri­sis which is largely one of confrontation with its own being. It is as if, having long watched itself approaching, it were now almost near enough to reach out and shake itself by the hand. Almost, but never quite.

 
The crisis has been most acutely registered in that realm of subject matter which may be termed the domestic. At a calculated risk of question-begging, I shall offer a filmic rather than an ethnographic definition of the domestic realm as comprising those areas of human activity which are (a) most difficult to shoot without interference, since they are not public, and (b) least sus­ceptible to reenactment, since they are not in any simple way repetitive. The purpose of this somewhat negative phrasing is to denote an area of personal, intimate though not quite clandestine behaviour in explicit contrast with those repeatable or public events which have formed documentary’s traditional sources— events on which the paraphernalia of filmmaking has been as­sumed not to exercise any significant influence. Repetitive ac­tions—such as the operating of a lathe—can be performed in front of the most cumbersome of cameras once the subject has overcome self-consciousness. With public events—such as po­litical speeches—it is different. We cannot say that the presence of the camera necessarily has no effect, for there is a whole cate­gory—once known fashionably but misleadingly as non-events —where the attention of the media is held responsible for their very existence. An early example might be the 1934 Nazi Party convention, which, we are told, was organised largely with a is in the nature of public events to change their nature according
to expectations of public response.

 
It was a consequence of the character of these traditional source materials—materials which contained already, in their be­ing rehearsable and/or audience-directed, something of the qual­ity of such events as exist only for the purpose of being filmed and of being wholly assimilated into the fabric of a discourse—that documentary seemed, in its classic period, to be concerned only with general truths of the human situation as opposed to the “in­dividual psychology” of acted fiction: dubious distinction being thus implied between people as individuals and as social crea­tures. It is not surprising, then, that the prime enticement offered by the new equipment should have been the possibility of open­ing up domestic behaviour to the record without first translating it into something other than itself (i.e., the scripted behaviour, or at least improvised performance, into which classical documentary had sometimes gingerly ventured in an attempt to transgress its limits); that the problems entailed should have been seen as clustering around the ideas of “influence” or “distortion” within the pro-filmic; and that those immediately exercised by such problems should have been the fieldworkers—directors, camera operators, and, above all, anthropologists already committed to a notion of academic “objectivity.”

 
While most practitioners would claim to be aware that film is not an open channel to “reality,” they have nonetheless tended to see the difficulties as centring on their own attitudes and the fear that these may somehow mar the neutrality of the material. Many of the prescriptions of ciné-anthropologists have rested upon an assumption, implicit or explicit, that the inevitable se­lectivity of shooting may be counteracted—or perhaps merely atoned for—by a refusal of selectivity in the editing: that the minimum of structuring will afford the maximum of truth. But the antithesis of the structured is not the truthful, or even the ob­jective, but quite simply the random. We may, indeed, find it a little puzzling that anthropology should have taken so readily to observational modes of filmmaking at all, when it might, as a sci­ence, have been expected to prefer the traditional patterns whose tight organisation of protofictional (or at any rate protodemonstrative) materials promises to raise the general principle above the vagaries of the particular instance. (Flaherty’s method, for ex­ ample, with its heavy use of reconstruction, may be said to have been based on the idea that the better we familiarise ourselves with the lifestyles of our subjects, the more of the elements of those lifestyles will be gathered into the realm of the repeatable —repeatable by virtue of our-the-filmmakers’ understanding of what constitutes their essentials.)

 
All events, at least in human affairs, are events perceived by somebody. What may seem to be at stake, from the filmmakers’ viewpoint, is to prevent their own perceptions from intervening between the viewer and the pro-filmic. But the filmmakers’ intentionality is not within the viewers’ ken; and their abdication of it will, conversely, do nothing to prevent the camera from transposing the world into what will function for the viewer as imagery. It is to the nature of this imagery that we must look for clarification.

 
The photograph is a physical imprint of the world. Like the pho­tograph, film stakes a claim on reality which has nothing to do with “realism” in any literary sense; and it is this claim which documentary aspires to fulfil. If it has proven notoriously diffi­cult to define documentary by reference to its constantly shifting stylistic practices, it is because the term “documentary” properly describes not a style or a method or a genre of filmmaking but a mode of response to film material: a mode of response founded, upon the acknowledgment that every photograph is a portrait signed by its sitter. Stated at its simplest: the documentary re­sponse is one in which the image is perceived as signifying what it appears to record; a documentary film is one which seeks, by whatever means, to elicit this response; and the documen­tary movement is the history of the strategies which have been adopted to this end.

 
A crucial fact about the definition of documentary as a mode of response is that it places the attribution of documentary significançe squarely within the province of the viewer. Any fiction film can—albeit with difficulty—be perceived as a documentary on its own making. Of course, it is perfectly possible for a film to “be” a documentary in that it will achieve greater coherence on this than on any other interpretation; and it is certainly more than possible for the makers to have intended it in this sense. Bu the fact remains that documentary is what we-as-viewers can perceive as referring to the pro-filmic, it being supposed that we can thus construe it as meaningful. However, if we are to treat documentary as so defined by the viewer’s perception, we must face up to certain difficulties which this raises concerning the identification of the images.

 
The photograph—once we are sure that it is a photograph— cannot lie. But it can be wrongly labelled. This so-called photo­graph of King George VI is in fact a photograph of someone else. If we accept that documentary is best defined as a way of per­ceiving images, we cannot evade the implication that it is blind to the falsity of labels. Documentary will be consequent upon what it appears to show, rather than upon what it necessarily does show; and the relationship between the two is a matter for the filmmakers’ ethics, inaccessible to the viewer. Yet the assump­tions which the viewer makes about this relationship, on the ba­sis of signals intended or unintended, will inform his perception of the film. To make a documentary is therefore to persuade the viewer that what appears to be is.

 
Film is not unique in this demand for a response based upon an ethical assumption which it cannot substantiate. Jazz, for ex­ample, may be considered documentary music in that it claims to be predominantly improvised and therefore to represent the ac­tual moment of musical creation. Likewise functional architec­ture, in its requirement that appearance should express use and structure, is assumed not to cheat. What appears to be a load­ bearing pillar must not be simply the outer facing of a ventilation duct. It is possible for documentary to be labelled as such in a bald, extrinsic manner with some caption or billing or publicity which, whilst overtly just requesting us to construe the images in a documentary sense, carries also the implication that they may be taken on trust. (Here again, Flaherty may offer us an example: for the myth of Flaherty, to which he never quite lived up, was taken as an extrinsic tag of “trustworthiness” under whose cre­dentials he felt able to engage in a degree of manipulation of the pro-filmic which most people today would consider a transgres­sion of anthropological ethics.) Most filmmakers, aware of the implicit circularity of this—“Believe me when I say that I’m tell­ing you the truth”—prefer to rely upon intrinsic signals as indi­cations of how the film is to be understood. The labels are swal­lowed. The ethics become buried in the style.

 
Documentarists, tracing their genealogy proudly from Lu­mière, like to believe that documentary is the “natural” form of cinema. But fiction film, like painting and literature, rests no spe­cial claims upon the provenance of its linguistic elements. It must surely be clear that it is documentary which is the paradoxical, even aberrant, form. True, the first films were of a “factual” na­ture; but the medium was not out of its immobile, one-shot in­ fancy when someone saw the possibility that it might be employed to signify something other than that which it recorded —this step being taken by Méliès, a prestidigitator. From now on, it was nonfiction films which were to be distinguished by a special name: actualities.

 
The problems latent in the idea of actuality become com­pounded at precisely the point where this name becomes inad­equate and must be replaced by the more evasive one, “docu­mentary”: the point at which the primordial image becomes articulated as language. (It is interesting, though futile, to speculate on how differently the syntax of film might have developed had Méliès’s irreversible step—the original sin of cinema—not been taken.) The difficulties are twofold. Firstly, this is the point at which it becomes possible for the articulations to be used— and perhaps inevitable for them to be perceived—as implicit in­ dications of the nature and status of the component images. Sec­ondly, there is clearly no sense in which the one-to-one relation of shot to prior event may be said to hold good for a structured sequence, let alone for an entire film. The activity of an editor is in this respect more akin to that of a painter than to that of a photographer, whose apparatus may be held to ensure a certain “objectivity” even when it does not produce a good—or even a recognisable—likeness. In what sense, then, is it meaningful to claim any privilege at all for such structures?

 
These two difficulties—the status imputed by structures to their elements and the status claimed for the structures them­selves—may be seen as the “vertical” and “horizontal” aspects of the problem. They are not entirely separable, as one example will suffice to demonstrate. Consider a group of synchronous shots taken from various distances, with or without simultaneous cam­eras, at a cricket match. An essential factor in this record, as in­ deed of the “atmosphere” of the occasion, will be the delay be­tween sound and vision as the ball makes contact with the bat. This delay will vary in proportion to our distance from the bats­man. If, then, we wish to cut from midshot to long shot immedi­ately after the batsman’s stroke, in order to show where the ball goes, the sound on the former shot will be almost instantaneous with the image, whilst the sound on the latter will be noticeably retarded. Thus the sound on the former may occur before the cut, whilst that on the latter may occur after it: in which case we should hear the impact twice. Conversely, if we cut from long shot to midshot to follow the subsequent movement of the bats­man, we ought not to hear it at all. What an editor will in fact do is to iron out these anomalies by laying all the sounds in ap­parently simultaneous synch with the pictures. In doing so, how­ever, he is allowing a syntactic relation between two elements— a relation whose claim to documentary privilege might properly be discussed in the context of alternative “horizontal” strategies —to determine a shift in the actual relation of the image to the pro-filmic. (It need hardly be added that the problem becomes more vexed when we are dealing not with true variations in cam­era position but with variations in the setting of a zoom lens from a static viewpoint.)

 
With this proviso in mind, however, let us examine first the “vertical” question. To face it directly: when, in our watching of a film, do we “know” that it is a documentary—i.e., that its constituent images are to be read in a documentary sense—and how? What is the nature of those buried stylistic signals to which we are responding? The answer to this question will vary from film to film and also, perhaps, from person to person; but what is probable is that these signals, in fact no more than reading instructions preferred to extrinsic labels for their very inexplic­itness, will take on, through our-the-viewers’ acquiescence in them, the misleading and contentious character of guarantees of the film’s veracity. The reasons for this lie in the nature of the documentary idiom.

 
Just as jazz and functional architecture are identified by cer­tain characteristics consequent upon their mode of manufacture, so most of what one might designate the stylistic indices of documentary have always been the product of practical constraints: the physical difficulties of filming, the demands of a theme not wholly tailored to the pregiven technology, or simply the usual inadequacy of the budget. These find their objective expression in: a predominance of location shooting; a graininess due to the absence of studio light; a customary linking of diverse im­ages through verbal narration; a woodenness in rehearsed performances due to the use of nonactors, and a consequent ten­dency for rehearsed material to be broken up into short units; a wobbliness in spontaneous material due to hand-holding of the camera, and a toleration in such material for temporary lapses of focus or framing and for imperfect continuity in the cutting of supposedly matching actions; and a tendency to define complementarity of shots (in a crosscut conversation, for example) ac­cording to the framings obtained by panning the camera from a fixed position.

 
In fiction film, by contrast, the stylistic characteristics are not directly determined by the exigencies of production. Thus, for example, a crosscut conversation will typically be constructed from an alternation of over-the-shoulder shots involving a relo­cation of the camera, as often as not in midsentence. Generally speaking, classic fiction style wilfully avoids the construction of its space from what might be thought credible positions for one casual observer. It is as if the very arbitrariness of the conven­tions were itself an index of the fictive disposition of the ele­ments—a signal that what actually happened during filming had no priority over the fact of filming, and need not concern us.

 
But if the idioms of documentary are not arbitrary, that cer­tainly does not mean that they are trustworthy. There is scarcely one of them (the only obvious exception being the unactorly performance) which has not at some time been appropriated by the fiction film, in which context it becomes an arbitrary signifier of realism. Indeed, it is difficult to think of a style of realism in fiction which does not call upon some quality thrust upon non­fiction by circumstance. It is almost as if verbal realism were to find it necessary to sport the misspellings, erasures, and shaky grammar of a hastily jotted dispatch. But the traffic has, of course, been two-way. Documentary has usually sought, in its more pre­meditated sequences, to approximate to the arbitrary conven­tions of fiction, since these constitute, after all, our received and shared film language for the articulation of time and space. The strangest transactions, however, are documentary’s borrowings from itself. We have all seen the shot where a reporter knocks on a door and is admitted, camera atilt, mike in picture, and hand­ held lights slewing drunkenly, to a house whose occupants show not the slightest surprise at this invasion. At this point the docu­mentary ethic is buried in more senses than one. (No doubt this sort of thing happens less in academic filmmaking than in television; but only in the most Utopian of fantasies can we neglect entirely the possibility of fraud.) What this amounts to is an at­ tempt to invest documentary realism with literary realism: a superimposition of idioms which, ethics aside, can lead only to a blurring of perception.

 
What we have seen in recent years has been a narrowing of the gap between the languages of documentary and fiction. On the documentary side, this convergence has resulted from precisely those technical developments which have been hailed as opening up the domestic realm to scrutiny, and from the skills developed by technicians in response to their challenge. Those same light­weight, silent-running cameras and recorders, plus film emulsions whose sensitivity obviates the need for extra lighting in most situations, have led to the production of films whose fluency of camerawork and naturalness of “performances” rival the most polished of studio achievements; and the high shooting ratios now frequently available for such filmmaking have opened up an unprecedented range of stylistic choice for the editor.

 
What “observational” technology in fact offers us, when we turn our attention from its mere data-gathering aspects, is a film grammar almost free of such technical limitations as have in the past characterised documentary, and have indeed virtually de­fined it in contradistinction to fiction. If documentary is a form fractured by paradox, what are we to think of a fulfilment whereby it divests itself of what have hitherto been, for the viewer, its de­finitive and delimiting features? There have, after all, been ear­lier casualties of the discovery that a stylistic label guarantees nothing: witness that “immediacy” which, as a cant word in the formative years of television, carried an implication that the qual­ity of being immediate, rather than simply of looking spon­taneous, might magically be conveyed across the airwaves. This concept did not survive the introduction of videotape recording. Does documentary face a similar dissolution?

 
Some people would argue that any distinction between docu­mentary and fiction diminishes rapidly toward zero as a film in­ creases in complexity, and that such problems of language and la­belling as have been seemingly posed or exacerbated by recent technical advances are no more than the reaffirming of this un­derlying contradiction. If they are right, then documentary is in­deed a false calling. It is here that we shift from the “vertical” to the “horizontal” dimension: from the question of the labelling of elements to that of the relation of a broad text (and, by implica­tion, of the elements considered as subdivisions of that text) to the pro-filmic. For to make documentaries is to engage with pre­cisely this paradox: to seek structures in which the aspect of film- as-record may retain (or perhaps reassume?) its significance.

 
The definition of documentary in terms of the viewer’s percep­tion—i.e., as material perceived as signifying what it appears to record—avoids some of the wilder consequences of attempted definitions in terms of productive method, which, with their stress upon the status of the pro-filmic, inevitably end up in dis­putes about whether this or that reality is as real as it might be. It is in this light that we may understand why the appropriation of the domestic realm—the conquest of an area hitherto “acces­sible” only to fiction—should constitute both a fulfilment and a crisis for documentary. The crisis, unfathomable from either of the extreme viewpoints grounded in the conception of film as either purely language or purely record, is one not primarily of field ethics—of how to register behaviour with the minimum of intrusion—but of how to structure material so that whatever it may become as experienced language may in some sense keep faith with its character as pro-filmic fact: for, whilst the attempt to ensure that things are what they seem will present itself to the location crew as an ethical problem, the implications will be pres­ent to the viewer only insofar as they manifest themselves in terms of the aesthetic. In the sense that the crisis is located in the transition from the filmmakers’ terms of reference to the view­er’s it is a crisis of editing, for it is at this interface that the editor works.

 
If, as has been suggested, the repetitiveness and public nature of the elements of classical documentary may be said to mark the outset of their own self-symbolism, it is clear that the reality con­trived upon such elements must take on the character either of a generalisation from a multitude of possibilities, superimposed and scarcely distinguishable, or of a symbolic entity toward which the pro-filmic was already in incipient project. Even when in­dividual behaviour was essayed, these guidelines were usually respected. The stilted speech of the ground crew awaiting the safe return of F for Freddie was a clear signal (an “intrinsic label”) that these were to be construed not as characters (acted) but as people (acting): that this record of their actions, though struc­tured according to the familiar canons of realism, was to be seen as articulating a demonstration in the Brechtian sense—a generalised statement about their own lives. Now, deprived of the sup­port of those conventions which no doubt at the time seemed the most grievous of limitations, condemned to a treacherous fluid­ity by the spontaneities of observed behaviour, documentary must ask itself what may be the structures which will arrest the settling of this fluidity into the ever-accommodating moulds of rhetoric.

 
Seen in this “horizontal” perspective, however, the prospect is not encouraging. A mere juxtaposition of two shots will dem­onstrate the proliferation of the difficulties. In an American Sen­ate committee hearing, a senior public official faces accusations of misconduct. As he listens to the development of a hostile line of questioning, we cut to his hands playing nervously with a pa­per clip, the voice of his interrogator continuing offscreen. This simple cut opens up a range of possibilities:

 
1. The sound continuity is genuine, and the second shot is also genuinely synchronous—i.e., it was taken with an­ other camera.
2. The sound continuity is genuine, but the second shot is a “cutaway” taken from a similar context elsewhere in the rushes.
3. Both shots are synchronous, but taken with one camera; and there is a concealed cut (perhaps not exactly matching the picture cut) in the sound.
4. The second shot is a cutaway designed to conceal the fact that the sound is discontinuous.
5. The cutaway is taken from a dissimilar context elsewhere in the rushes.
6. The cutaway was reenacted after the event by the charac­ter in the film.
7. The cutaway was set up with an actor.
 

 
And so forth. These may be said to represent varying levels of ethical probity—assuming, that is, that some ethical claim has been implicitly made—or, alternatively, to represent a range of statements whose verbal equivalent would be somewhat as follows:

 
a. This man is made nervous by this question.
b. This man is made nervous by this type of question.
c. This type of man is made nervous by this type of question.
 

 
But the implicit signals whereby we-as-viewers evaluate the material will, even if we consent to trust them, be too coarse to enable us to discriminate between such alternatives. Not only will we be unable to judge the truth of the film’s statement; we will be unable even to determine with any precision what sort of statement it is. However carefully the filmmakers may have matched the nature of the cut to the nature of the statement they wished to be seen as making, however scrupulously they may, under the banner of vérité, have selected their options only from the upper reaches of the above lists, neither the sincerities nor the choices made in sincerity will be truly legible in the product. From the standpoint of the viewer, the practices of “observational” cin­ema cannot be legitimised by the ethics of fieldwork. Yet it is only from the standpoint of the viewer that documentary can be understood.

 
How, then, are we-as-viewers to take such a sequence? If we perceive it as a transparency (selecting, that is, from the upper parts of the above lists), we stand in danger of being cheated; if we perceive it as an argument (i.e., selecting from the lower parts), then the aspect of film as record, with its implication of uniqueness and contingency, dwindles into insignificance and the particular becomes only the exemplar of the abstraction it ar­ticulates. The latter option leaves us little better off than we were in the days of classical, 35mm documentary.

 
This problem can only be compounded by the use of film as a mode of disquisition (and I am not convinced that retreat into the idea of film as mere research tool is any more than an evasion of it) in anthropology, whose whole bias is to regard the pro-filmic —in the form of ritual, kinship, mythic representation, every­ day behaviour—as itself a structure for the articulation of social meanings. If an event is itself an element in a linguistic structure,what is the status of a film which linguifies it into an element of an altogether different discourse—and, worse still, tries to make this discourse accessible to those who do not speak the original “tongue”?

 
Consider another example of what occurs in the editing pro­cess: an example taken, for simplicity’s sake, from our own cul­ture. The event is an argument between two people in a small office. The argument lasts for an hour and a half, out of which the camera is running for approximately fifty minutes, panning be­tween the two people and occasionally shifting position to favour one or the other. The attitudes of the characters emerge only slowly, with much repetition of key points, though with a grad­ual overall rise in the emotional temperature. The relevance of this sequence to the theme of the film does not justify allowing it more than about twelve minutes of screen time: but this, irksome though it may be, does no more than stress the inherent selec­tivity of cutting. The problems confronted by the editor there­fore present themselves in the form of such questions as: Does this sequence, whilst avoiding undue repetition, reproduce the spiral nature of the interaction? Does it do justice to both posi­tions whilst also conveying the subtleties of the psychological tactics in use on both sides? Does it respect the integrity of each participant, in the sense of not allowing a change of emotional state to appear unmotivated (so that anger, for example, might come across as mere petulance), or of presenting someone’s line of argument in a form less rational—or even more rational— than that which it took in the actual debate? And beneath these, of course, lie more fundamental questions. Are we trying to be fair to the people as individuals or to their strategies, or to the in­stitution of whose ethos they are the temporary embodiment?

 
Are we seeking to demonstrate the arguments used or to clar­ify the actual intellectual positions which these partly express but partly mask? Is the tedium of the event’s repetitiousness— a quality it may possess only for the observer uncommitted to either viewpoint—something which should be retained or avoided? Such things need to be asked not just in a liberal spirit of impartiality, but in an attempt to ground the film’s putative meanings—even, where appropriate, its polemic—in a correct apprehension of the world. Throughout the process, the editor is engaged in a curious mental exercise: to attempt, from the rushes and from the testimony of those present, to form an intuitive im­pression of the event as if it were firsthand experience (granted that all such experience is itself partial and selective); and then, in settling upon the presumptive happening “behind” the mate­ rial, to use that material to say it.

 
The difficulty with all this lies not in the fact that it is a mat­ter of personal judgment (as of course it is), nor that people are having scripts written for them out of their own words, and per­formances drawn from the repository of their own unguarded gestures (for it cannot be otherwise), nor even that, in the pro­cess of compression, certain “themes” may have to be sacrificed to the coherence of others (the given becoming the data); but in something less obvious. In our anxiety to keep faith with the event as we have understood it, in our careful attention to the mutual respect of the two characters and the behaviour appro­priate to their relative status, in our assessment of the degree of eye-to-eye contact which may be permitted across a given cut, the alacrity with which one party may press home an advantage or the pause to be allowed before the other climbs down or gives way to fury, how much interruption is permissible, how much non sequitur will be tolerated—in all these judgments we are re­sorting to a multiplicity of social behavioural codes: and we are resorting to them not in the sense of requiring to encode them as pro-filmic matter to be demonstrated, but in the sense of recruit­ing them for use as filmic codes in the articulation of the sequence.
The dangers of such a procedure in the portrayal of unfamil­iar societies, whose codes of body language, etiquette, and proto­col lie outside the competence of editor and viewer, need scarcely be emphasised. I have myself required expert help in the edit­ing of a conversation between two Tuareg, where the constant microscopic adjustments to the height of the veil carry a great weight of social significance. But it is not simply a question of “getting it right.” What was in effect happening here was that we were allowing the known conventions of dialogue cutting to feed significance back, as it were, into the gestures which had cued the cuts, so that these gestures might be understood retrospectively by the untutored viewer. And the deeper danger lies precisely in the elision of such known conventions—concerning matching of eyelines, continuity of dialogue over cuts, and so forth—with the capacity of film to swallow the world into its own syntax, since these conventions have been developed specifically for the artic­ulation of the world as realist fiction.

 
A more extreme example may be taken from a film about a video-dating service. Here, in a sequence where a man watches a videotape of a potential girlfriend, I found myself able to impart an emotional coherence and momentum to the action by em­ploying intuitively understood conventions of eye contact and its avoidance; but in this case the interaction thus narrated from the documentary elements bore a purely illusory relevance to those elements, since the glances of the man were directed not at the girl herself but at a television screen, whilst those of the girl had been directed not at the man, nor even at her unseen interviewer, but at a video camera. What was happening in this sequence, then, was the deliberate narrative prefiguring of a relationship which was to develop later in the story—an excursion into that treacherous terrain between documentary and fiction which it was our conscious intention in this film to explore. But, whilst one might hesitate to use such methods in a film claiming strictly observational status, we must face the fact that the alternative to using these social codes to imply a relationship might have been to allow them, by default, to deny one.

 
To put it crudely: we have learned that, when a midshot of someone lowering a cup is followed by a close-up of a cup being placed on a saucer, this is to be construed as signifying not two separate actions but one continuous one. Such a convention— some twenty years younger than the cinematograph itself—func­tions to efface the pro-filmic (where there actually were two ac­tions) in the projection of a hermetic reality, a closed diegetic world defined only by the narrative which calls it into being to in­habit it and, though masquerading as everybody’s world, intelli­gible only from the standpoint of its narrator.

 
Furthermore, it is arguable that recourse to such conventions serves not only to define a quasi-literary realism but to inhibit the documentary response altogether: for it is difficult to see what, other than their cumulative presence, may be said to alert the viewer to the fact that material is to be construed in a fictional mode—that we are to see not Christopher Lee mounting some plaster steps in a low-key-lit studio but Count Dracula returning to his castle. Just as, in the documentary response, failure to con­strue an element as signifying what it appears to record leads to a breakdown of meaning, so, in the fictive response, perception of an element as recording other than what it appears to signify is consigned to irrelevance. In other words, it is not merely inci­dental to but definitive of fiction that its nature as record should be negated; and one of the functions of received film syntax is precisely to ensure such negation.

 
In our case, however, the fiction is not a true fiction, acknowl­edged by the viewer as such, but the insinuation of a demonstra­tive position—a pseudo-narrative omniscience achieved by the elision of filmic and pro-filmic languages—into the viewer’s con­struction of the film’s meaning: a replacement of documentary’s once frank demonstrativeness by a demonstrativeness covert and shifty: an attempt to pass off a metareality as a reality of some transcendent make. The viewer will refer the film back to the events filmed; but such return will constitute not the fleshing out of understanding from the fund of his experience but the super­imposition of fictive representation upon the disorder of the pro- filmic. One might posit a form of cryptofiction: a mode in which, though the relation of the material to a prior world is intellec­tually acknowledged, such relationship remains marginal or ir­relevant to the meaning we-as-viewers attribute to the whole. Conversely, it is possible that such meaning may be reinvested in the world in ways the material does not warrant: that an argu­ment may be concealed within the structure of the diegesis such that our construction of the latter will entail endorsement of the former.

 
 
Ought we then to regret the invention of the equipment which, erasing the signatures of documentary’s previous (if lim­ited) credibility, have enabled us to weave our dubious assertions into the semblance of unanswerable reality? Perhaps. But the problem as we have stated it offers at least a clue to its own solu­tion, hinting at strategies which must be adopted if documentary is to escape its former demonstrative limitations, its confinement to the general, without being willy-nilly sold into realism. What are needed, broadly speaking, are methods whereby the various strands of discourse—the referential nature of the images, their demonstrative disposition, the construction of narrative conti­nuities in time and space, the filmic and extrafilmic codings— may be denied elision and offered as separable to the viewer’s scrutiny. Many approaches are possible. Indeed, the creativity of future documentary must consist largely in exploring them. One—perhaps not too promising—might consist in pushing the received conventions to the point of parody so that, whilst still functioning to articulate the material, they would be perceived in their arbitrariness. Another would be to employ methods of selective jump-cutting whereby one theme—the logic of an ar­gument, perhaps, or the local narrative of one character’s ac­tions—would be articulated as a continuity whilst the remainder respected the discontinuity of the pro-filmic. Where commen­tary must be used, it might take the form of two voices disagree­ing with each other as to the proper interpretation of the evi­dence. If cutaways are needed, perhaps they should be graded differently or wear different clothing from synchronous shots of people listening, as a mark of their different grammatical status. (I once heard an anthropologist-filmmaker being taken to task by a colleague for having constructed a bargaining sequence by in­ tercutting material of a stallholder and customer shot on differ­ent days. A distinction in grading between the two components might have satisfied the requirements of both differentiation and synthesis.)

 
An interesting pendant to these considerations is that even the received conventions of fiction are not immutable, and do not always retain their significance when the responsive context is changed. This fact may act against us or in our favour. On the one hand, I have made the mistake of trying to use a dissolve in a documentary for its classical fiction purpose of indicating a telescoping of narrated time (which in documentary means a telescoping of pro-filmic time), only to find that viewers simply failed to read it in this sense—or even to notice it at all. On the other hand, it seems likely, once the documentary response is firmly established, that a continuity cut which in fiction would certainly signify a single action may take on the force of a jump cut signifying two.

 
Some of the above suggestions may, I confess, have their ori­gin in the editor’s preference for the scratched, unmatched pic­ture and nonequalised sound of the cutting copy over the harmo­nious fluency of a finished print. However, I am not here trying to outline all the possibilities but merely to suggest that the pu­ritanical caution advocated by many anthropologists as an ap­propriate attitude to their records is unnecessary. There are cer­tain jokes which we assume to be true stories because there would be no point in telling them if they were not. Perhaps filmmakers should aim toward forms of construction which, in achieving this sort of credibility, would circumvent the need for stylistic labels. Back in the early 1950s I saw a television programme about George Bernard Shaw, doubtless shot with the full panoply of 35mm equipment and its attendant hordes of technicians. A mid­ shot of Shaw sitting down at his typewriter was followed, in clas­sic fashion, by an over-the-shoulder shot as he began to type. What he typed was, “I don’t normally behave like this at all.”

 
We have only to look at some of the best in ethnographic film to see both the meaninglessness of claims to supralinguistic ob­jectivity and the subtlety of the methods whereby documentary may yet insist upon the pro-filmic. In a film such as Lorangs Way, by David and Judith MacDougall, we may perhaps respond first to its formal discipline and to the sureness of its control. We may be inclined to point out that the long-sustained shots, whose ex­clusion of offscreen activity becomes at times almost oppressive, constitute as much the imposition of a filmic rhythm as would constant cutting to the objects of the characters’ gaze. Yet it is also true, not only that the presence of the camera is acknowl­edged in each sequence, but that it is acknowledged by successive sequences in contrasting ways—as companion during a walking track, as interlocutor during a formal interview, as casual nui­sance in an occasional verbal jibe, as maintaining-a-respectful- distance in a long telephoto hold—and that the juxtaposition of these modes of acknowledgment sustains a constant interroga­tion of the status of the film unit, both technical/aesthetic and so­cial, which inhibits the semantic closure to which realism con­stantly tends. And it does this, furthermore, in a way which can have no precise equivalent in fiction.

 
In short, then, we must beware of treating film from an Olym­pian standpoint as if, solely by declaring ourselves outside the process, we were able to span that dichotomy between filmmak­ers’ perspective and viewer’s perspective which is the concrete presence of film. We may ask ourselves only what options face the filmmakers for the structuring of their perceptions as re­corded, and what options face the viewer for the deciphering of this structure as a text. Indeed, as an extreme statement of this position one might suggest that the filmmakers’ responsibility is fulfilled in an attempt to reverse the polarity of the filmic signs so that their sense is toward their material sources—their object matter understood as their subject matter. Certainly it is the case that, since documentary reality is the viewer’s construction, any suggestion of interpretations being “forced on us” represents an abdication of responsibility by the viewer. Wilfully or by over­sight, some materials may be wrongly labelled. Some things may have been less rehearsed or more rehearsed, less spontaneous, less calculated, less uninfluenced by the camera’s presence than we-as-viewers suppose them to have been. But there is no sharp demarcation between the misunderstandings of documentary and the misunderstandings of life. And the documentary stance is essentially one of interrogation.

 
We have suggested ways in which complex structures may draw upon, rather than efface, the aspect of film as record; but we have not yet found a justification, within the viewer’s perspective, for those “observational” practices which, in the filmmakers’ perspective, represent themselves as exploring the potential of the new technology in the appropriation of the domestic realm. (By “observational practices” I mean the extreme precautions frequently taken to minimise the crew’s influence upon the pro- filmic and/or, within the acknowledged fallibility of the medium, to signal the degree of such influence to the viewer: such prac­tices as lengthy acquaintance with the subjects before the start of shooting and/or the inclusion of crew members or camera gear in the shots.) Anthropologists have been mistaken in imagining that vérité might herald a return to the prelapsarian innocence of the actuality. But have they been altogether wrong in the meth­ods they have evolved in this belief?

 
Let us look again at the dichotomy film-as-record/film-as- language with whose subjective aspect we began. If documentary were merely record, then editors would not be needed to order it, since to grant significance to the order in which records are presented is to impute to it a linguistic nature; yet if documen­tary were language pure and simple, editors would not be needed to manipulate it, since there would be no meanings generated other than those commonly available—to film crew and viewers alike. Clearly these twin aspects of the medium are not to be un­derstood as alternatives: for, quite apart from its possible articula­tions as fiction, the record can function as record only to the ex­tent that it is legible as such. Of course, there may be mementos so poignant and personal that the linguistic element scarcely rises even to the rudimentary; and we may well, at the other extreme, care to ask exactly what “record” can mean to someone who was not present at the original event. Whether at the cellular level of the photograph or the complex one of structured communica­tion, whether in one-to-one relationship with the pro-filmic or otherwise, an act of decipherment is what is called for. As view­ers, in electing to perceive a film as documentary, we do not re­ject a fictive option for a known nonfiction, but rather select a mode of apprehension in full knowledge of our own ignorance.

 
There are in The Battleship Potemkin sequences in which min­imal human activity—sailors asleep in hammocks, others on watch—is intercut with details of the ship and with seascape ma­terial such as buoys, gulls, other vessels, and views ashore. At our present distance in time we may construe such material as a por­trayal of Odessa harbour either in 1905, when the story is set, or in 1925, when the film was made. This distinction, between the fictive and the documentary construction, is subtle but absolute: and it would remain absolute even if the action of the fiction had been set in 1925 also. It depends not upon our prior knowledge of the look of Odessa at these dates but upon the mode of our in­terrogation of the images. A reality, like a fiction, must be articu­lated. But the difference between an account and a narrative lies, for us-as-recipients, in the relationship we posit between the lin­guistic elements and the total discourse. In fiction, the elements are exhausted in the production of the overall meaning of the text; and anything which cannot be read as contributing to this meaning is consigned to a limbo of insignificance. In documentary, by contrast, the elements are seen as always exceeding their contribution to any given meaning; and they remain always open to scrutiny either for their own sakes or for their potential in the generation of new meanings oblique, peripheral, or even antagonistic to the text as understood.

 
As will be recognised, the analogy of an “account” is inade­quate beyond a certain point: for in a verbal account of some­thing—a piece of journalism, say—the excess by which we in­terpret the prior world as outspanning its description is present to the reader only as a subjective coloration to the words, a sense of their striving after what they cannot wholly comprehend; whereas in documentary that excess is present in the images, in their potential always to reveal—under different interroga­tions—aspects of the pro-filmic hitherto unremarked. In this re­spect, then, the verbal affinities of documentary are with poetry, since the resistance of its elements to total absorption in the dis­course represents a resistance against the drift toward pure symbolism. Verbal poetry seeks to dissolve the symbolic fixity of the word by undermining the rules of syntax—either relaxing them, as in free, surreal, or rhapsodic verse, or subordinating them to other and more arbitrary structures (alliterative, metrical, quan­ titative, syllabic… ) which, whilst preserving, mock them. Freed of the absolute grip of those grammars with which its symbolic precision is held in mutual support, the word offers the full range of its ambiguities and connotations for the generation of a flux of meaning which will be set in resonance by the reader’s experi­ence. All film aspires toward the poetic in that it has neither a wholly predetermined syntax nor a precise, delimited symbolic vocabulary. But to grant a film documentary sense is to respect in its images the density, the plenitude of the pro-filmic: a pleni­tude which defies its reduction toward the symbolic and thus de­fies also, by implication, its articulation into a simple, linear state­ment approximating to the condition of prose. Documentary faces us with the paradox that, whilst in its elements it is capable of falsehood, since its labels are swallowed, in its totality it is not capable of falsehood, since its articulations are those of poetry. Thus whilst its elements may be misread though not miswritten, its totality—conceived as prose but perceived as poetry—may be miswritten but not misread.

 
Very well, perhaps this formulation is a little exaggerated. I do not wish to suggest that there is no sense in which documentary may be read as prose. To the extent that it uses unambiguous constructions to make statements capable of falsehood, clearly it may. By editing, I could show the “wrong” side carrying a vote or winning a battle; and even those who insist that there is no ab­solute—or at least absolutely knowable—“real” event against which its “representations” can be judged would surely agree in finding such editing reprehensible. But today’s lie, differently re­garded, may be tomorrow’s evidence. A notorious film of Pyg­mies building rope bridges becomes, when we are told that they do not build such bridges and had to be taught to do so by the filmmakers, a film about Pygmy adaptability and helpfulness. Old documentaries are constantly being ransacked for new com­pilations; yet such recycling seems always to enrich rather than diminish them. Documentary, unlike fiction, welcomes its own displacement.

 
Another way of putting this would be to say that documentary always exceeds its makers’ prescriptions. (If it did not, ethnographic films might as well be made in the studio with actors.) It is such a recognition which lies behind the wish expressed by many anthropologists that every last frame of ethnographic rushes should be catalogued and preserved in a gargantuan Monde Imaginaire for the benefit of researchers yet unborn—an idea which at first may seem as crazy as making a collection of good sentences in the hope that somebody will one day shuffle them into a book. Resistance to this recognition, on the other hand, is frequently encountered in the form of politically reac­tionary claims of privilege for the source of discourse, whether personal or (by surreptitious proxy) institutional. Superficially reasonable demands that our films be comprehensible are often in effect demands that the viewer be browbeaten into shar­ing our understanding of them. Documentary’s images are, ide­ally, not illustrative but constitutive. They are constitutive of the viewer’s meanings, since it is the viewer who constitutes them as documentary.

 
It has always been so. Of course, in the post-Méliès era, the documentary impulse had to satisfy itself with what amounted to an act of retrenchment, sustaining the Lumière actuality-values in this limited area and that—the locations, “real” people, confinement to the “facts” of known incidents—all of which became, as it were, tokens of a fidelity to be rewarded at some later consummation, endlessly deferred.

 
But if the technical developments of the past decade-and-a-half have announced that consummation, the announcement has been widely misunderstood. What the new equipment can re­cover of actuality is not its innocence but its plenitude: or rather, it can extend to most realms that plenitude which was previously available only in those of public or of nonspontaneous behav­iour. “Observational” practices increase the margin of “excess” whereby the film outstrips its makers’ intentions: and it is by this argument—by this argument alone—that they may be justified within the viewer’s frame of reference. In our example of the hands fiddling with the paper clip: once the viewer has accepted our assurance of its significance, it is its significance, not our assurance of it, that will be effective. The plenitude of the im­age, its polyvalency, is experienced by the viewer as a play of connotations. Just as the ethics of filmmakers are experienced as aesthetics by the viewer, so the anthropologist’s objectivity translates into ambiguity: and the “real-life” density commonly attributed by viewers to such film is our experience of active engagement in the generation of meaning. All film is a trace of the world: and, whilst it may be true that the very articulations whereby we are enabled to perceive it as such commit us to eth­nocentricity, this truth may at least arguably be suspended in that zone of flux at the leading edge of communication where poetry is forever congealing into prose, insight into dogma. If not, docu­mentary is unjustified.