Version of record first published: 17 May 2010.

Transcription

1 This article was downloaded by: [The University of Manchester Library] On: 19 September 2012, At: 06:09 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK Visual Anthropology: Published in cooperation with the Commission on Visual Anthropology Publication details, including instructions for authors and subscription information: The ethnographic pact and documentary film Jean Pierre Olivier de Sardan a b a Professor at the Ecole des Hautes Etudes en Sciences Sociales, Marseille b Director of research, Centre National de la Recherche Scientifique, France Version of record first published: 17 May To cite this article: Jean Pierre Olivier de Sardan (1999): The ethnographic pact and documentary film, Visual Anthropology: Published in cooperation with the Commission on Visual Anthropology, 12:1, To link to this article: PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

2 Visual Anthropology, Vol. 12, pp Reprints available directly from the publisher Photocopying permitted by license only 1999 OPA (Overseas Publishers Association) N.V. Published by license under the Harwood Academic Publishers imprint, part of The Gordon and Breach Publishing Group. Printed in Malaysia. The Ethnographic Pact and Documentary Film 1 Jean-Pierre Olivier de Sardan The opposition between visual anthropology ("ethnographic cinema") and documentaries is basically a false problem, as far as both are based on a "realist pact", with ethical and technical dimensions. The "ethnographic pact" is only a variant of this realist pact. It is through the complex system of choices common to all documentaries, of which a detailed description is given, that the ethnographicity of a documentary manifests itself, with its emicness and its descriptiveness. It is evident that the term visual anthropology actually covers two distinct domains, whatever the possible connections or bridges that may link them [Heider 1976, Olivier de Sardan 1982]. On the one hand, visual anthropology describes a group of new fools and methods of ethnographic research, allowing the gathering together of a corpus of a particular kind (images and sounds) which in itself constitutes an archive (but for what consultation, and by whom?) and which must, like all research data, be gone through and analysed (but how and why?). The entire question, moreover, lies there: what problems, for what scientific objectives, demand or justify recourse to such a body of data and provide methods for analysing it? One can easily define some specific areas of investigation: gesture, technology, social interactions, rituals, and so on. At the same time, what is lacking is methodological and epistemological thinking based on empirical models for ways of approaching such bodies of audiovisual research findings in their various aspects (other than the often denigrated but nevertheless necessary one of illustration). That is not our subject here. On the other hand, visual anthropology refers to a particular type of communication product, otherwise known as ethnographic film, and about which the usual question is, what relationship does it bear to the documentary? It is this second aspect of visual anthropology that concerns us here. J.-P. OLIVIER DE SARDAN is professor at the Ecole des Hautes Etudes en Sciences Sociales, in Marseille, and director of research at the Centre National de la Recherche Scientifique, France. He has published six books about West Africa (among them: Les sociétés songhay-zarma, and Anthropologie et développement, Paris: Karthala), has made several ethnographic films, and has written some articles about the epistemology of anthropology, in Critique of Anthropology, Enquête and L'Homme. 13

3 14 J.-P. Olivier de Sardcm ETHNOGRAPHIC FILM AND DOCUMENTARY FILM The opposition between the two is largely a false problem. None of the critics who have tried to distinguish between these two filmic (or videographic) genres has been able to demonstrate a true difference between them. It will be enough to take a quick look at the criteria: these refer respectively to the subject, to the discourse, to authorship and to technique. The first two are only invoked outside the world of visual anthropology, the second two are only of interest to certain specialists. (a) Are ethnographic films distinct because they are connected with the portrayal of so-called primitive peoples? But for a long time ethnographers have ceased to define their field in these old exotic terms, and many films by ethnographers are made in a modern and/or Western context, even though film and television documentary has not given up those escapades in the tropics. (b) Is it then the form of academic discourse (which, if one wanted to be caustic, one could describe as the boring and didactic character of certain films) that defines audiovisual ethnography? But the pedantic quality of a commentary, like the heavy-handedness of a production, can hardly define a genre, nor can it set it apart, even if it is true that too many video products issuing from the cameras of ethnographers are amateurish and incompetent. This is simply a problem of professional training, nothing more nor less. (c) Can one then say that it is necessary and sufficient that a film be made by an ethnographer, or with his or her collaboration, to qualify as an ethnographic film IColleyn, in Auge and Colleyn 1990]? Aside from the obvious tautology of this definition, one can easily show that the emblematic films of ethnographic cinema do not conform to it, starting with those of the totemic ancestor, Flaherty. (d) Finally, are there production techniques specific to ethnographic film, like those sometimes cited, such as the sequence-shot and non-intervention, which are incontestable signs that a film is ethnographic [Young 1975]? Unfortunately here too the skills, the underlying motivations and the stylistic approaches for using both nonintervention and intervention, quick montage as well as the sequence-shot, have for a long time been part of the repertoire of documentary filmmakers of all kinds. This brief stocktaking makes one thing dear: nothing allows us to place the ethnographic film in obvious opposition to the documentary. The ethnographic film is part of documentary, because it depends upon the same constraints upon its forms of communication (in brief: the cinematic language) and upon the same requirements as to its purpose (giving an account of a known or unknown reality). Neither the subject, the text, nor the techniques used create an a priori boundary between the two. But one cannot get by with such a conclusion quite so easily. Is there not an incontestable, even if vague very vague "family resemblance" among a large number of ethnographic films? Conversely, is there not a long history or a strong tradition in documentary which can never be regarded as anything but ethnographic? Must not one admit that even if ethnographic film is a part of documentary, it is nevertheless a particular style of documentary? But following on from that, how can one define that "particularity" if one cannot rely upon the existence of any

4 Ethnographic Pact and Documentary 15 differentiating criteria? Isn't it a paradox to suggest that ethnographic film is a sort of subgroup of documentary if one cannot clearly define the boundaries of that subgroup? It seems to me that one can resolve this paradox if one considers the ethnographicity 2 of a film as an indicator which influences the documentary, as other indicators (for example, its political militancy, or its impressionism) also influence it or color it. More precisely, the notions of an ethnographic pact and of choice of directorial method may allow us to explore the ethnographicity of a film, reading it as a documentary film. It is this that I propose to attempt here. But before speaking of the ethnographic pact it is necessary to speak of the realist pact. THE REALIST PACT AND THE ETHNOGRAPHIC PACT One can easily show that even the distinction of fictional film from documentary film (much more solid than that between ethnographic and documentary film) cannot be founded on radically differentiating indicators, such as type of subject, discourse, authorship or technique. The documentary film utilises the entire range of devices which are used in fiction (without of course being able to take advantage of all the same means but this is a matter of scale): the making and manipulating of images, as has been known and stated for a long time, are based on the same cinematic processes in all genres. Despite certain naive notions which have long been stripped from cinima verite, documentary film cannot escape from the obligatory law of the "cinematic lie." In fact, one could argue that these processes are nothing but the elementary laws of cinematic language, which compose and select images (shots) in the first instance, and then reselect, reorganise and recompose them in the second (in editing), and that it is at some point beyond the application of these laws that fiction and documentary radically diverge from one another. Isn't it, for example, at the level of the actors and themes that, despite their minimal common language, the main differences between fiction and documentary emerge? Actors and staging 3 in fictional cinema: no actors and no staging in documentary cinema. But it is not so simple as that. The people in a documentary film are certainly not Hollywood actors, but their "naturalness," their "unprofessionalism" and their "spontaneity" still cannot disguise this plain fact: they act! But they do something different from acting other roles. They act their own roles instead of playing the roles of people in a script. As for staging, it is far from absent in documentary, and one can find its traces at every level, from the basic unit of the shot (where one asks a person to enter the frame at a signal from the camera operator) up to where one would expect it least: in filming, without any interference by the crew, a collective ritual, which is in no sense performed for the camera, but which represents not a renunciation of staging but a deliberate choice of a "found" mise-en-scene, that of the ritual itself, which we have long known to be at least, if not above all, a collective performance enacted for its own participants or for others. One can in fact reapply, in relation to documentary, the demonstration which Lejeune applied brilliantly to a particular "literary genre": that of autobiography

5 16 J.-P. Olivier de Sardan [Lejeune 1975]. To summarise his point: no narrative technique, no grammatical or stylistic process, no editorial or typographic arrangement in other words not one objective indicator enables us to differentiate between the autobiography and the hovel. And yet... and yet, they are not the same thing. Lejeune resorted to the notion of an autobiographical pact. This is because the author of an autobiography makes a promise to the reader that this will really be an autobiography, that this is to be read as an autobiography and not a novel. There is a responsibility on the part of the author not to create confusion along the way by using methods which would break this tadt contract. Basically, exactly the same thing occurs in the documentary film: the director makes an analogous pact, which one may call a realist pact. "This documentary film which you see gives an account of the truth, and the images which I show you are the products of reality and not the effects of a fiction." It matters little that, among all the images, those of documentary have been manipulated by the techniques of shooting and editing: it is enough that we agree to ignore this, and that as a result any "effects" (or naiveties) of bad production do not put us off, or at least not very much. In this respect, ethnographic films adhere to nothing more than a particular form of the realist pact, which one may call an ethnographic pact. Seen from this perspective, the various inadequate criteria which we examined at the beginning, and which we were unable to use to define ethnographic film as a distinct genre of documentary, now resurface and occupy a new position, although a relative and unstable one. They are no longer differentiating indicators but indices of the ethnographic pact. Specialised knowledge of an exotic subject, for example, which transforms the strangeness of a distant custom, or the pedantry of a commentary, or the actual presence of the researcher these constitute three indices which signal or reinforce the ethnographic pact as a particular form of the realist pact. Can one characterise more precisely the realist pact and its variant, the ethnographic pact? The one, like the other, has first of all an ethical dimension: if the cinema is nothing but an illusion, the author nevertheless guarantees that this illusion is "true" in other words, that the lies of cinematic language and documentary expression are necessary mediations between two truths: the truth perceived by the spectator and the truth of the situation to which it refers. If the images are "manipulated", they are not completely fabricated, and they nevertheless "represent" the reality to which they refer, they preserve a trace of it. The cinematic illusion appropriate to documentary itself guarantees a fidelity to the reality to which it refers. This is the ethical component in the realist pact, and the ethnographic pact conforms to it. In a certain way it reinforces it: it postulates a realist expertise, that is, it guarantees a knowledge which in theory should be more profoundly based upon the reality to which it refers than that which a layman might have, a knowledge which can express itself through the various signs that make up the language of cinema [cf. below]. But there is also a technical dimension to the realist pact: it is in fact best not to break the cinematic illusion, since this tends to leave a misleading trace of the technical processes by which filmmaking manipulates images and sounds. Not only must the illusion be true, it must also be believable. It is the skill of the director which underwrites the technical guarantee of the realist pact. Even more, the

6 Ethnographic Pact and Documentary 17 ethnographic pact respects this component of the realist pact and in fact willingly reinforces it. The minimum obvious intrusion of production techniques (the absence, for example, of ad hoc professional music), more than usual thoroughness of observation, a slower editing style all these are technical effects (for this is very much a matter of technical effects) which emphasize the realist pact. Through these two elements the ethnographic pact can be seen as but an accentuation of the realist pact in other words, documentary simply displays here the indices of the realist pact more strongly than usual, which leads in general (although not necessarily) to a reduction of other indices. Thus the subjectivity of the author, aestheticization of the subject, and polemical forms of rhetoric all tend to be reduced. But one may introduce a third element which can be a part of the ethnographic pact and which one could call emicness. This becomes necessary in order to guarantee that the spectator will not simply be shown what has happened, but that he or she will achieve a particular level of understanding which is dose to the perceptions and the logic of those who see it, so as to open a window on the "point of view" of the "natives." To give at least minimal but accurate access to the "world view" of the subjects filmed is one of the specific aims of the ethnographic pact, which clearly corresponds to one of the primary objectives of anthropology in general (specifically, the concern to provide an "emic" representation of social facts and behavior). To take a well-known example: Rouch's film Les Maitres fous. It is a documentary film and, because of the realist pact, the viewer accepts that the scenes are "true": those possessed do not significantly change their behavior because a camera is there, the possession is not simulated, there are no frills attached to the product. The ethical dimension of the realist pact is respected, as is the technical dimension: the skills of the director give us the impression that we are actually present at the ceremony. Perhaps less important, but proceeding from this, we can tell that the possession of these immigrant workers in the slums of Accra has not been organized for the needs of the film, and that Rouch has not employed his "actors": in every sense, the possession has taken place as it should, and the actors in the film are well and truly in trance. But there is something more: there is the sense that an ordinary tourist, or a passing journalist, could not have filmed here in this way... That the authenticity of these scenes is better guaranteed than in previous films about other cases of possession... That there is behind these images a comprehension and a familiarity (particularly when one thinks of the very early date of this film)... That the ritualised logic which supports these acts, and which animates these faces, becomes almost understandable to a viewer whose own culture must be at the very opposite extreme of the one he or she is watching... It is this rather diffuse sense, which one finds again and again in documentaries whose style varies as much as their subject matter, their techniques, and their degree of success, that I propose to call the ethnograhic pact. How does this sense enter into images which vary so much in their nature? How does this ethnographic pact, in its varied dimensions (especially its accentuation of the realist pact and of emidty) "come about" in a film? By what indices does it reveal itself? It is here that we must introduce another, more tangible, analytical dimension,

7 18 ].-P. Olivier de Sardan which views documentary as a system of embedded choices. It is in fact through these choices, common to documentary films of all kinds, that the ethnographicness of a documentary film manifests itself. MICRO-CHOICES AND MACRO-CHOICES All film is a complex system of choices, made throughout the process of production: choices of subject, selection of natural settings, writing, filming, editing... At each of these stages, the constraints and opportunities which exist create a set of possibilities among which one must decide: constraints and opportunities linked to cinematic language, to the norms of the local society, to the institutional situation, to the budget of the production, to stylistic conventions, and so on. These choices overlapping, continuous, multiple are certainly often implicit, that is to say, made not as a result of careful deliberation and reasoned decisions. But for all that they are no less real or influential. Their implicit character may result from two opposite causes: It may result from technical and narrative incompetence (so frequent among ethnologists who launch themselves upon a production without training), which makes these choices appear not to be choices at all, because the range of possibilities is not understood and the inventory of directorial skills has not been mastered. It may, on the contrary, result from technical and narrative competence which in the world of filmmaking often corresponds to a formula of "how it is done," reinforced by "on the job training" which remains undigested and unexamined, and which expresses itself in a rhetoric of feeling / in the name of which these choices are made spontaneously and intuitively. But they are no less influential for all that. The origin of these choices is therefore by no means obvious. One of the tasks visual anthropology should certainly set itself is to examine them and gauge their effects: documentary can content itself with a level of debate familiar in cinema clubs, or it can remain an impressionistic kind of criticism, or value only the kind of ineffable and subjective discourse associated with the creator or artist; but, on the other hand, one can imagine an inevitable development of it as a subdiscipline oriented towards the social sciences, which tries to create a body of work through systematic and reasoned analysis. Here one touches in passing upon a third meaning of the term visual anthropology (other than the gathering and analysis of a body of audiovisual material, and the production of ethnographic documentary films): the study of the modes of knowledge and of the mise-en-scene of the real, seen through the processes of documentary film (which one might call, at the risk of seeming pretentious, an epistemology of the realist pact...). The choices made in a documentary film work, and therefore in an ethnographic film, can be analysed according to two different scales: on the one side, microchoices (framing of a shot during filming, joining together two shots in editing,

8 Ethnographic Pact and Documentary 19 for example); on the other, macro-choices (deciding upon a subject, or whether or not to use a commentary, for example). The micro-choices have been widely studied by what might be called the School of Nanterre, founded by Claudine de France [de France 1982; Comolli 1983]. The macro-choices have been less well explored. These are the choices of direction, whereas the micro-choices are those of recording images and sound. The privileged level of the micro-choice is the shot. The privileged level of the macro-choice is the sequence, and the film itself as an articulation of sequences. Through these macro-choices (but through micro-choices as well) the realist pact is born and reproduced. And it is through them that the realist pact eventually takes its form as an ethnographic pact. DIRECTORIAL CHOICES AND THE ETHNOGRAPHIC PACT I shall restate here three main types of choice: it is through these that one establishes a strategy for direction and upon which the ethnographicness of a documentary film is constructed. The Subject Obviously not everything is "good to film" even though almost everything is filmable. The construction of a filmic subject offers certain analogies to the construction of a scientific subject, but it obviously differs in that its methods and its language are not literally comparable. It is most unusual for a "good" research subject to be an equally "good" film subject. A documentary filmmaker, even if he is also an ethnographer, constructs a subject differently than a researcher, even if he is also a filmmaker. The ethnographic filmmaker (or ethnographer/director team) should here be distinguished from a director without any pre-existing anthropological plan or objective by the fact that the subject is most often extracted from previous social research. In other words, the subject benefits from a very particular kind of knowledge based on professional anthropological practice over a long period. This has two effects: on the one hand, the subject is taken out of a selection of materials which was not for the most part organised to function as a film, since the choice of the subject generally precedes its selection as a subject by the documentary filmmaker; on the other hand, this selection values certain aspects of reality generally not of interest to an ordinary filmmaker and orients the film, because of the choice of subject, in an "ethnographic" direction. But it is also clear that a documentary filmmaker who spends an entire year living among the people he or she films is in a position of participant observation quite analogous to that of an ethnographer... The choice of a film subject becomes, to the extent that an ethnographer participates in making it, the expression of a compromise between an ethnographic interest and a narrative interest, this latter referring to different parameters of the subject, such as its novelty, the uniqueness of the scenes which compose it, the beauty or photogenic qualities of the locations, the events, the characters, etc. Of course, in documentary in general a narrative interest is rarely the only concern, but

9 20 J.-P. Olivier de Sardan it is even more rarely joined to an ethnographic interest based upon a "deep" knowledge. The Narrative Web To make a film on one subject or another is not to predetermine how to deal with it. Every documentary and ethnographic films are no different in this than any others is a story about the real and takes its point of departure from it. It therefore must have a line of direction, or a "transport mechanism" [MacDougall 1978]. We know that time often fulfills this role. Chronological structure is a convenient and often-used line of direction. There are certainly others with a thematic basis, such as geography, family life, etc. But chronological structure remains the most frequently used, and it is certainly this which permits the greatest ethnographicness, at least of the so-called "simple" form. It can in effect be analysed as a simple type or complex type. Let us call simple chronological structure a chronological structure based on the unity of a limited period of time, functioning rather like the unity of place: the prototype of this is what one might call a "social event." It is a narrative structure frequently used in documentaries of an ethnographic kind: technological processes, like ritual processes, have produced the material for innumerable "ethnographic films." This is almost to say that the line of direction also functions as the subject. The subject is in a sense its own line of direction, for to name it is already to give an account of it. Put otherwise, to describe a social event is already to recount it. The linearity of the subject is such that, even if it is obviously broken into many sub-sequences, it can also be read as a single sequence. It has the quality of certain one-sequence-films. This is a matter, as one can see, of a particularly convenient compromise between ethnographic interest and narrative interest. In films in which a technological process or a ritual constitutes at the same time the subject and the line of direction, the ethnographic description transforms itself relatively easily into a filmic description, which is in itself a filmic account. One could say the same of subjects treated less often, which are equally important and are strongly descriptive, consisting of sequences of social interaction: a domestic scene, a children's game, a market transaction... There, as distinct from technological processes or rituals, it is difficult to propose these "free" interactions as complete films, despite their self-sufficiency (unity of action, of time and of place), since they lack the mechanical rigor of technological processes and are also without the collective mise-en-scene of ritual processes. They tend therefore to constitute sequences, scenes (seen parenthetically) within other subjects. As for a complex chronological structure, it is a chronological structure which cannot function directly as an account and which must be constructed as such: a life, an era... Far from adopting the preceding descriptive attitude (not unlike that of the ethnographer), the director must remain divided in time and space, so as to invent a transport mechanism that can jump diachronically as well as synchronously, weaving its own line of direction out of more scattered spatiotemporal fragments. The chronological linearity of the account is no longer a linearity resulting from the fuming. The film need not be less rich ethnographically, but this comes about differently, through a different set of choices. The narrative

10 Ethnographic Pact and Documentary 21 structure (a use of portraits, an account of an enquiry, a historical survey, etc.) moves farther away from the descriptive. Filmmaking Options At the same level of choice I group a series of choices that are in fact of different orders but that taken together have a powerful role in giving a film its personality, its color, its style and its possible ethnographicity. One can identify four series of these. 1. Filmmaker/Film Subject Interactions This is a question of the level of intervention by the film crew in relation to the "actors," i.e. the people who pass through or give lif e to the documentary. At one extreme one of course finds complete non-intervention, which was once the claim of the "observational" film [Young 1975]. Even if this does not eliminate the manipulation of images inherent in all documentary filmmaking (occurring particularly at the editing stage), and even if it does not eliminate the possible modification of the behavior of the subjects (the effects of filming), or make claims to the status of "cinema-truth" (dnfrna v&rite), it nevertheless produces methods of filmmaking which reinforce the ethnographic pact, because one could call it an effect of direct recording. At the other extreme, one finds strategies of staging and deliberate acts of miseen-sctne by the director. Sometimes these will entail methods of fictionalization or direction of the actors (even if they are playing their own roles, the minimum requirement of the realist pact), sometimes the restaging of events or lapsed rituals by those who know about them, and sometimes suggesting to people in the film that they act out their routine activities in an improvisational way. Most often, directors place themselves between these two extremes and adopt a moderate form of intervention: asking someone to wait for a signal to enter the frame, making an artisan repeat a technical step, arranging the background decor of an interview, etc. This last example obviously raises the question that is always at issue in relation to the effect of fuming: to what extent does this sort of thing introduce a bias and risk breaking the realist pact? The interview is in effect a means of maximizing the effect of filming and making it so obvious that it poses no problem, since it is the camera itself that both confronts and is addressed by the subject. By developing this strategy beyond the interview (in the strict sense), whole films can be made on the principle of a companion-camera with which the people in the film can converse (for example, to Leacock or Depardon). The effect of direct recording is thus guaranteed here. But the arguments for this are not always based upon efficacy: they are also ethical, and for this reason "dialogue" is sometimes preferred to an observational approach [MacDougall 1975], for the latter approach, by contrast, often tends to make the subject forget the presence of the camera. But in the case of certain rituals (a person in trance, or a mask), military maneuvers (a march), festivals (a carnival) or political events (a demonstration), this approach allows the ethnographic pact to function at its highest level.

11 22 J.-P. Olivier de Sardan Apart from these cases, there are those instances in which one asks the subjects to behave "as if" the camera were not there, although the situation is one in which they could hardly ignore it. The filming effect obviously varies considerably, depending upon the nature of these situations: a blacksmith can no doubt continue to "act as if" while hammering on his anvil without much trouble, but one can be sure that no lover can "be natural" while courting his girlfriend under the eye of the camera. It is situations such as these, half forbidden to documentary, that threaten to break the realist pact situations that subject the "actor" to contradictory pressures (a double bind) of the sort: "But behave naturally!" No single form of filmmaker/film-subject interaction has a monopoly on ethnographic truth, but one can conjecture that there is less of a problem in interactions between filmmakers and their subjects than in interactions between film subjects. And as a result, how one decides to treat the first sort of interaction depends very much upon how one intends to treat the second. To the extent that interactions between people are primarily matters of social relations outside the concerns of the film, one can conclude fairly reliably that the documentary will tend to minimize these in favor of action, on the one hand (such as largely functional interactions, technology and ritual, in particular), and speech, on the other (such as interviews, or encounters between celebrities and their audiences). When showing interpersonal relations, whether transient or habitual, a certain feeling of proximity is created which is shared with the viewer. And does it matter whether the very special mood this creates results from a participatory camera or a camera which is completely ignored, provided it works? 2. Commentary and Dialogue Whether to have commentary or not, interviews or not, dialogue between characters or not these choices are in part a function of the subject, in part a function of the line of direction of the film, and in part quite separate from either of these. A major problem, crucial to all documentaries (but even more so when ethnographic concerns are joined to narrative ones), is the management of information essential to the viewer. It is even more crucial if the linguistic and cultural distance between subject and viewer is very great (although it is never completely absent, since all documentary necessarily contains an element of discovery), and this problem reaches its peak when the subject is particularly exotic. The number of possible solutions is then limited: to provide such information outside the film, to incorporate it in the film as written text, or to incorporate it verbally. Information outside the film can be provided in a leaflet or a booklet distributed with the film (at screenings or public presentations) or, more fully, in a joint videocassette/book publication [cf. Auge and Colleyn 1990]. Unfortunately, these instances remain quite rare and do not involve broadcast television. Written textual information within a film consists primarily of captions, annotations and explanatory titles, either at the start or scattered throughout the film. However, the most common method is the verbal one. It is at this point that the choice of commentary arises, which has long been the only method of delivering explanatory and contextual information, and which all genres of documentary have at various times used and abused; and also the interview, which can also assume some of this function. (MacDougall has even described this as "interior

12 Ethnographic Pact and Documentary 23 commentary" [MacDougall 1975], an option which has been equally used and abused.) In this regard, it seems to be not the use of one or another method of disseminating information (or its abuses) that distinguishes ethnographic films from other forms of documentary, but rather the content of the information conveyed: a priori, if the information comes from a team including an anthropologist, for example, it will be more fully contextual [Heider 1976]. It is true that an anthropologist from an intellectual and academic background will tend, not without risks, to recapitulate his field research and spell out everything verbally, and that ethnographic concerns in this case may run the risk of swamping whatever narrative interest the film may have. But it would be wrong to identify an ethnographic film only by the content of a knowledgeable commentary. On the contrary, an ethnographic film often manifests itself by its silence. How is this possible? The answer becomes clearer if one remembers what was mentioned above as the descriptiveness of certain subjects (especially, social events) as an ethnographic index. It is not only the descriptiveness inherent in certain subjects (we are now at the level of "choice of subject") which feeds this ethnographic index, it is also in the descriptiveness of the position assumed by the filmmaker, manifested in an economy of words (we are now at the level of "filmmaking options"), for to give the viewer the illusion of being introduced into "the interior" of a world not his or her own, as a familiar guest, but also with a sense of separateness, is one of the ways in which documentaries create, imperceptibly, and without fanfare or words, their ethnographicness. 3. Bias in Film Style Here we are concerned more with the actual filming than with editing. This is a matter of the micro-choices that are made, to the extent that these serve as the site of a bias, or constitute a deliberate strategy of microchoices in the design of the film, favoring such a type of framing or such a type of shot, this type of linking or that type of rhythm. At the time of filming the choices are obviously what sort of shots and sound recordings to make. The biases involved may be of an ethnographic kind. Alan Lomax, making a film on dance, thus shoots in wide shots in order to show all the movements from head to foot. Other filmmakers may use the sequence-shot as a way of ensuring fluidity or descriptiveness. In editing there can be a strong time-sense or weak time-sense (giving one or the other more importance), or a minimizing of visual effects (i.e. distortions of the image), or a refusal to use illustrative sound, or the choice of a moderate or slow rhythm (a refusal of quick editing) all prominent indicators of the ethnographic pact. 4. Paradigms and Points of View Anthropologists do not escape ideology, and their films, to judge from the evidence, are far from expressing the disinterested point of view that characterizes the assumptions about, and current stereotypes of, the intellectual. Their discipline is itself increasingly splintered and divided. They are confronted by differing concepts of reality. But these perspectives are in fact fairly limited in number, and films made by anthropologists (which they have themselves directed) or on which they have collaborated (as writers or scientific advisors) bear the traces of these perspectives.

13 24 J.-P. Olivier de Sardan Minimizing or maximizing the exoticism of what one shows, presenting it as familiar or unusual, dramatizing it or making it commonplace these are, on the one hand, scientific choices and, on the other, ideological ones, which inevitably result from the orientation of the anthropologist, among those who have a significant role in the working out of a documentary. Presenting an enchanted image of local communities one has filmed, or emphasizing their internal contradictions, or underlining their difficulties and the oppression of which they are victims, or stressing their inventiveness and resourcefulness all these appear explicitly or implicitly in a film at the level of the varied choices that produce it. Neither the non-anthropological filmmaker nor the anthropological filmmaker can escape from one or another of these biases. But perhaps it is the anthropological filmmaker who is more used to challenging "commonsense" evidence, particularly when it involves non-western cultures and requires avoiding the traps of exoticism. From this point of view, the "anthropological project" can express itself through a thousand different filmmaking methods, making tangible, visible, audible and perceptible the logic that underpins the behavior of the social actors who are filmed, and by refusing to sensationalize or overdramatize cultural differences. This is not, or need not be, just a question of "commentary" but rather the approach to filming and editing as well. Certainly restraint, collaboration with those being filmed and respect for their customs are not the exclusive province of anthropologists, but they are at the heart of it to a greater extent than in other professions. CONCLUSION Having completed this journey, is it now possible to suggest a minimum definition of the "index of ethnographicness," granting that there is no dear boundary which separates the subgroup of "ethnographic film" from the genre of "documentary film" of which it forms a part? Throughout the complexity of detail and the filmmaking options which define it, with its subject and its line of direction, the sound and the color of documentary film, is there not perhaps, "despite everything", an "ethnographic style"? one that I for my part would willingly organise around two properties which are central to the ethnographer's method (less in the conception than in what one practises) and which can also be expressed but differently in images and sounds (as distinct from other characteristics of social science which seem, in themselves, inseparable from scientific thought and its written expression: challenging of sources of evidence, theoretical debate, conceptual analysis, etc.): emicness, the expression of indigenous perspectives [Olivier de Sardan 1998]; and descriptiveness, the careful observation of behavior, customs and interactions. Emicness and descriptiveness are certainly not the only properties which anthropology pays attention to, but they nevertheless reflect what one might call either an "anthropological culture" or an "anthropological sensibility." All this being not the monopoly of professional anthropologists, but rather allied to a set of skills which anthropology has developed, and with which it has surrounded itself.

14 Ethnographie Pact and Documentary 25 The particular attention paid to describing human actions and interactions, the particular attention to the meanings which the people filmed attribute to their actions and interactions these constitute in the last analysis the ethnographic index which imperceptibly transforms the realist pact into the ethnographic pact. It is one form of respect accorded to filmed reality but there are of course others as well... (Trans. from the French by David MacDougall) NOTES 1. First published in French ("Pacte ethnographique et film documentaire") in Xoana 2, (1994): ethnographicité. 3. mise en scène. 4. The English word feeling is used in the original. REFERENCES Augé, Marc, and Jean-Paul Colleyn 1990 Nkpiti. La rancune et le prophète. Paris: Editions de l'ehess. Comolli, Annie 1983 Les gestes du savoir. Paris: Université Paris X. France, Claudine (de) 1982 Cinéma et anthropologie. Paris: Maison des Sciences de l'homme. Heider, Karl 1976 Ethnographic Film. Austin: University of Texas Press. Lejeune, Pierre 1975 Le pacte auto-biographique. Paris: Seuil. MacDougall, David 1975 Beyond observational cinema. In Principles of Visual Anthropology. Paul Hockings, ed. Pp The Hague: Mouton Ethnographic film: failure and promise. Annual Review of Anthropology, 7: Olivier de Sardan, Jean-Pierre 1982 Production écrite et production audio-visuelle. Geste et Image, (special issue) Emique. L'Homme,147: Young, Colin 1975 Observational cinema. In Principles of Visual Anthropology. Paul Hockings, ed. Pp The Hague: Mouton.