Uploaded 16 April 1999
Robert Rosenstone’s essay raises at least two questions that should be of eminent concern to professional historians. The first is that of the relative adequacy of what we might call “historiophoty” (the representation of history and our thought about it in visual images and filmic discourse) to the criteria of truth and accuracy presumed to govern the professional practice of historiography (the representation of history in verbal images and written discourse). Here the issue is whether it is possible to “translate” a given written account of history into a visual-auditory equivalent without significant loss of content. The second question has to do with what Rosenstone calls the “challenge” presented by historiophoty to historiography. It is obvious that cinema (and video) are better suited than written discourse to the actual representation of certain kinds of historical phenomena – landscape, scene, atmosphere, complex events such as wars, battles, crowds, and emotions. But, Rosenstone asks, can historiophoty adequately convey the complex, qualified, and critical dimensions of historical thinking about events which, according to Ian Jarvie, at least, is what makes any given representation of the past a distinctly “historical” account?
In many ways, the second question is more radical than the first in its implications for the way we might conceptualize the tasks of professional historiography in our age. The historical evidence produced by our epoch is often as much visual as it is oral and written in nature. Also, the communicative conventions of the human sciences are increasingly as much pictorial as verbal in their predominant modes of representation. Modern historians ought to be aware that the analysis of visual images requires a manner of “reading” quite different from that developed for the study of written documents. They should also recognize that the representation of historical events, agents, and processes in visual images presupposes the mastery of a lexicon, grammar, and syntax – in other words, a language and a discursive mode – quite different from that conventionally used for their representation in verbal discourse alone. All too often, historians treat photographic, cinematic, and video data as if they could be read in the same way as a written document. We are inclined to treat the imagistic evidence as if it were at best a complement of verbal evidence, rather than as a supplement, which is to say, a discourse in its own right and one capable of telling us things about its referents that are both different from what can be told in verbal discourse and also of a kind that can only be told by means of visual images.
Some information about the past can be provided only by visual images. Where imagistic evidence is lacking, historical investigation finds a limit to what it can legitimately assert about the way things may have appeared to the agents acting on a given historical scene. Imagistic (and especially photographic and cinematic) evidence provides a basis for a reproduction of the scenes and atmosphere of past events much more accurate than any derived from verbal testimony alone. The historiography of any period of history for which photographs and films exist will be quite different, if not more accurate, than that focused on periods known primarily by verbal documentation.
So, too, in our historiographical practices, we are inclined to use visual images as a complement of our written discourse, rather than as components of a discourse in its own right, by means of which we might be able to say something different from and other than what we can say in verbal form. We are inclined to use pictures primarily as “illustrations” of the predications made in our verbally written discourse. We have not on the whole exploited the possibilities of using images as a principal medium of discursive representation, using verbal commentary only diacritically, that is to say, to direct attention to, specify, and emphasize a meaning conveyable by visual means alone.
***
Rosenstone properly insists that some things – he cites landscapes, sounds, strong emotions, certain kinds of conflicts between individuals and groups, collective events and the movements of crowds–can be better represented on film (and, we might add, video) than in any merely verbal account. “Better” here would mean not only with greater verisimilitude or stronger emotive effect but also less ambiguously, more accurately. Rosenstone appears to falter before the charge, made by purists, that the historical film is inevitably both too detailed (in what it shows when it is forced to use actors and sets that may not resemble perfectly the historical individuals and scenes of which it is a representation) and not detailed enough (when it is forced to condense a process that might have taken years to occur, the written account of which might take days to read, into a two or three-hour presentation). But this charge, as he properly remarks, hinges on a failure to distinguish adequately between a mirror image of a phenomenon and other kinds of representations of it, of which the written historical account itself would be only one instance. No history, visual or verbal, “mirrors” all or even the greater part of the events or scenes of which it purports to be an account, and this is true even of the most narrowly restricted “micro-history.” Every written history is a product of processes of condensation, displacement, symbolization, and qualification exactly like those used in the production of a filmed representation. It is only the medium that differs, not the way in which messages are produced.
Jarvie apparently laments the poverty of the “information load” of the historical film, whether “fictional” (such as The Return of Martin Guerre ) or “documentary” (such as Rosenstone’s own The Good Fight ). But this is to confuse the question of scale and level of generalization at which the historical account ought “properly” to operate with that of the amount of evidence needed to support the generalizations and the level of interpretation on which the account is cast. Are short books about long periods of history in themselves non-historical or anti-historical in nature? Was Edward Gibbon’s Decline and Fall, or for that matter Fernand Braudel’s The Mediterranean, of sufficient length to do justice to its subject? [1] What is the proper length of a historical monograph? How much information is needed to support any given historical generalization? Does the amount of information required vary with the scope of the generalization? And, if so, is there a normative scope against which the propriety of any historical generalization can be measured? On what principle, it might be asked, is one to assess the preference for an account that might take a hour to read (or view) as against that which takes many hours, even days, to read, much less assimilate to one’s store of knowledge?
According to Rosenstone, Jarvie complemented his critique of the necessarily impoverished “information load” of the historical film with two other objections: first, the tendency of the historical film to favor “narration” (Rosenstone himself notes that the two historical films he worked on “compress[ed] the past to a closed world by telling a single, linear story with, essentially, a single interpretation”) over “analysis”; and, second, the presumed incapacity of film to represent the true essence of historiography, which, according to Jarvie, consists less of “descriptive narrative” than of “debates between historians about just what exactly did happen, why it happened, and what would be an adequate account of its significance.” [2]
Rosenstone is surely right to suggest that the historical film need not necessarily feature narrative at the expense of analytical interests. In any event, if a film like The Return of Martin Guerre turns out to resemble a “historical romance,” it is not because it is a narrative film but rather because the romance genre was used to plot the story that the film wished to tell. There are other genres of plots, conventionally considered to be more “realistic” than the romance, that might have been used to shape the events depicted in this story into a narrative of a different kind. If Martin Guerre is a “historical romance,” it would be more proper to compare it, not with “historical narrative” but with the “historical novel,” which has a problematic of its own, the discussion of which has concerned historians since its invention in much the same way that the discussion of film today ought properly to concern them. And it ought to concern them for the reasons outlined in Rosenstone’s essay namely, because it raises the specter of the “fictionality” of the historian’s own discourse, whether cast in the form of a narrative account or in a more “analytical,” non-narrative mode.
Like the historical novel, the historical film draws attention to the extent to which it is a constructed or, as Rosenstone calls it, a “shaped” representation of a reality we historians would prefer to consider to be “found” in the events themselves or, if not there, then at least in the “facts” that have been established by historians’ investigation of the record of the past. But the historical monograph is no less “shaped” or constructed than the historical film or historical novel. It may be shaped by different principles, but there is no reason why a filmed representation of historical events should not be as analytical and realistic as any written account.
***
Jarvie’s characterization of the essence of historiography (“debates between historians about just what exactly did happen, why it happened, and what would be an adequate account of its significance”) alerts us to the problem of how and to what purpose historians transform information about “events” into the “facts” that serve as the subject matter of their arguments. Events happen or occur; facts are constituted by the subsumption of events under a description, which is to say, by acts of predication. The “adequacy” of any given account of the past, then, depends on the question of the choice of the set of concepts actually used by historians in their transformation of information about events into, not “facts” in general, but “facts” of a specific kind (political facts, social facts, cultural facts, psychological facts). The instability of the very distinction between “historical” facts on the one side and non-historical (“natural” facts, for example) on the other, a distinction without which a specifically historical kind of knowledge would be unthinkable, indicates the constructivist nature of the historian’s enterprise. When considering the utility or adequacy of filmed accounts of historical events, then, it would be well to reflect upon the ways in which a distinctively imagistic discourse can or cannot transform information about the past into facts of a specific kind.
I do not know enough about film theory to specify more precisely the elements, equivalent to the lexical, grammatical, and syntactical dimensions of spoken or written language, of a distinctly filmic discourse. Roland Barthes insisted that still photographs do not and could not predicate – only their titles or captions could do so. But cinema is quite another matter. Sequences of shots and the use of montage or closeups can be made to predicate quite as effectively as phrases, sentences, or sequences of sentences in spoken or written discourse. And if cinema can predicate, then it can just as surely do all the things that Jarvie considered to constitute the essence of written historical discourse. Moreover, it should not be forgotten that the sound film has the means by which to complement visual imagery with a distinctive verbal content that need not sacrifice analysis to the exigencies of dramatic effects. As for the notion that a filmed portrayal of historical events could not be “defend[ed];” and “footnote[d],” respond to objections, and “criticize the opposition,” there is no reason at all to suppose that this could not in principle be done. [3] There is no law prohibiting the production of a historical film of sufficient length to do all of these things.
Rosenstone’s list of the effects of historians’ prejudices against “historiophoty” is sketchy but full enough. He indicates that many of the problems posed by the effort to “put history onto film” stem from the notion that the principal task is to translate what is already a written discourse into an imagistic one. [4] Resistance to the effort to put history onto film centers for the most part on the question of what gets lost in this process of translation. Among the things supposedly lost are accuracy of detail, complexity of explanation, the auto-critical and inter-critical dimensions of historiological reflection, and the qualifications of generalizations necessitated by, for instance, the absence or unavailability of documentary evidence. Rosenstone seems to grant the force of Jarvie’s claim that the “information load” of the filmed representation of historical events and processes is inevitably impoverished when he considers the question of whether a “thinning of data” on the screen “makes for poor history.” While pointing out that film permits us to “see landscapes, hear sounds, witness strong emotions…, or view physical conflict between individuals and groups,” he seems unsure whether historiophoty might not “play down the analytical” aspects of historiography and favor appeals to the emotive side of the spectator’s engagement with images. But, at the same time, he insists that there is nothing inherently anti-analytical about filmed representations of history and certainly nothing that is inherently anti-historiological about historiophoty. And, in his brief consideration of the film documentary, Rosenstone turns the force of the anti-historiophoty argument back on those who, in making this argument, appear to ignore the extent to which any kind of historiography shares these same limitations. [5]
He grants, for example, that, although the film documentary strives for the effect of a straightforwardly direct and objective account of events, it is always a “shaped”–fashioned or stylized–representation thereof. “[W]e must remember,” he writes, “that on the screen we see not the events themselves . . . but selected images of those events.” [6] The example he gives is that of a film shot of a cannon being fired followed by another shot of an explosion of the (or a) shell some distance away. Such a sequence, he suggests, is, properly speaking, fictional rather than factual, because, obviously, the camera could not have been simultaneously in the two places where first the firing and then the explosion occurred. What we have, then, is a pseudo-factual representation of a cause-effect relation. But is this representation “false” thereby, that is to say, is it false because the explosion shown in the second shot is not that of the shell fired in the first shot but rather is a shot of some other shell, fired from who knows where?
In this case, the notion that the sequence of images is false would require a standard of representational literalness that, if applied to historiography itself, would render it impossible to write. In fact, the “truthfulness” of the sequence is to be found not at the level of concreteness but rather at another level of representation, that of typification. The sequence should be taken to represent a type of event. The referent of the sequence is the type of event depicted, not the two discrete events imaged, first, the firing of a shell and, then, its explosion. The spectator is not being “fooled” by such a representation nor is there anything duplicitous in such a rendering of a cause-and-effect sequence. The veracity of the representation hinges on the question of the likelihood of this type of cause-and- effect sequence occurring at specific times and places and under certain conditions, namely, in the kind of war made possible by a certain kind of industrial-military technology and fought in a particular time and place.
Indeed, it is a convention of written history to represent the causes and effects of such events in precisely this way, in a sequence of images that happens to be verbal rather than visual, to be sure, but no less “fictional” for being so. The concreteness, precision of statement, and accuracy of detail of a sentence such as, “The sniper’s bullet fired from a nearby warehouse struck President Kennedy in the head, wounding him fatally,” are not in principle denied to a filmed depiction either of the event referred to in the sentence or of the cause-and-effect relation that it cites as an explanation. One can imagine a situation in which enough cameras were deployed in such a way as to have captured both the sniper’s shot and the resultant effect with greater immediacy than that feigned in the verbal representation and, indeed, with greater factual precision, inasmuch as the verbal utterance depends on an inference from effect to cause for which no specific documentation exists. In the filmed representations of this famous event, the ambiguity that still pervades our knowledge of it has been left intact and not dispelled by the specious concreteness suggested in the provision of the “details” given in the verbal representation. And if this is true of micro-events, such as the assassination of a head of state, how much more true is it of the representation in written history of macro-events?
For example, when historians list or indicate the “effects” of a large-scale historical event, such as a war or a revolution, they are doing nothing different from what an editor of a documentary film does in showing shots of an advancing army followed by shots of enemy troops surrendering or fleeing, followed by shots of the triumphant force entering a conquered city. The difference between a written account and a filmed account of such a sequence turns less on the general matter of accuracy of detail than on the different kinds of concreteness with which the images, in the one case verbal, in the other visual, are endowed. Much depends on the nature of the “captions” accompanying the two kinds of images, the written commentary in the verbal account and the voice-over or subtitles in the visual one, that “frame” the depicted events individually and the sequence as a whole. It is the nature of the claims made for the images considered as evidence that determines both the discursive function of the events and the criteria to be employed in the assessment of their veracity as predicative utterances.
Thus, for example, the depiction, in Richard Attenborough’s film Gandhi, of the anonymous South African railway conductor who pushed the young Gandhi from the train, is not a misrepresentation insofar as the actor playing the role may not have possessed the physical features of the actual agent of that act. The veracity of the scene depends on the depiction of a person whose historical significance derived from the kind of act he performed at a particular time and place, which act was a function of an identifiable type of role-playing under the kinds of social conditions prevailing at a general, but specifically historical, time and place. And the same is true of the depiction of Gandhi himself in the film. Demands for a verisimilitude in film that is impossible in any medium of representation, including that of written history, stem from the confusion of historical individuals with the kinds of “characterization” of them required for discursive purposes, whether in verbal or in visual media.
Even in written history, we are often forced to represent some agents only as “character types,” that is, as individuals known only by their general social attributes or by the kinds of actions that their “roles” in a given historical event permitted them to play, rather than as full-blown “characters,” individuals with many known attributes, proper names, and a range of known actions that permit us to draw fuller portraits of them than we can draw of their more “anonymous” counterparts. But the agents who form a “crowd” (or any other kind of group) are not more misrepresented in a film for being portrayed by actors than they are in a verbal account of their collective action.
***
Too often, discussion of the irredeemably fictional nature of historical films fail to take account of the work of experimental or avant-garde filmmakers, for whom the analytic function of their discourse tends to predominate over the exigencies of “storytelling.” Rosenstone cites a number of experimentalist films that not only depart from but actually seek to undermine the conventions of commercial (especially the Hollywood variety of) filmmaking. A film such asFar from Poland , he points out, not only does not feature storytelling at the expense of analysis but actually brings under question the conventional (nineteenth-century) notions of “realistic” representation to which many contemporary historians, analytical as well as narrational, still subscribe. He specifically likens the work of experimental filmmakers to that of Bertolt Brecht in the history of the theater. But he might just as well have likened it to the work of those historians of the modern age who have taken as their problem less the “realistic representation” of “the past” than what Jarvie himself calls the question of “what would be an adequate account” of “what exactly did happen, why it happened, and . . . its significance.” This is surely the lesson to be derived from the study of recent feminist filmmaking, which has been concerned not only with depicting the lives of women in both the past and present, truthfully and accurately, but, even more important, with bringing into question conventions of historical representation and analysis that, while pretending to be doing nothing more than “telling what really happened,” effectively present a patriarchical version of history. The kind of experimentalist films invoked by Rosenstone do indeed “subvert” the kind of “realism” we associate with both conventional films and conventional historiography, but it is not because they may sacrifice “accuracy of detail” in order to direct attention to the problem of choosing a way to represent the past. [7] They show us instead that the criterion for determining what shall count as “accuracy of detail” depends on the “way” chosen to represent both “the past” and our thought about its “historical significance” alike.
Footnotes:
[First published American Historial Review 93, no.5 (December 1988): 1193-1199. References in notes to Robert Rosenstone’s article refer to Rosenstone’s contribution to that issue of AHR, and to pagination in the printed issue of AHR, not to its reproduction electronically in Screening the Past]
[1]Edward Gibbon, The History of the Decline and Fall of the Roman Empire (London, 1776-88); Fernand Braudel, The Mediterranean and the Mediterranean World (New York and London, 1972).
[2]Robert A. Rosenstone, “History in Images/History in Words: Reflections on the Possibility of Really Putting History onto Film,” AHR, 93 (December 1988): 1174; I.C. Jarvie, “Seeing through Movies,” Philosophy of the Social Sciences , 8 (1978): 378.
[3]Jarvie, “Seeing through Movies,” 378.
[4]Rosenstone, “History in Images/History in Words,” 1175.
[5]Rosenstone, “History in Images/History in Words,” 1178-80.
[6]Rosenstone, “History in Images/History in Words,” 1180.
[7]Rosenstone, “History in Images/History in Words,” 1183.