(Visual Sociology 9:1, 1994)


by Bruce Jackson


Documentary artists have long paid witness to and fixed in the amber of their vision lives ignored by historians and nearly everyone else. What would we know or what diminished sense would we have of pre-smallpox Mandan life on the upper Missouri, mid-19th century London poor, slumdwellers of turn-of-the-century New York, tenant farmers in depression Alabama, prisoners in Massachusetts' Bridgewater State Hospital for the Criminally Insane or HIV positive gay black men were it not for the documents provided us by George Catlin, Henry Mayhew, Jacob Riis, James Agee and Walker Evans, Frederick Wiseman, and Marlon Riggs?

The New Social History has discovered in the past what documentarians in all media have for a century examined in the immediate present: that history exists at all levels of society, not just the level that writes political memoirs; and that our documents are themselves creations driven by aesthetic and political criteria.

Perhaps the most important lesson of the New Social Historians is that history belongs to those about whom or whose documents survive. Ken Burns's The Civil War miniseries on PBS (1990) is grounded in public and private letters, diaries, and photographs; the words of the illiterate or images of the unphotogenic are absent. The Australian scholar Shane White was able to write about physical and behavioral characteristics of 18th century black Americans by examining newspaper classified ads for runaway slaves, the forum in which whites would be least likely to exaggerate or caricaturize (1991). The name of Gilgamesh is well known because he was the subject of a series of heroic poems not long after his reign and many of the tablets on which those poems were inscribed have survived; poetry about Gilgamesh's immediate predecessor and successor did not survive and we have no stories to tell about them.

Ordinary people now document the notable moments of their lives in sync-sound video using inexpensive lightweight handheld VCRs capable of producing tapes of a visual and audio quality that only a decade ago required a van packed with expensive highly sophisticated equipment. For the first time since the only means of documentation were notebook or sketchpad, ordinary people share with the privileged sophisticated means of documentary recording. The results are sometimes spectacular. Had not George Holliday been standing on his patio trying out his brand-new camcorder, Rodney King would have been just one more drunken black man convicted of speeding and resisting arrest. The PBS film memoir about physicist Richard Feynman (Richard Feynman: The Last Journey of a Genius, 1989)  was shot largely with a home camcorder. In Africa and central and South America now, and in the Soviet Union during the last few years before the breakup, unimpeachable political messages were sent and events documented and shared using portable video. Many of those tapes still exist; they are facts on the way to becoming history.


The American Historical Review devoted much of its December 1988 issue (93:5) to a forum on film and history. In the keynote piece, Robert A. Rosenstone, upon whose research both the commercial film Reds (1981) and the documentary The Good Fight (1983) had been based, complained that, "The Good Fight—like many recent [documentary films]—equates memory with history; it allows veterans of the Spanish war to speak of events more than four decades in the past without calling their misremembrances, mistakes, or outright fabrications into question" (1180). He also complained that when he wrote the film's narration the directors frustrated his desire to include

the issue of possible Stalinist "terrorism" in the ranks. Their objections were as follows: they could find no visual images to illustrate the issue and were adamant that the film not become static or talky; the topic was too complex to handle quickly, and the film—as all films—had so much good footage that it was already in danger of running too long. This decision to sacrifice complexity to action, one that virtually every documentarist would accept, underlies a convention of the genre: the documentary bows to a double tyranny—which is to say, an ideology—of the necessary image and perpetual movement. And woe to those aspects of history that can neither be illustrated nor quickly summarized. [Rosenstone 1180]
Rosenstone's objections to the fictive and documentary films based on his work are interesting, but at least equally interesting are the questions his comments suggest about history. If the documentarian doesn't replicate the historian's vision or ask the historian's questions, is it because documentary fails at that work or because documentary does other work instead? What can historians and other social scientists learn from documentarians?

Hayden White, the philosopher of history and author of Metahistory, pointed to the key defect in Rosenstone's argument with documentary:

 ...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." [Hayden White1193]
Historians like Rosenstone privilege print. They may prize film for what it can show them about something, but they fault film for its apparent lack of analytical complexity. They thereby miss the point of film. A sound film is not just words with accompanying pictures; film is more than data that can be summarized in words. Film is a way of knowing, which means it has essential aspects that are not verbal. It is like music or food or sex, which also have aspects that cannot be delivered in words: no matter how good the writing about, it's never the thing itself. Talk about film, as Trinh Min-Ha puts is, is always "beside, not in" (27 February 1973).

All social science is about something; documentary film is about something and is something. Hearts and Minds is not only a powerful emotional construct, it is also a keen examination of real things that happened. Good films don't present ideas; they are ideas given body.


Documentary films are created in an inverted funnel of declining possibility. The filmmaker begins with the materials and subjects of the world, which are infinite in scope. From that infinitude one selects a subject, then an approach. These decisions are in part predicated on practical concerns: the amount of money determines the size of the crew, the amount of footage that can be purchased, shot and processed, the number of location and editing days, the sophistication of the sound mix. Thousands of microdecisions are made for aesthetic reasons, technical reasons, weather reasons, background noise reasons, time-to-go-to-dinner-now reasons, access and permission reasons, sore back and feet reasons. Every decision forecloses others and the possibilities become ever narrower.

In films based on real-time events, filmmakers are limited by the arrow of time: they return from the field with a fixed amount of data, a fixed world of possibility: so many feet of exposed film and audiotape. (They can go back and do pickups, but it's never the same, and that just changes the moment at which the world of possibility is fixed, not the fact or limitation of it). They may manipulate those images and sounds, but they will have no other images to do things with.

The first part of the editorial process is a discarding of shots and sounds that didn't work: bad focus, over or undermodulated sound, a thumb in front of the lens or an unacceptable hair in the gate. As scenes are juxtaposed with other scenes, footage that once seemed indispensable becomes irrelevant or redundant and it too goes into the outtake cans. Shots you risked your life to get just don't work when bracketed by shots that editing shows must come before and after, and they too become outtakes, filmic detritus. The 25 hours of footage brought home from location becomes 20 hours and then 10 hours and ever less until it gets to the final length.

Some filmmakers employ the character of an external narrator as a major organizing device; others use only voices of people who are part of the world depicted. This matter of narrators may seem a minor difference, merely a stylistic choice, but from a filmmaker's point of view the differences are profound. There is no parallel division in the writing of history or sociology; annotated interviews or diaries aren't the same thing.

A film without a narrator is composed of nothing other than primary material: footage made in the field, found in archival searches, or interviews. The work of editing such a film is entirely predicated on decisions previously made and images previously captured. Films with narrators are far easier to edit than films without narrators, for the obvious reason that the narrator's voice can utter facts not documented in the images and soundtracks, and it can provide continuity and transitions where none otherwise exist. The narrator's mellow voice of reason can say, "And now we'll turn to a consequence of that decision...." Or, while a field of lilies waves prettily on the screen, "they all died within the next two years and their graves were unmarked."

 Richard Ellison, executive producer of Vietnam: A Television History bragged that his series had

No fancy intercut editing, no emotive music, no omniscient narrator. Plainness is in the interests of the philosophical objective, which is to manipulate the viewer as little as possible. The archival film is what it purports to be; sources are identified when necessary; contradictory viewpoints are clearly articulated; conclusions and value judgments are expressed by participants and interview subjects, not by the program makers. [Frisch 161]
Maybe yes and maybe no. It depends what is done with that archival film, for only the filmmaker knows what footage is rejected and what the real principles of admission or selection were. Absence of the omniscient narrator and emotive music doesn't mean the product is morally or politically neutral; it just means that whatever is being done is being done the hard way rather than the easy way, that it's being done with film technique rather than talk technique. If you're going to deceive in a narratorless film, you must use the images of the real world to do it, which is harder to do. Harder, not impossible.

Filmmakers who use narrators pay a price for taking the easy way: narrated films date far more quickly than films without narrators. That is because voices in primary footage, the "real life" material depicted in a film, are always perfectly contemporary and therefore consonant with the images and action in the film: people interviewed in 1936 talk the way people in 1936 talked. But narrator's scripts and voices are outside the action, they are written and pitched in terms of ideas and styles current when the film is edited, so with time they become more and more dissonant. The narrator's profound and wise words that seemed to apt to the producers when a film was made in 1970, may now seem to us naive, tendentious, pedantic, and even soporific.

Classic examples of narratorless presentation are Leni Riefenstahl's Triumph of the Will (1935), which has only a few title cards to identify the scenes, and Frederick Wiseman's Near Death (1990). One polar opposite is Thames Television's The World at War (1974-75) with Lawrence Olivier's magisterial baritone making it all make sense for us, telling us what happened and what to think. Another is a typical "60 Minutes" sequence or "Frontline" broadcast, in which the narrator's voice is primary, and most images are simply illustrations of what the narrator is saying to us.

A comforting voice of authority can help people trust the images and voices in the film and thereby make the unfamiliar or uncomfortable more accessible. On the other hand, a narrator distances the viewer from the subject, puts the people pictured in the third person, and creates a mediating position between "them and us." Without the narrator, film has a far greater chance of achieving the "I-thou" intimacy, of giving the viewer a sense of direct contact with the event or subject. That sense is of course deception: the direct contact is always with the film itself.


The key fact missed most often by social scientists utilizing documentary films for data, is this: documentary films are not found or reported things; they're made things. They differ from fiction films in that they bring with them questions of accuracy linked to an external reality. But knowing about that linkage doesn't take us very far: you look at my film and say "It's not the way it seemed to me" and I respond, "Well, that's how it seemed to me." We need to remember that documentaries don't document just an event; they also document the perception of an event, and that perception is enacted in the film itself. That double documentary aspect is what permits multiple documentaries to be made about the same event or person and for each to have validity and utility. Right and wrong exist somewhere else.

Presumably, if two physicists set out to measure and weigh a chair and have at their disposal well-calibrated instruments they will arrive at the same numbers. But if you and I each document a football game or executive conference or war with identical cameras we will not produce the same photographs or films. A "scientific" inquiry provides a statement about an object or condition; the documentary inquiry produces an object that has its own condition.

This perception of the nonfiction thing made and having its own validity is recent, but it helps us re-see contemporary reports of the past. Jacob Riis matters now not just because of the overt content of his images, but also because of what his pictures say about seeing. Henry Mayhew matters because of what he told us about the London poor 150 years ago; his work also matters because of what it tells us about how things could be documented and conceptualized then. (Even Charles Dickens, who was there, depended on Mayhew for the flavor and facts of London poverty.) The rightness or wrongness of the referred object is only part of the conversation. The documents become texts that can themselves usefully be read at levels beyond or apart from the immediately obvious. This was Michel Foucault's great lesson. Where once we might have faulted whole classes of old documents for antecedent moral or political contamination, Foucault taught us to take that antecedent as a given of every made object, as part of the content rather than subverter of it. The reader/viewer/critic's task is to discern the operative antecedent and thereby transcend it.

The only reality of a film is the reality of the film you see. The reality you see depends in some measure on who you are and what you know, on what codes you can read. The reality the film is about is in another plane entirely. This is as true for the most complex documentary as the simplest and most direct news footage. The late documentary filmmaker Emile de Antonio said that the greatest advance in television was the inclusion of C-SPAN on basic cable as a free service. C-SPAN, he said, permitted people to see political figures being themselves at great length and usually without editing. "It's the only place on television," he said, "where you have something like direct access to what is going on. Everything else is edited. On C-SPAN, people reveal themselves."

When we see any film or video, we're never seeing the thing itself; we're seeing a flattened and framed and (except for C-SPAN) edited version of it. You never saw Jack Ruby shoot Lee Harvey Oswald in the basement of Dallas police headquarters (unless you were in the basement of the Dallas police headquarters). You saw a small grainy picture, or a tape of that small grainy picture. The physical reality in that underground driveway was totally different from the reality of your livingroom or bedroom or wherever you've had the TV the many times you've seen that action replayed. TV transmission flattens audio amplitude so what is soft is brought up and what is loud is brought down, and the real noise was stereophonic and it went to higher and lower frequencies than the tiny speakers in most TV sets can utter. There were odors of cordite and fear-sweat and exhaust fumes from the car Oswald never reached. And there was the terror experienced by everyone there of not knowing what was going to happen next: we know it was Jack Ruby firing a snub-nosed revolver and only Jack Ruby, but when that first shot rang out and echoed in that cement and tile space no one had the security of that knowledge. None of that noise or odor or terrific confusion was present on our TV sets: what happened next was we called out, "Hey Fred" or "Hey Alice, they just shot that Oswald guy," and we watched until the replays and commentaries got boring and then we did something else in the same rooms we'd been in when it all started.

Which is to say that the news or documentary footage we see may be grounded in the real, but in its mode of existence is far closer to The French Lieutenant's Woman or The Man Who Would Be King than to real life.


The postmodern Reflexivity question now so much an issue in social science and among social science critics of documentary work has been a common subject of discussion among documentary filmmakers for at least 60 years. The Soviet filmmaker Sergey Tretyakov, for example, made this comment in a symposium published in New Lef in 1927:

I think that to distinguish fiction from nonfiction (the terminology is arbitrary), we must keep in mind that there is a gradation in the falsification of the elements of which the film is made. I define falsification as the arbitrary distortion, the displacement of genuine elements. We find this distortion, first of all, in the material (what to photograph?—the selection of what we need from the whole mass of things before us); secondly, the distortion of the material by the placement of the camera and the selection of lighting; and thirdly, by the director's montage. [Tretyakov et al 30]
Both documentary and fiction films are products of skilled artifice. Both are made, edited, subject to selection and choice by an intelligence outside the events we subsequently see on the screen. In both, directors and camera operators decide when shots shall start and stop, when it's been gotten well enough to go on to something else; editors subsequently decide what portion of the exposed footage shall be used and in what order.

But there is a difference in intention that modulates all those choices, a difference similar to firing a gun at a target and at a person. In both instances the same techniques and technology apply; the intended results are radically different, and those intended results determine the kind of weapon and charge used, when the trigger is pulled, and what happens to the target and shooter afterward. In film, the difference is this: fiction film exists to take us out of the non-film world; documentary exists to increase our intimacy with it. Editors of fiction films work very hard to see that nothing breaks the fabric of the aesthetic experience, the internal harmony of the film; editors of documentary films work very hard to see that the thing out there survives the editorial process, that nothing destroys the connection to the quotidian. That difference is radical.

John Huston was proud of his ability to limit the damage editors could do to his films: he shot fewer alternate takes and cover shots than any other Hollywood director. But even fiction films by less competent directors are made in the shooting and tuned in the editing. Documentary films, except for those where the visuals are shot only to fit a script already written, come into being in the editing, which is why most documentaries have a higher shooting ratio than most fiction films and why many documentary filmmakers do their own editing, or work intimately with their editors.

The date of a documentary film is at least as much the date of the editing as the date of the event, the date when the images were made. The images in Alain Resnais's Night and Fog are from the 1930s and early 1940s, but the political and aesthetic ideas of order informing the film are from 1955. The view of Lincoln Brigade volunteers in The Good Fight was 1983 not 1936. Except for a few scenic shots and the comforting narratives provided by Shelby Foote, the images and words in The Civil War were entirely from the 19th century, but the sensibility driving the film is 1990; it surely would have been a different film had it been made at the height of the Vietnam war rather than deep in the Reagan\Bush presidencies. .

And this is as it should be, as we expect in a history book but often forget in a film because the events in a film seem to be happening before our eyes. They aren't. Nor did the film happen before the eyes of the people who shot the film and recorded the sound. They were just getting the components that would be there for the editor. Their film and sound strips bear the same relation to the finished film as the stoneground flour I buy at the mill and yeast I buy at the co-op has to the French bread I will serve at dinner this evening. Those elements are essential, but they are only possibilities until the sorting and measuring and mixing and baking occur.


Perhaps you remember the famous lines from Archibald MacLeish's "Ars Poetica": "A poem should not mean,/ But be" (1926). We can talk about meaning and structure and all that, but none of those conversations matter unless the poem first works as a poem. Ditto for film or any other documentary work: if it doesn't first work in terms of its own aesthetic (as a film, as a photograph or suite of photographs, say) the discussions are merely academic.

Earlier, I listed a number of factors influencing choices made during filming. During editing, one of the major forces is the film experience itself. This is totally exclusive of the truth value of the documentary being made. There are practical reasons for filmmakers' or editors' insistence on making major decisions in terms of film exigencies. It's not just a matter of filmmakers privileging their craft and art over truth and logic and authenticity, and neither is it the blanket "tyranny" and "ideology" for which Rosenstone faults us. Our audiences know a great deal. They know when a film is well or clumsily made, they know when the film is letting them experience something before their eyes and when the film is forcing them to make excuses. When the audience has to make excuses the linearity of film time is sheared, and the spell is broken.

Peter Fonda tells a story of his father spending hour after hour in the garden practicing a fast draw for a movie. A friend said, "This is crazy, Henry. Just have them undercrank the shot." That is, shoot it slower so on projection the movement will appear faster than it really was. Henry Fonda said he couldn't do that because the character was wearing a Phi Beta Kappa key on his watchchain and the key would swing faster than it should. "So what," his interlocutor said, "nobody's going to notice that." "I know that," Fonda said, "but they'll feel it."
 Feel it they will. And more: without even reflecting on the matter, they feel the difference between kinds of films even if they can't say what those differences are. Sit a group of ordinary television watchers down and show them an ethnological film about Eskimo dancing and they'll watch it one way; sit them down and show them a 60 Minutes segment on Eskimos and they'll watch it another; and show them a fictive film about Eskimos and they'll watch it yet another way. They code-shift exactly as they code shift for different kinds of verbal language, effortlessly, without conscious thought. Perhaps not one in a hundred or thousand could explain or detail the differences in those filmic modes, but nearly all will know them. Satisfying that knowledge is part of the filmmaker's formidable task. Documentary filmmakers work in the language of film. Documentaries that are just illustrated lectures are almost invariably lousy films.

Scholars in the social sciences and humanities, I should point out, do exactly the same thing. We accumulate far more primary data than we ever put in front of editors of learned journals or books. We all edit out our rough stuff and tune our public presentations to the expectations and skills of our intended audience. Production of documentary objects, just as production of scholarly objects, is a collaborative enterprise, in which the mind of the user is ever in the mind of the maker.


The New Social History has been paralleled by similar movements in anthropology and folklore and sociology. There has been, I think, a reconceptualization and revitalization of the provenance of "truth" in the social sciences, a tendency to take comfort in the aesthetic freedom of the humanities rather than the deceptive mathematical or clinical certitude of the physical sciences.

I think that documentary filmmakers and scholars can profitably co-exist if both keep in mind that we may look at the same or similar things but we are not speaking the same language when they we tell or show everybody else what we think we saw and came to understand. We can coexist if we both understand that the physical objects we produce have profoundly different and equally legitimate ways of making the complexities of the world accessible to our readers and viewers. Just because we come up with radically different statements when we examine what seems to be the same thing doesn't mean either of us is wrong. It means only that the world is as complicated as we thought it was all along and that we need as many kinds of responsible vision as we can get.

Finally, I should note that I've been more than a little uncomfortable generalizing so much about things every one of which is different from every other. Edmund Burke once said that Sir Joshua Reynolds "was a great generalizer" and that "this disposition to abstractions, to generalizing and classification, is the great glory of the human mind." William Blake's comment on that was: "To Generalize is to be an Idiot To Particularize is the Alone Distinction of Merit—General Knowledges are those Knowledges that Idiots possess" (Erdman, 630).

I hope I haven't offered you generalizations that idiots possess. I think the generalizations I've made about documentary work are true, but I know that what makes this work interesting, what makes it valuable, what makes the best of it windows into the heart as well as the mind, is the stuff that you have to deal with when you take them one at a time, just like people.

(An early version of this paper was given in the NEH Humanities Series entitled "Rethinking the Past: The Lessons of the New Social History" at the Johns Hopkins University, 31 October 1990. I want to thank Diane Christian, Sarah M. Elder, Michael Frisch, Bobbie Louise Hawkins, and Lawrence W. Levine for their suggestions.)


Erdman, David V., ed. 1970. The Poetry and Prose of William Blake. Garden City: Doubleday. 630

Frisch, Michael. 1990. Shared Authority: Essays on the Craft and Meaning of Oral and Public History. Albany: SUNY Press.

Trinh, Minh-ha. 27 February 1993. Comments after screening of Shoot for the Contents, SUNY/Buffalo.

Rosenstone, Robert A. 1988. "AHR Forum: History in Images/History in Words: Reflections on the Possibility of Really Putting History onto Film." American Historical Review 93(5): 1173-1185.

White, Hayden. 1988. "Historiography and Historiophoty." American Historical Review 93(5): 1193-1199.

White, Shane. 1991. Somewhat More Independent: The End of Slavery in New York City, 1770-1810. Athens: University of Georgia Press.

"Symposium on Soviet Documentary: S. Tretyakov, V. Shklovsky, E. Schub, and O. Brik." 1979. Trans. Elizabeth Henderson. In Lewis Jacobs, The Documentary Tradition, 2nd ed.: 29-36.

copyright 1994 Bruce Jackson