On the Documentary

Last day in Johannesburg. On the terrace next to our hotel I take a photo of one of the approximately 15 chairs around a long, fixed table. A chair cushion has been fitted onto a tin bucket. Three wooden bars screwed onto the bucket serve as legs. In order to increase the chair’s stability, the bars have been bent outward. Probably under steam—the same method by which the components of Vienna coffeehouse chairs are bent. The handle has been left on the bucket. This handle is used to pick up the chair and carry it indoors when the place closes.

I’m afflicted with a fever for detail. I want to adjust an imbalance.
—Harun Farocki’s diary, April 20141

Think of Alan Turing, who avoided defining intelligence, seeking instead the moment when a test subject could no longer tell the difference between communicating with a person or a machine. What is held to be intelligence can be considered as such.

By the same token, I want to avoid defining what documentary is and what distinguishes it from fiction. Instead, I want to look at what is effectively documentary or non-documentary for film itself.

On German television for a while now there has been an advertisement that appears before the news and ends with a man in a white coat talking into the camera. It’s amazing that people still believe that a white coat can establish credibility—what’s new about this video is that the camera isn’t fixed, but sways back and forth slightly.

Apparently the camera isn’t attached to a tripod but is probably placed on a Steadicam mount. There is no practical reason for this, only a rhetorical one. So, why wasn’t this film shot on a tripod?

—Much is filmed today with floating mounts such as Steadicams in news broadcasts and reportages. The lightly swaying image is meant to indicate that the event being documented has occurred spontaneously and singularly, rather than something staged for the camera.

—The viewer is intended to believe that the man is not a paid performer, and that he speaks voluntarily in favor of the slimming product.

—I noticed this for the first time in the 1990s. In the series NYPD, I think. At the beginning of a shot there was often a moving-in, as if the camera had noticed almost too late that something was happening that needed capturing.

A few years ago I noticed in an episode of the famous German police drama series Tatort that many of the shots began with the camera zooming in on something, as if it found the frame only a little too late. An acquaintance working in post-production told me there was a software that could produce this effect quickly and easily.

In The Death of Mr. Lazarescu (Cristi Puiu, 2005), Lazarescu sits in a chair and makes a phone call. After a while he shifts his weight. The camera pans in the direction in which Lazarescu’s torso has moved and back again. I play back this camera movement again and again, like one does with footage of a controversial football moment. I come to the conclusion that Lazarescu’s small movement can’t convince the cameraman that the actor wants to stand up and move to the left. The cameraman simply pretends not to know what the protagonist is going to do and to read his movement incorrectly. In other scenes in this film it can clearly be seen that the actors’ movements have been agreed on and rehearsed—this is particularly so in the scenes in which Lazarescu is brought to the overcrowded emergency rooms of several hospitals. A carefully developed choreography entangles the one with the many. There are many feigned false pans in this film.

In the feature film The Death of Mister Lazarescu (2005), director Cristi Puiu’s camera is choreographed to take on the improvisational quality of documentary-style filming.

Something similar happens in The Son (2002), a film by the Dardenne brothers: the camera briefly follows a person, then corrects itself and returns to the leading actor, who then makes a move and is followed. Olivier is a carpenter working in a training workshop for juvenile offenders. We have already seen that the news of the arrival today of a new trainee has disconcerted him. After instructing one of his charges, he slips away and hurries to the office to watch furtively through a window as the new trainee is admitted. We soon find out the significance of the new arrival. Five years before, he had broken into Olivier’s car to steal a radio, without noticing Olivier’s child inside it. The child held on to him, and he strangled it to death.

Olivier knew that the person who killed his child was in the building. He wanted to go to work as usual—but then changed his mind and went to the office. The corrective pan, briefly to the trainee and then back to Olivier, corresponds to Olivier’s change of mind. The camera says: but let’s not tell more about this everyday business, let’s come back to Olivier, whose mind is elsewhere. The pan expresses a cinematic nervousness. It can be understood as a stylistic device, like the way a swish pan is sometimes used to express that something unexpected is happening.

A pan like this marks the presence of the narrator. A noticeable narrator gets in the way of immediacy.

In feature films—classical feature films—the camera anticipates. In the documentary, the camera pursues. In the classical feature film, the camera knows the staging—the screenplay, the construction plans for the studio, and the rehearsals with the stand-ins and actors. The camera knows the production text and speaks it without faltering. This fluency corresponds to the continuity that applies to the sequence of shots.

Our new video camera can do something extraordinary: it gives us back something we missed.2If someone starts to speak or comes in through the door, and the camera is only recording on the third word or by the time the door is already half-open, it adds the first word or the door beginning to open to the recorded footage. This works because it continually shoots and stores in standby mode.3

Let’s ignore the fact that the cameraman Ingo Kratisch, with whom I have worked for a long time, and I are secretly proud of being able to predict the order of events, and that the three-second bonus this camera gives us starts to further devalue this craft. The question is why the documentarian aims to present a complete sentence or movement, and why he wants to prove his foresight. Meanwhile feature films and television series increasingly use stylistic elements originating in documentary film, such as wobbly cameras or pans and zooms to correct the image crop. These are used to pretend that the camera cannot predict—or cannot completely predict—the events happening in front of the camera.

Eddie Sachs was a racecar driver who repeatedly tried to win the Indianapolis 500—until he died in a crash. The director Richard Leacock made a film about him. A few years ago, at an event in the Austrian Film Museum in Vienna, I heard Leacock say, or rather assert, that he had never held a conversation with Sachs. He just sat next to him in his car, and Eddie began to talk to him. Leacock was certainly aware of his own sophism; for him it was one of the rules of Direct Cinema not to conduct interviews. Direct Cinema looks for events that take place like a narrative and don’t need commentary from their protagonists. However, in Leacock’s film The Chair (1961), about a man sentenced to death and his lawyers’ ultimately successful fight for a pardon, there is in fact a commentary. It was apparently inserted on the request of the television channel, which didn’t want to subject its viewers to a film entirely without commentary.

There are many writers whose names are known by anyone who has ever read a book about literature—even without reading the authors themselves. Likewise, anyone who has read half a book on documentary film knows that Leacock developed Direct Cinema along with D. A. Pennebaker, the Maysles brothers, and the producer Robert Drew, but has almost certainly never seen a film by Leacock.

However, here we’re talking about controlled uncertainties—comparable not to the real but rather to the feigned stumble, to the actor’s deliberate slip of the tongue. Documentarians want the effect of imperfection but without a demonstration of their clumsiness.

Albert and David Maysles pose between “Big Edie” and “Little Edie” Bouvier Beale, the subjects of their 1976 documentary Grey Gardens.

We documentarians often make Direct Cinema films. We look for events that occur as if they had been staged for a film. At the same time, we have to prove that we have found something and recorded it without writing or staging it. We might montage a sentence without the first words, or film a door half open—preferably not due to constraint, but to calculation.

On the evening after a documentary shoot I often take special pleasure in the spoils. Having filmed a good scene proves we were able to predict when and where a particular event would take place. In this sense we’re more similar to the hunter than to the artisan who makes something. We don’t make pictures; we take them.

When we filmed a military exercise at the Twentynine Palms Marine base in 2010, in an Iraqi or an Afghan maneuver village made of cargo containers covered with plastic tarp, one of the extras told us he and others were going to carry out an armed attack at 18:00 hours. He discreetly showed us the direction from which his commando would appear. We set up the shot and let the camera run. We framed an open-air table and a few men eating in the foreground. A couple of minutes after six, two “insurgents” arrived and fired wildly. The men at the table ran away, although one of them returned to take his food with him.

We weren’t surprised by the attack: we didn’t jerk the camera when we heard the shots, and we didn’t switch it on too late and miss the beginning. We knew the script. For the three-day training maneuver there is a “scenario” that lists for this date: 16:00: “Bank robbery” (we had missed it); 17:00: “American guards are approached and harassed by the locals” (we shot some of this); and 18:00: “Insurgents carry out an armed attack and then flee into a mosque.” We made a documentary film, even a Direct Cinema film.

People interrupt each other during documentary shooting; it isn’t exactly predictable who will speak when. The camera often pans to someone who doesn’t speak. This also has an acoustic effect. The boom swings over, and the first words of the sentence are quiet. One always tries to equalize this in the mix, even though it would be a nice effect. A sound that only fades in after a while would give a more subtle signal than a camera that corrects the frame or sharpens the focus.

So the camera has to be aimed at the door or the right person before the shot begins. I’m reluctant to tell someone to start doing something specific at our command. When asked what distinguished a fictional film from a documentary, I used to say that at least I knew the difference between the directors in each genre: the one could earn a swimming pool with his work work, and the other couldn’t. I soon had to concede that Michael Moore could certainly earn pools with his films.

This text appeared in French in the March 2015 edition of the journal Trafic: Harun Farocki, “À propos du cinema documentaire,” trans. Pierre Rusch, Trafic no. 93 (spring 2015). This version was translated from the original German by Michael Turnbull.