3. Mise-en-scène and Its Discontents

Russian Ark

Keeping this convoluted history in mind, I want to examine now at what is overlooked by the historical tendency to associate mise-en-scène with the long take, and to oppose it to editing. Because I believe that these traditional oppositions and associations limit our understanding of the richness and artistry of the cinema.

For starters, let’s look more closely at Bazin’s argument that the long take is better than the short one for representing reality. A commonly heard argument here (one still hears it today) is that whenever a filmmaker cuts, he or she is guiding the viewer’s attention, and forcing them to look at particular things in particular ways. By way of contrast, Bazin argued that long takes allowed viewers more freedom—they could look where they wanted. This contributed to the idea that long takes are somehow more respectful of film viewers, and as such require more sophisticated viewers. Over time, this created the kneejerk association that long takes are somehow smarter than shorter ones (an idea that lives on in the attacks on Michael Bay).

But is this argument necessarily true? There are many reasons to doubt it. For one thing, all shots, long and short, are equally artificial. It simply isn’t the case that as a shot gets longer, it somehow gets truer. To think that way overlooks the artifice of the long take.

For instance, Bazin’s arguments about how long takes were more respectful or less manipulative than shorter ones don’t always hold up to scrutiny. As it turns out, there is nothing stopping long takes from being just as composed and manipulative as shorter ones. Directors have many tools at their disposal to direct the viewer’s attention through the long take, just like they do in shorter ones. Composition can be, and often is, a means for directing attention. So, too, are performances and camera movements. In other words, there’s no reason to assume that mise-en-scène is any less “manipulative” than editing.

This point is well made by David Bordwell in his article “Widescreen Aesthetics and Mise-en-Scène Criticism” (1985, available here as a PDF). In particular, Bordwell observes how Otto Preminger’s use of widescreen was celebrated by certain critics operating in the Bazinian tradition. He relates how two critics writing in the Bazinian tradition, V.F. Perkins and Charles Barr, praised a scene in River of No Return (1954) in which Marilyn Monroe’s character, Kay, drops her valise while boarding a raft. As the scene continues, we see the valise drifting away in the background of other shots. The argument here is that Preminger has left it up to the viewer to see this detail, even as the action continues in the foreground. Both Perkins and Barr celebrated Preminger’s employment of long takes and deep focus, arguing that they gave his films a kind of naturalism, transparency, and subtlety.

But Bordwell argues that this isn’t the case at all. In his analysis, he notes how the film actually employs several devices that function to draw attention to the disappearing valise:

“When Kay drops the valise she glances frantically toward it and cries out, ‘My things!’ Harry shouts, ‘Let it go!’ At the same moment, the camera pans sharply to the right to reframe the valise, and a chord sounds on the musical track. Our attention to the drifting bundle is just as motivated. For one thing, the bundle is initially centered when Matt and Henry pass. Furthermore, Preminger has anticipated this camera position a few shots earlier, when matt ran to the edge of the bank. It is common for a classical film to establish a locale in a neutral way and then return to this already-seen camera setup when we are to notice a fresh element in the space. We thus identify the new information as significant against a background of familiarity. As a fresh element in a locale we have already seen from a comparable vantage point, the bundle becomes noteworthy. In sum, Preminger’s staging of the scene stands out because it avoids editing, but it uses other means to draw our attention to the bundle—centering, the return to a familiar setup, and the repetition of cues for the bundle’s loss.” (22–3).

So much, then for Preminger’s supposed naturalism and transparency. His long takes and use of deep focus—hallmarks of Bazinian realism, and supposedly free of manipulation—turn out to be saturated with artifice, and highly manipulative. (Preminger is hardly the only example one can find of this—Citizen Kane, for instance, also uses composition and sound cues to focus its viewers’ attention, in addition to its celebrated usage of long takes and deep focus.)

Another problem with the critical tendency to oppose long takes and editing is that it ignores the many ways that those two techniques commonly work together. This point is well made by Brian Henderson in his 1976 essay “The Long Take,” which seeks to deconstruct the false binary between Bazin’s long take and Eisenstein’s montage. For one thing, Henderson points out that long takes still have duration, beginnings and endings, and as such still employ editing—they’re edited together. What’s more, even a director like Max Ophuls—truly a master of the long take if there ever was one—rarely assembled his films out of nothing but long takes. Thus, a film with many long takes may also feature shorter ones, and those shorter takes may in fact come between long takes:

“The present article takes its chief emphasis from the fact that the long take rarely appears in its pure state (as a sequence filmed in one shot), but almost always in combination with some form of editing. […] Most analyses of long take directors and styles concentrate on the long take itself and ignore the mode of cutting unique to it—what we call below the intra-sequence cut. But such cuts or cutting patterns (one could even speak of cutting styles) are as essential to the long take sequence as the long take itself.” (316)

Throughout the essay, Henderson patiently draws attention to these problems in order to ultimately argue that a film criticism that simply opposes long takes and editing is bound to overlook the crucial role that editing plays in defining the long take, and sequences of long takes. His goal is to point out an area of filmmaking that has largely gone unstudied. Sadly, the tendency to diametrically oppose mise-en-scène with cutting prevails nearly forty years later, leaving a fascinating realm of cinema still largely unstudied. At the present moment in popular film criticism, the championing of long takes has once again risen to something of a fetish. It receives a disparate amount of attention despite the fact that the long take is but one element of filmmaking, no better or worse than any other.

A related problem is the tendency to measure the length of a film’s takes by calculating the Average Shot Length, or ASL. This value is calculated by dividing a film’s running time by the number of shots it contains. And ASL is a very useful value in many respects. For one thing, when one surveys many different films, ASL can give a general sense of how rapidly films are cut in a given place or time. Thus, one can say that the average rate of cutting in Hollywood cinema has increased throughout the sound era. Or, one can attempt to catalog which contemporary Hollywood films feature the longest ASL’s—as I did here, in an earlier article for PressPlay.

But ASL also leaves out a lot of information, especially when one is analyzing specific films. Alfonso Cuarón’s Gravity, for instance, has an ASL of roughly 35, since it features 156 shots in 90 minutes. (I’m speaking approximately here; I’m going off reported numbers, and haven’t performed this analysis myself. Also, I don’t know the actual runtime of the film, sans credits. The exact ASL, however, is beside the point.) Having in hand an ASL of 35 doesn’t mean that every shot in Gravity is 35 seconds long—or indeed that any shot in Gravity is 35 seconds long. The opening shot, for instance, is at least twelve minutes long—meaning that, on average, the remaining 155 shots have an ASL of 30 seconds. And for every shot longer than that, there means there are also shots shorter than that. What of them? Are they any good? What is their relationship to the longer shots in the film? Or is Gravity only good during its long shots? (And if that’s the case, then why?)

The fetishizing of long takes is part of a larger, long-running problem in film criticism, which as a whole is arguably less critical than it pretends to be. As David Bordwell has expressed it:
“Instead of asking how films work or how spectators understand films, many scholars prefer to offer interpretive commentary on films. Even what’s called film theory is largely a mixture of received doctrines, highly selective evidence, and more or less free association. Which is to say that many humanists treat doing film theory as a sort of abstract version of doing film criticism. They don’t embrace the practices of rational inquiry, which includes assessing a wide body of evidence, seeking out counterexamples, and showing how a line of argument is more adequate than its rivals.” (“Articles“)

Put another way, the fascination with the long take risks becoming entirely symptomatic, and uncritical. What makes a movie good? Long takes! How do you know which movies are the best? Why, just check which ones feature the longest takes! This is a totally dumbed-down type of film criticism, where all we need do is calculate ASL’s in order to rank all the movies ever made.

I don’t want to imply that long takes aren’t important, or don’t feature a special relationship with mise-en-scène. Certainly we should be sensitive to the unique challenges and properties posed by the long take, and how it presents its content to the viewer. No film better illustrates this than Aleksandr Sokurov’s feature Russian Ark (2002), whose 96 minutes of footage consist of a single take. I myself watched the film twice in a row in the theater, something I’ve rarely done—but Russian Ark is truly an atypical film.

However, is Russian Ark somehow more realistic than films that feature editing? Hardly. It’s worth remembering, a la Henderson, that the 96-minute-long shot, and the film itself, still has a beginning and an end. When compared with a person’s life—or even a single day—it is still but a miniscule slice of time, unable to compete with actual lived experience.

What’s more, we would do well to remember Bordwell’s analysis of Preminger. The film as a whole, rather than being some transparent documentation of reality, is entirely contrived. The single take carries us from room to room, and from scene to scene. We go where Sokurov takes us. And the man isn’t just wandering the Winter Palace of the Russian State Hermitage Museum with a camcorder, capturing whatever reality he finds there. Instead, he’s organized everything that we see. His camera movements and framing glide along very differently than we people do, being balanced by a Steadicam. And they continuously direct our attention, focusing it to particular aspects of the spectacle. Meanwhile, everything that appears on screen is the product of meticulous design and rehearsal. And we aren’t even seeing the first take, but the fourth! (Goodfellas’s Steadicam passage through the Copacabana is similarly no more real or less artificial than any other shot in any other film ever.)

Along these lines, associating mise-en-scène exclusively with long takes perpetuates the bias toward long takes, since they then seem to have a special quality (mise-en-scène!) that’s lacking in shorter ones. Because, I mean, if cutting eliminates mise-en-scène, then aren’t they inherently worse? But short takes do have mise-en-scène, and understanding the connection between the mise-en-scène and the montage is extremely important. To put it another way, if montage is the study of the interrelation of shots, and all shots possess a mise-en-scène, then montage is also the study of the interrelation of mise-en-scènes. This is a topic just as worthy of serious critical attention as the study of individual long takes. What is needed, overall, is a critical approach to cinema that seeks to relate the various parts to the whole (as we find in the works of critics like Bordwell and Henderson).