Although Marguerite Duras (1914-1996) is among the most celebrated writers of the 20th century, her film work has long been hard to see. (She notched up an astonishing 19 titles as director beginning in 1967, most of them feature length.) At last, a French label called Benoît Jacob Vidéo—named after the Paris street on which she lived and run by her son Jean Mascolo "to save and protect a large part of the cinematic patrimony" of the writer—has re-released three of its seven previous VHS packages on DVD: India Song (1975), Les Enfants (1984), and the four interrelated shorts from 1978-'79, Césarée, Les Mains négatives, Aurélia Steiner (Melbourne), and Aurélia Steiner (Vancouver).

From disc to disc, the quality and quantity of the bonus features varies, not to mention the absence or presence of subtitles in various languages (English, Spanish). But I was particularly impressed by what emerges if you try to access any particular scene or moment of India Song, which deserves its reputation as Duras’s greatest cinematic work: every single shot in the film—some of them lasting several minutes, it’s true—gets its own chapter! Moreover, each chapter number comes with a reproduction of the corresponding page in Duras’s very literary découpage (in French) of the piece.

And why not? This delightful division of India Song shows exactly how a film can be—should be—broken down into its constituent parts. With the caveat that, as each film’s formal structure is necessarily unique in its fine details, the division, each time, would necessarily be different. And herein lies a world of possibilities that has scarcely been touched by the producers of DVDs, even among the most distinguished and scholarly labels.

When VHS began to be phased out for DVD, a common expression of relief among critics, teachers, and fans of cinema was: "Now we can find a scene instantly, without all that fiddly back and forth in fast speed!" Indeed, this notion of the detachable, “anthological” scene, instantly accessible and repeatable, became, for a little while, the veritable rallying cry of a new wave of 21st-century cinephilia: all of us had become, in our lounge rooms, lightning “VeeJays” of the cinematic fragment, sampled at the speed of thought.

But who is doing the preliminary cutting-out of these scenes, fragments, and samples for us, before we even get to them? As with so much of the DVD revolution, the technological and industrial realities of what is delivered—what we purchase—are rather more banal than the best-case scenarios that our utopian imaginations have feverishly conjured.

"Chaptering" is among the least studied—and most universally used—of DVD operations. It is, very frequently, a highly frustrating business, whether for scholars or ordinary consumers. Let us take a random, typical example: Ang Lee's stirring Crouching Tiger, Hidden Dragon (2000), from Columbia TriStar Home Entertainment. The chapter index will take you to a series of still images, each standing for a certain "slice" or segment of the film, of indeterminate length. Unless you have already memorized the movie, these visual aide-mémoires are perfectly useless, unhelpful, and unenlightening; invariably, you will have to start skipping through the chapters manually to find the bit you seek. It may still be quicker than fast-forwarding an entire VHS copy, but it is hardly a new-technological "open sesame."

Let’s say you want to see, or study, the marvelous "tavern fight" scene in this film. You cannot access it in a single chapter; it is strung out over two. This is the commonest absurdity we find in chapter divisions: the simplest, most obvious and transparent units in a film’s constructions—its scenes, meaning its unified time-space "blocks"—are regularly violated, and to no better purpose (one presumes) than to zero in on certain special-effects or moments of high action that might punctuate that scene.

The cheaper the DVD authoring, the sloppier its divisions. Some discs give the impression that they have been chopped up on a strictly mathematical basis: five parts of roughly 20 minutes for a 100-minute film, for instance. A common, slightly more elaborate tactic is to give each chapter a jaunty little title—but again, the purpose of this is usually less than clear. Indeed, the zany, surrealistic charm of many such chapter titles would serve as an excellent exercise for a screenwriting class: imagine the plotline that would coherently string together these poetic chapter titles (sampled from the Magna Pacific release of Mark Lester’s 1997 horror thriller Misbegotten, written by B-film maestro Larry Cohen):

1. Fully Insured
2. Wanting a Baby
3. The Donation
4. Insemination
5. Who Is She?
6. For Poetry’s Sake
7. Curious About Him
8. You Both Belong to Me Now

In a DVD, there is—theoretically—no limit to the number of ways a film can be encoded so as to be accessed in sections. It is only convention—a rather dull, purely instrumental convention—that dictates that we have only one pathway, via chapters arranged according to linear plot unfolding, into the movie. In the present situation, it is bizarre that the mark of a quality DVD label—a Criterion or an Edition Filmmuseum (Vienna) or Intermedio (Spain)—is that they give us (mercifully) a chapter division that actually corresponds to the film’s own clear scene division. But so much more remains to be done.

There is a complete disconnect between the practice of chaptering and its theory—partly because that theory, as it was developed and practiced throughout the 1960s and ’70s under the title of “segmentation,” has been largely forgotten today. From Christian Metz to Raymond Bellour, and with much tweaking all over the world by such expert narratologists as Brian Henderson, films were minutely carved up into their constituent units—and all the divisions required elaborate (sometimes laborious) metacritical justifications. What is the difference between a scene and a sequence, as the collocation of several scenes? How do we tabulate passages built out of the alternation between two or more scenes? What about the force and meaning of classic transitional devices—an overlap or superimposing dissolve, as opposed to a stark fade-to-black? At stake, finally, was an entire conceptualization of narrative and its material, cinematic articulation. Bellour spent several years, in seminars, carving up Vincente Minnelli's Gigi (1958) in this way: from its largest-scale sections to its smallest-scale instants, he used this process as a way to notice and explore a film’s formal shapes, its insistent repetitions and meaningful transformations. But we can only dream of a new DVD edition of Gigi from Warner Bros. that would apply Bellour’s painstaking segmentation to its chapter breakdown.

That’s a real pity, because DVD, as a format for viewing and studying, has the potential to bring the element of process—and discovery—back into the lively intellectual adventure of segmentation. The problem with reading an essay or book that depends on a segmentation—usually summed up in an imposing table—is that one simply takes it as given; what’s missing are the steps, the experiments, that led to the choices made. Although, to my knowledge, it has yet to be done, a DVD could be arranged propositionally, offering one possible version (and hence understanding) of a film’s structure, or indeed several. And the viewer would always have the raw material—the film watched and experienced in its integral, uninterrupted form—against which to evaluate and question this proposition.

If there is a broad-brush, slightly vulgar echo of yesterday’s careful work on segmentation today, it is to be found in screenplay-writing manuals: all those arguments about whether the typical mainstream, classical, story-driven film of today should be divided into three acts, four parts, eight sections, or some other easily applicable system. But wouldn’t it be (at the very least) interesting to have this very diversity of divisions made evident, or tested out, within the chapters of a DVD? If we could see, at a glance, what the latest Spielberg or Shyamalan looks like when submitted to the analytical tools of Syd Field, Linda Aronson, Christopher Vogler, or Kristin Thompson? Perhaps these experts could even be invited, in an audio commentary to justify their slicings of a narrative into these large-scale parts. Remember, there is no technical reason for any DVD to ever restrict itself to only one such commentary.

Let’s further imagine a myriad of ways that a film can be segmented, and hence accessed, beyond strict narrative determinations —and let’s expect them all, in an ideal DVD edition, to be made available to us at the press of a button. Wouldn’t it be great if students of film style could line up, for public demonstration, all the close-ups or tracking shots in a film—without having to re-edit the whole damn thing themselves? If they could array certain gestures—upturned hands, or slouching postures—in a suitably histrionic silent film? If the present-day theorists of “the moment” in cinema, cut free of narrative or even thematic strictures, could be given the liberty to conjure such instants from the “support” of the film itself, without having to go to Godard’s extent (in Histoire(s) du cinéma and its spin-offs) of extensive avant-garde appropriation and reworking?

One occasionally stumbles upon a particularly ingenious and creative DVD segmentation. In the case of Punch-Drunk Love (2002, Columbia TriStar), for example, I can only imagine that director Paul Thomas Anderson himself made the radical decision to slice up certain sections of the film according to its musical score, not its traditional, narrative scene divisions. Look at and listen to, most strikingly, the transition between chapters 13 and 14: the split comes in the middle of a continuous event (Adam Sandler’s character cracking up in his office) but marks the appearance of a special musical "moment," the quotation of Shelley Duvall’s song "He Needs Me" from Robert Altman’s Popeye (1980). This decision—annotated nowhere on the disc, since Anderson eschews audio commentaries—reveals an aspect of the properly post-classical logic of a sublimely nutty and hyper-formalist movie: the music, sometimes straddling several scenes and battling it out in the mix with other sound sources, truly drives and structures the whole.

Let us broaden this discussion beyond the particular issue of segmentation. Talk of "multiple pathways" and the like will probably arouse an aversion in some DVD producers toward "pedagogic" or scholarly editions of films—presumably deemed a turnoff in the general marketplace, suitable for and desirable to only a small, elite, niche group. (In a parallel sphere of culture, teachers of literature complain that today’s students are increasingly resistant to reading heavily annotated editions of classic poems, plays, and books.) Often, this stingily "pragmatic" logic limits what can be conceived. It is said, for instance, that a Histoire(s) du cinéma DVD cannot be adequately subtitled in English, or indeed any language other than French, for the screen would be filled to bursting with words rendering, all at once, Godard’s voiceover narration, text printed on the screen, and also sometimes the dialogue from clips that plays underneath, or alternately, in the mix. It’s a fair enough argument—and one assumes it is precisely the rationale behind the very piecemeal subtitling currently offered to us by Gaumont. But that logic vanishes if we posit the subtitling as a series of “passes” or layers that could be called up separately—plus, as the special Japanese edition provides, another layer to identify the many quotations (of all kinds) in the work. Good news: the Australian company Madman has recently committed itself to a future edition of Histoire(s) that reconstitutes its multiple layers for English-language audiences.

Ultimately, where do ubiquitous "special editions" end, and "scholarly editions" begin? The line is becoming increasingly blurred, as the recent Universal 50th Anniversary release of Welles's Touch of Evil (1958) shows. What, paradoxically, allows some DVDs to load up on intellectual extras is the built-in indifference factor: bonus tracks are a strictly take-it-or-leave-it affair. No one is forced to watch them; their mere presence on the menu, or filling up an added disc, is enough to provide the necessary “surplus value” deemed attractive for consumption. And so we must maintain pressure on DVD companies, in order to exploit what is indifferent and make it essential.