Lessons with Bazin: Six Paths to a Poetics

October 2018

This essay was commissioned by the German journal montage a/v, and it appeared in volume 18, no. 1 (2009), 108–128. I present it here in revised form. Thanks to Dan Morgan for his helpful suggestions.

From the 1910s into the 1960s, filmmakers, critics, and intellectuals created a distinctive tradition of writing about cinema. These writers were seeking to understand the nature, functions, and resources of film as an art form. Georg Lukács, Riccioto Canudo, Louis Delluc, Leon Moussinac, Rudolf Arnheim, and other writers contributed to this tradition, as did filmmakers like Eisenstein, Kuleshov, Pudovkin, and Dziga Vertov. These theorists advanced widely different positions, but they held some ideas in common. For instance, they tended to assume that the medium had an expressive essence that set it apart from other art forms. For Delluc and his followers, that essence lay in the mysterious quality of photogénie. For the Russians and many of their admirers, cinema’s expressive core was to be found in the technique of editing. For Arnheim, the essence of film was its abstraction from perceptual reality.

Another feature of this trend was a prescriptive bent. Both filmmakers and film theorists tended to judge some techniques superior to others, usually because those techniques were in harmony with the purported essence of cinema. For example, many thinkers held that a lengthy, static shot constituted merely passive recording. The result was “theatrical” and hence “unfilmic.” By contrast, editing was ipso facto a good thing because it manifested what cinema was uniquely equipped to do—juxtapose moving images in time and space, with a freedom unavailable to the drama. Essentialism and evaluation meshed: The best films tended to be the most “cinematic” ones.

André Bazin’s theorizing fits snugly into this tradition. His writings reflect on the fundamental nature of film, its artistic resources, and its social and political effects. His original contribution to the tradition lay in his arguments for mechanical recording as a virtue, not a deficit to be overcome. Contrary to the dominant strain of the 1920s and 1930s, Bazin believed cinema’s essence to lie in its photographic capacity, its power to lay bare the phenomenal reality of the world. As a result, he argued that the “unfilmic” technique of the long take, conjoined to camera movement and depth of field photography, respected that reality. It was editing, the technique once held to be the supremely filmic technique, that introduced a level of artifice with which directors would have to contend.

The originality of Bazin’s thought challenged other traditional beliefs as well, perhaps most radically the notion that a theory of cinema ought to be wholly occupied with the question of film as an art.1 Still, he did not reject certain assumptions of the standard position. Like his forebears, Bazin posited an ontology of cinema that defines its intrinsic nature: in his view, its recording capacity. Like his forebears, he based value judgments about films and filmmakers on the extent to which they respected that nature. In these respects he remained a classical film theorist.

Suppose, however, we imagined another strain in film theorizing. Imagine reflection and research that do not take Cinema as a whole as its object. This project would concentrate instead upon certain periods, genres, styles, trends, or other delimited phenomena. Imagine as well that this investigation holds in suspension the idea of a cinematic essence, recognizing film’s distinct (if not unique) resources while also exploring cinema’s many affinities with other media. Imagine as well that this inductive inquiry seeks out regularities of cinematic subjects, forms, and styles of a certain places or periods. Imagine that the researcher also tries to discover how certain effects are reliably produced by those regularities. And imagine that this enterprise is descriptive, analytical, and explanatory, not (at least primarily) evaluative.

What would differentiate this from the study of trends in style and form undertaken by art historians or musicologists? The search for broader principles—be they explicitly canonized rules or items of expertise that filmmakers know tacitly. I take it that this search can be labeled a poetics: a systematic inquiry into the materials, forms, and constructive principles of filmmaking within various traditions.

The most salient instance of this enterprise in the classical era of film theorizing is the anthology Poetika Kino (1927). Here prominent literary critics affiliated with the Russian Formalist school joined filmmakers to propose some general principles governing film structure and style. Granted, the essays don’t wholly avoid evaluation, and some make passing declarations about what the essential conditions of cinema might be. Yet to a surprising extent for their period, the writers try to elucidate principles of plot construction, stylistic texture, and spectatorial uptake.

Given this perspective, we can read other classical theorists as offering as a sort of poétique malgré lui. For example, Arnheim valorizes significant form as a basis for cinematic art. We do not have to accept his theory of value, or his ontology of film art, to recognize that he discovered broad principles of expressive form, such as the geometricization of shot composition and the importance of changing image gestalts in the course of a shot or sequence. These principles can help us pick out important pictorial trends, from silent cinema (Keaton, Sternberg) through to modern times (Tarkovsky, Sokurov). Likewise, Eisenstein’s taxonomies of types of montage (metric, rhythmic, tonal, etc.) may apply well to his own practice, but the principles he enunciates can be found in several filmmakers, even Ozu.2 Eisenstein’s pedagogical exercises offer an even more explicit conception of poiesis, or active making, than his better-known “official” theorizing. Vladimir Nizhny’s collection of classroom sessions, Lessons with Eisenstein, is a rich source of “practical theory,” the sort of fine-grained attention to creative choices that feeds directly into a film poetics.3 In sum, when theorists base their claims on inductive inferences and empirical claims, a poetician can take those as prods, hypotheses, or bodies of evidence warranting further inquiry.

Bazin, of course, is traditionally thought of as a realist. Yet much in his work points toward a film poetics. Reading him from this angle forces us to bracket off some of his most original assumptions about photographic realism and the strongly evaluative conclusions he arrives at. For many readers, this is too much to surrender. But I think that the effort is worth making. My own research is deeply indebted to his work, so in a sense this essay is a tribute. He might not agree with how I’ve treated his ideas; I wish he were still living to criticize me. In any event, here are what I take to be some major lessons that Bazin’s work offers to someone pursuing an empirical and historical poetics of cinema.4

Matters of Method

Bazin taught us several methodological lessons. Perhaps the most striking is his propensity for analyzing shots and scenes in exquisite detail. In his journalistic reviews, of course, such description would have been out of place, but in his longer essays, such as his classic study of William Wyler, and in his book on Orson Welles, he plunged into fine-grained analysis of a sort that was almost unprecedented in film writing.5 Most such passages are too long to quote here, but I can’t refrain from a sample, from Bazin’s account of The Best Years of Our Lives.

This scene is set in a bar. Fredric March has just convinced [Dana] Andrews to break off with his daughter and urges him to call her immediately. Andrews gets up and goes toward the telephone booth located near the door, at the back of the room. March leans on a piano in the foreground and pretends to get interested in the musical exercise that the crippled sailor (Harold Russell) is learning to play with his hooks. The camera’s field of view shows the keyboard of the piano dominating the foreground, March and Russell in the middle distance, the whole barroom around them, and, quite distinctly, Andrews in the far distance, tiny in a telephone booth. This shot is clearly built upon two dramatic poles and three characters. The action in the foreground is secondary, although interesting and unusual enough to demand our keen attention, since it occupies a privileged place in the composition. Yet the true action, the one that constitutes at this exact moment a turning point in the story, develops almost secretly in a tiny rectangle at the back of the room—in the left corner of the screen.6

And so on for six more paragraphs of nuanced description!

This level of detail remains extraordinary. Most discussions of film today, in both the popular press and academic work, don’t try for it. (Perhaps that’s partly because after all these years both critics and researchers remain largely uninterested in visual style.) Bazin had a sharp eye. He counted shots across entire films and used a stopwatch to time scenes.7 As a result, he was the only critic I know to realize that Hollywood films typically had an average shot length between nine and twelve seconds.8 Bazin’s precision is all the more remarkable in that he was largely working from memory and notes. He did not have our access to archives and editing tables, let alone video copies. He wrote his final essay, a penetrating analysis of the courtyard murder in Le Crime de M. Lange, after watching the film in a television broadcast.9 If he had done nothing else, Bazin would be remembered today for his pioneering efforts in close visual analysis.

Analysis can degenerate into shapeless description if there are no concepts to mold it. Bazin’s explications gain solidity because firm ideas organize the details. The broadest of those concepts—profondeur de champ, the long take, cinematic narration, and the like—are commonplaces today, but it’s worth remembering that he made them important tools for every critic and theorist. In most cases, however, he didn’t originate them. Where did they come from? From other writers, but also from film artisans. Hence a second methodological lesson for poetics: Listen to the filmmakers.

In early 1941, Gregg Toland wrote an article called “Realism for ‘Citizen Kane.’” He claimed that he and Welles had agreed on a visual approach that obliged them to shoot in what we would now call long takes. Apart from chiming with Welles’ theatrical experience, the sustained shot would avoid cuts.

We tried to plan action so that the camera could pan or dolly from one angle to another whenever this type of treatment was desirable. In other scenes, we pre-planned our angles and compositions so that action which ordinarily would be shown in direct cuts would be shown in a single, longer scene—often one in which important action might take place simultaneously in widely-separated points in extreme foreground and background…. Welles’ technique of visual simplification might combine what would conventionally be made as two separate shots—a close-up and an insert—in a single, non-dollying shot.10

What permitted this, Toland explained was his technique of “pan focus,” the tactic of keeping all planes of action in sharp relief. Later in 1941, he published a more general piece on the creative work of the cinematographer, and he restated this theme.

Hitherto the camera had to be focused either for a close or a distant shot, all efforts to encompass both at the same time resulting in one or the other being out of focus. This handicap necessitated the breaking up of a scene into long and short angles, with much consequent loss of realism. With pan-focus, the camera, like the human eye, sees an entire panorama at once, with everything clear and lifelike.11

The idea of optical realism, the emphasis upon sustained shots (called “scenes,” as they often were called at the time), the possibility that a single depth shot could contain the equivalent of two or more closer shots, and of course the technique of “deep-focus” cinematography—all these tenets of Bazin’s aesthetic are here in rudimentary form.

Citizen Kane opened in Paris on 3 July 1946. Its technical innovations had already been discussed in the French cinephile press of the time, and Toland’s second 1941 essay was translated in 1947.12 All critics were primed to notice the film’s technique, and Toland’s rationale, which was part of the film’s publicity campaign, gave critics several important concepts with which to work.13 For more on the campaign, see my weblog entry here.

Similarly, in early 1947 William Wyler published an article about the making of The Best Years of Our Lives, and this teases out another implication of Toland’s signature style.

Gregg Toland’s remarkable facility for handling background and foreground action has enabled me over a period of six pictures he has photographed to develop a better technique of staging my scenes. For example, I can have action and reaction in the same shot, without having to cut back and forth from individual cuts [shots] of the characters. This makes for smooth continuity, an almost effortless flow of the scene, for much more interesting composition in each shot, and lets the spectator look from one to the other character at his own will, do his own cutting.14

French critics knew Wyler’s essay, and they picked up his idea that a densely composed shot in depth gave the spectator a certain amount of freedom. In February of 1948, Alexandre Astruc declared that profondeur de champ “obliges the spectator's eye to make its own technical découpage, that is to find for itself within the scene those lines of action usually delineated by camera movements.”15 In the same month, Bazin’s appreciative essay, “William Wyler ou le janseniste de mise-en-scène,” quoted the passage above as evidence that Wyler allows the spectator “to perform the final cutting [l’opération finale du découpage] himself.”16

It takes nothing from Bazin’s originality to acknowledge that these concepts were circulating in his community. For he grasped their implications more deeply and developed them more imaginatively than anyone else. He quickly extended the idea of depth staging to Renoir’s 1930s work, again following up comments from the director.17 He distinguished Welles’ baroque use of depth from the more neutral and uninflected setups of Wyler. And as I’ll show shortly, he developed the concept of “the viewer’s découpage” of the action in highly original ways. In all, Bazin’s work offers an exemplary instance of how a poetics of cinema can pick up hints from filmmakers’ explanations of their creative choices, test those statements against the films, and elaborate the ideas in ways that point out broader principles of construction.

This brings me to my third methodological lesson, one we might call constrained generalization. Film theorists, as I’ve suggested, tended to think globally, seeking out laws that characterize cinema as a whole. Film historians tended to think more locally—sometimes too locally. The historiography of Bazin’s day treated film’s development as owing everything to national cinemas. Maurice Bardèche and Robert Brasillac’s Histoire du cinema (published in 1935), the standard book on the subject for Bazin’s generation, treated history as largely a matter of national cinemas, each with an indigenous movement or trend at a given period. The same plan was expanded to epic dimensions in Georges Sadoul’s Histoire générale du cinéma, volumes of which were being published as Bazin was starting his career.

Bazin saw continuity where these senior historians saw change, and he posited common tendencies across different cultures. Some of these tendencies were frankly speculative, as when he posited that a universal yearning for a simulacrum of reality led cinema to be invented in different countries at about the same time. More concretely, he showed that a many postwar directors in Europe, England, and the U.S. faced a common problem: How to adapt plays and novels in ways that respected, indeed acknowledged, their specific identity as dramatic or literary texts? This led him to study how modern cinematic technique could highlight conventions characteristic of other media: the theatricality of theatre in Olivier’s Henry V and Melville’s Enfants terribles, the use of ellipsis and tense structures in Diary of a Country Priest.

Senior historians had claimed that “film language” developed along a straight line, with each artist contributing to a growing fund of expressive resources. But Bazin saw a clash of norms. He identified an international film style, derived from American continuity editing, that was slightly modified with the coming of sound. Against this he located a more distinctive trend, one that was coming to fruition in his own day. Current filmmakers seemed to be rejecting the American découpage-based style in favor of something more faithful to the continuity of time and space. This trend he saw in Renoir, Wyler, Welles, and the Neorealists—linked not by country or culture, but by a shared urge to render the phenomenal world.18 Bazin’s account is thus neither universal nor narrowly local. It is fixed firmly in history but acknowledges that disparate artists can arrive at similar solutions to shared problems. By pointing out how filmmakers in different countries could converge on similar norms, he managed to create middle-level generalizations that avoided vacuity and invited refinement.

Empirical discoveries

To these three methodological lessons—scrutinize the film; listen to the filmmaker; and build constrained generalizations—I’ll add three substantive ones. The first is directly related to the search for middle-range trends.

Elsewhere I’ve argued that Bazin engineered a counter-history of film style suitable for the mature sound cinema. Through close analysis and an ingenious deepening of arguments sketched by his contemporaries, he revised the standard account in subtle and far-reaching ways.19 The product of this effort is best seen in his essay “L’Évolution du langage cinématographique,” assembled at the end of his life from three earlier essays. Other essays, along with his books on Welles and Renoir, supplement and clarify these.20

The “Évolution” essay is usually seen as offering historical support for Bazin’s larger theory of cinema. If cinema’s ontology is based in photographic recording, the history of cinema enacts a Hegelian unfolding: the medium’s development gradually reveals its true essence. It goes this way: Two tendencies of the silent cinema, sheer recording (e.g., Lumière actualites) and artifice (abstract montage), vied with one another for supremacy until an unstable equilibrium was reached in the early sound cinema. That international sound style banished the extremes of montage seen in Griffith, Gance, and Eisenstein. In the “classical découpage” of the 1930s the phenomenal reality of space and time was respected to a considerable degree. But then Renoir, Welles, and Wyler lifted the dialectic to a new level. They fulfilled the promise of cinema’s ontology through techniques like camera movement and profondeur de champ, which record phenomenal reality. These creators “surpassed the surpassing” by building into the densely composed shot all the details that would have been assembled through classical découpage. Editing now took its rightful place, used when it is necessary for more abstract storytelling (e.g., montage sequences summarizing a passage of time). The new style of the period after 1939 was the fulfillment of cinema’s essence.

Here Bazin’s evaluative preferences come to the fore. The best films are those early films that respected reality (works of Murnau, Flaherty, von Stroheim, Dreyer) and those contemporary films that delivered cinema from decades of flirtation with stylized artifice (works of Renoir, Wyler, Welles, the Neorealists).

The difficulties with Bazin’s conception of film’s essence are well-known, and the dialectical trajectory he traces can be faulted on historical grounds. For example, classical découpage is not a creation of the sound era. It was forged in the 1910s (only partly thanks to Griffith), and it was stabilized before the “heresies” of German Expressionism and Soviet Montage emerged. Nevertheless, if we decouple Bazin’s grander theoretical commitments from his investigation into the evolution of film style, we find that he has sketched a rich research program in film poetics.

His claims can be tested, corrected, expanded, or rejected in the light of further information. Has he accurately characterized the tendencies at work in the periods he sets out? What motivates the choice of his exemplary filmmakers? If we sample a wider range of films than he could, do his claims stand up?

To summarize drastically, I think that contemporary research would revise Bazin’s schema along the following lines. There was much more variety in the first dozen years of filmmaking than he could have known. Editing, close-ups, camera movement, and other “advanced” techniques were developed and, rather surprisingly, sometimes abandoned. By the 1910s, two broad stylistic norms emerged in the fiction film. European filmmakers and some in other countries, including America, refined a tableau style based in lateral and depth staging. Meanwhile, in the course of the 1910s most U.S. filmmakers abandoned the tableau and embraced continuity editing as the controlling technique. The years afterward saw the triumph of the American editing style in most film industries, although that style was subject to some local variation. I offer a video lecture tracing this argument here.

To continue my revision: The coming of sound ratified continuity norms, and camera movement was integrated into them in ways that fulfilled and expanded traditional functions. Longer takes and depth staging, to be found in every decade and many countries during the silent era, were also assimilated into the continuity style, not only by Renoir, Welles, and Wyler but also by Ophuls, Mizoguchi, Hawks, and many other filmmakers. Deep-space arrays and deep-focus filming became international trends in the 1940s and in the 1950s, at least in black-and-white work, but such imagery remained embedded in a framework of classical continuity editing. In sum, Bazin’s claims about dialectical breaks and freedom of visual exploration depend on a selective use of evidence. From the early 1910s on, international film style in mass-market cinema shows a pattern of development in which new technical devices are harnessed to powerful norms of cinematic construction.21

My account of this newer story is itself schematic, and it isn’t invulnerable to criticism. It will doubtless be revised in future research. My point is simply that Bazin offered us a new framework based on acute observation and bold conjectures. We need not accept his ontological premises or his prescriptive conclusions to find his empirical claims informative and his inductive inferences plausible. By testing and recasting this framework, the study of the history of film style has made progress.

So let’s take Bazin’s transnational story about the evolution of style as a first substantive lesson in poetics. A second lesson bears more on causes and functions. How do we explain the process of continuity and change we observe in Bazin’s scheme, or the recasting of it that I sketched? What conditions created the “evolution of film language,” and what resulted from it?

The influential historians of Bazin’s day offered two principal explanatory strategies. Bardèche and Brasillach, as befitted their right-wing alliances, saw national schools as informed by a sort of volksgeist or national character.22 Sjöström, they tell us, “flooded his sober plots with a sort of radiance, with a sort of nostalgia and all that atmosphere for which the Scandinavians have created an untranslatable word—Stemming.”23 Sadoul, man of the left, tended more towards a broadly economic mode of explanation, particularly when discussing the class appeals of early cinema and the cartelizing conduct of Hollywood.

Bazin admired Sadoul, and he took much from Bardèche and Brasillach (without acknowledging them). Yet he often preferred more transcendent explanations. The most famous instance is his claim, in “Le Mythe du cinéma total,” that the earliest inventors were fulfilling a universal dream of capturing reality. This tendency to posit a spirit or will to form, whereby individual historical actors fulfill a mission inherent in the medium, can be seen in the “Évolution” essay as well. The stylistic tendencies—“faith in the image” versus “faith in reality”—become historical forces in themselves rather than convenient labels we apply to a welter of individual decisions and institutional pressures.

Granted, Bazin was sensitive to certain technological constraints, especially of lens design and lighting. (Are these also the inheritance of Toland’s remarks on Kane?) But his comments treat these as chiefly encouraging or inhibiting the flowering of cinema’s true essence. For example, Bazin claims that the lenses of early filmmaking were designed to take in a wide view of the set, as filmmakers covered scenes in a single distant shot. Lenses of longer focal length, he implies, were developed in tandem with cutting patterns that isolated figures and threw backgrounds out of focus. “Progress in optics is closely linked with progress in editing.”24 Again, however, Bazin was subject to limitations of his period and limited knowledge of silent film history, so his remarks, however pregnant with implications, tend to treat “technical progress” as another reified force.

Abandoning Bazin’s Hegelian evolutionary scheme, later researchers have been able to disclose more concrete causal forces. In The Classical Hollywood Cinema (1985), two collaborators and I sought to show that the development of classical découpage sprang from a mix of institutional and individual action. The American cinema developed a mode of production that, in order to be routinized, based its work processes on the continuity scenario. That document served as a blueprint for other phases of what Marx called “serial manufacture.” The script broke scenes into shots for maximum control during shooting and editing; it had the equally important effect of creating a cinema of dynamic storytelling. At the same time, the discourses of the industry—chiefly the trade and professional press—converged on conceptions of quality, continuity, and other features that provided fairly clear-cut goals for filmmakers. Adjacent Hollywood institutions, like service firms and professional associations, affirmed these goals as well, even as they steered them to their own aims.

In sum, the system of aesthetic norms we call “classical film style” was created by a convergence of particular institutional forces offering historical agents a restricted set of choices, a weighted set of solutions to recurrent problems. Once the system was stabilized, new technologies, new stylistic devices, and new modes of organization could be adapted to fit the traditional structures of production and stylistic expression.

Whatever the virtues and faults of these arguments, The Classical Hollywood Cinema attempted to provide specific accounts of how institution-driven norms and craft practices shaped stylistic choices in a particular era. It is to these factors that a poetician can look for proximate causes of artistic continuity or change. Once more, Bazin has bequeathed us a set of questions rather than answers. By looking beyond national character and class-based interests, Bazin reminded us that other causes might have played a role.

Bazin floated not only causal but functional explanations for the stylistic changes he observed. This tendency accords well with a poetics of film, which seeks to explain how constructive principles are designed to elicit particular effects from spectators. He assumed that we are tacitly sensitive to style, and that even subtle stylistic choices are registered, perhaps subliminally, by spectators. He argued that showing a child in the same frame as an approaching lion was more arousing than creating the impression of that action by intercutting shots of a child and shots of a prowling lion.25 Filming an action in a full shot will serve to create a greater sense of concrete spatial and temporal realism.

Most theorists in the classical tradition assumed that pictorial devices functioned to guide the spectator’s attention. Writers who saw editing as the central film technique insisted that it worked to draw the viewer’s eye to what was most important in the scene at any moment. Bazin did not question this assumption, but he refined it in two ways. First, he suggested that classical découpage was in this respect basically an extension of theatre, enlarging this or that aspect of a homogeneous scene, much as if the viewer were scanning the stage and using opera glasses to study a part of the action. More importantly, the density of the deep-focus image, according to Bazin, worked against the linear, univocal effects of editing. Cutting forced us to look here, then there; but Wyler and Welles designed their shots so that spectators had more “freedom” to discover levels of significance, to “make their own découpage,” as we’ve seen.

Further, this activity offers a new kind of experience. Citizen Kane and The Magnificent Ambersons use deep focus to increase the viewer’s psychological investment in the action.

Obliged to exercise his liberty and his intelligence, the spectator perceives the ontological ambivalence of reality directly, in the very structure of its appearances…. [Deep-focus filming yields] a realism that is in a certain sense ontological, restoring to the object and the décor their density of being, the weight of their presence; a dramatic realism which refuses to separate the actor from the décor, the foreground from the background; a psychological realism which brings the spectator back to the real conditions of perception, a perception which is never completely determined a priori.26

Once established, this perceptual realism can yield a daring game of vision. What is in the foreground should be more important than action in the distant background, but in The Best Years of Our Lives, Wyler makes the foreground action at the piano distract us from what is taking place in the telephone booth. In The Little Foxes, the key action of Herbert Marshall collapsing on the staircase takes place in the background, but Wyler throws it out of focus. Bazin remarks that this makes us strain to see it: “This artificial blurriness augments our feeling of anxiety: as if over the shoulder of Bette Davis, who faces us and has her back to her husband, we have to discern in the distance the result of a drama whose protagonist nearly eludes us.”28

Bazin has selected some striking, probably atypical examples, but he shows that the hypothesis of guiding the spectator’s attention around the frame can yield great insights. We can go on to study how staging such scenes calls on more basic principles of steering the viewer’s visual exploration (frontality, position in the picture format, and so on). We can also study something that Bazin neglected: the ways in which dialogue can highlight figures in a densely composed shot. Moving beyond the individual scenes, we can analyze how in earlier scenes in Little Foxes and Best Years Wyler prepares us to notice these climactic compositions.29 Nonetheless, it was Bazin who made them prototypical cases—tough, if extreme, examples that demand explanation.

My account of Bazin’s contributions to poetics might seem to have favored only one dimension of inquiry, stylistics. The emphasis reflects both my own interests and one of Bazin’s abiding concerns. But poetics has many other dimensions: the study of genre, of thematics, and of overall form, or “composition” in its broadest sense. Bazin made contributions to all of these areas of inquiry, but I want to conclude my survey of his curriculum by examining yet another dimension of poetics. Besides his effort to analyze and explain film style, Bazin proved truly pioneering in examining new strategies of narrative construction that arose in the postwar period.

Bazin was captivated by the Neorealist films that were pouring into Paris at about the same time as American imports were. “It could well be that, today, Italy is the country where the understanding of film is at its highest,” he wrote in 1948.30 Most critics celebrated the fact that Neorealist films turned a new light on the nation’s social problems and did so by using nonactors and shooting on location. Bazin went further. He greeted these technical innovations part of that evolution toward realism that he saw as the most progressive trend in contemporary cinema. And another hallmark of the movement was its attitude toward cinematic narrative, an attitude that was in harmony with the contemporary American and European novel.

Bazin finds, for instance, that the new Italian films display remarkably fragmented plots. Paisá is a film of episodes, treating its stories as more or less incomplete; even within each episode, the action is presented obliquely, and many dramatic issues remain unresolved. “The technique of Rossellini undoubtedly maintains an intelligible succession of events, but these do not mesh like a chain with the sprockets of a wheel. The mind has to leap from one event to the other as one leaps from stone to stone in crossing a river.”31 Hollywood’s dramatic arcs and tight linkage of cause and effect yield to a sheer succession of events, with scenes dwelling on conventionally undramatic moments. “The story unfolds without regard for the rules of suspense, its only resource being a concern with things themselves, as in life.”32 De Sica’s Ladri di bicyclette “unfolds on the level of pure accident: the rain, the seminarians, the Catholic Quakers, the restaurant—all these are seemingly interchangeable, no one seems to have arranged them in order on a dramatic spectrum.”33 In such formulations, Bazin showed himself as sensitive to narrative constuction as he was to visual style.

Bazin’s discussions of Neorealism laid bare several principles that would shape the narratives of postwar cinematic modernism. The temps morts of so many 1950s and 1960s dramas were anticipated in La Terra trema, in which Visconti builds his scenes out of “blocks of reality.” “A fisherman rolls a cigarette? No ellipsis compresses the operation; we see the whole thing. It will not be reduced to its dramatic or symbolic meaning.”34 The meandering, unpredictable string of encounters in Rossellini’s Voyage to Italy is a logical step in the Neorealist approach to plotting.35 Bazin did not live to see Hiroshima mon amour, with a second half centering on a couple drifting through the city in a nocturnal pas de deux, or L’Avventura and La Notte, with their languid, apparently aimless strolls.36 But he would surely have recognized these wayward excursions as a logical extension of what he found in Ladri di bicyclette: “It would be no exaggeration to say that Ladri di bicyclette is the story of a walk through the streets of Rome by a father and his son.”37 Dwight Macdonald sloganized the insight: “The Talkies have become the Walkies.”

Bazin invested the narrative experiments of Neorealism with an ontological and moral weight stemming from his commitments to phenomenological realism. Once again, those of us who don’t share those commitments can nevertheless weigh the justice of his observations about large-scale form. Bazin’s observations have proven very helpful in characterizing what we call in English “art cinema.”38 Bazin’s analyses of Neorealism point up principles of construction basic to an entire narrative tradition that was forming under his eyes.

Beyond matters of methodology, beyond the empirical discoveries he bequeathed us, Bazin shows us that any theory which takes into account the craft practices of filmmaking, the formal and stylistic patterning that films display, and the implications of filmmakers’ creative choices can help answer the questions posed by a poetics. All this isn’t to say that we should strip off the “poetics” from Bazin’s theory and toss away the rest. His theoretical arguments are fascinating and remain relevant to the digital age. His criticism is lyrical and eloquent. As a researcher, however, I want to explore some paths he did not follow to the end. Barred from university posts by a chronic stammer, he became an inspiring teacher, and we are still his students.

1 : Dudley Andrew has argued persuasively that Bazin offers not a theory of film as art but rather a perspective on film as a medium (though one that harbors unexpected aesthetic dimensions). That medium remains interesting and valuable when it is used for sheer recording, as in Thor Heyerdahl’s Kon-Tiki. See Andrew, “André Bazin’s ‘Evolution,’” in Defining Cinema, ed. Peter Lehman (Rutgers: Rutgers University Press, 1997), 88.

2 : See my Ozu and the Poetics of Cinema (Princeton: Princeton University Press, 1988), Chapter 6.

3 : See Vladimir Nizhny, Lessons with Eisenstein, trans. and ed. Ivor Montagu and Jay Leyda (New York: Hill and Wang, 1962). My argument for this as an exercise in poetics can be found in my The Cinema of Eisenstein (Cambridge: Harvard University Press, 1994), Chapter 4.

4 : For a related argument placing Bazin in a poetics-centered tradition, see my Poetics of Cinema (New York: Routledge, 2007), 14–16.

5 : Again, Eisenstein got there ahead of him, in his pedagogy and in such essays as “‘Eh!’ On the Purity of Film Language,” in Selected Works, vol. I, Writings, 1922–1934, ed. Richard Taylor (Bloomington: University of Indiana Press, 1988), 285-295; and the essays collected in Selected Works, vol II, Towards a Theory of Montage, ed. Michael Glenny and Richard Taylor (London: British Film Institute, 1994).

6 : Bazin, “William Wyler ou le janseniste de la mise en scène,” Qu’est-ce que le cinéma? vol. 1 (Paris: Cerf, 1958), 166.

7 : See Dudley Andrew, André Bazin (New York: Oxford University Press, 1978), 227.

8 : Bazin, “William Wyler,” 163.

9 : See Andrew, Bazin, 233.

10 : Gregg Toland, “Realism for ‘Citizen Kane,’” American Cinematographer 22, 2 (February 1941), 54, 80.

11 : Gregg Toland, “The Motion Picture Cameraman,” Theatre Arts 25, 9 (September 1941), 652–653.

12 : See Gregg Toland, “L’Opérateur de prises de vues,” La Revue du cinéma no. 4 (1 January 1947), 16–24.

13 : See for example Jacques Manuel’s comments on the fluid continuity and depth of field achieved by The Magnificent Ambersons in “Essai sur le style d’Orson Welles,” La Revue du cinéma no. 3 (December 1946), 56.

14 : William Wyler, “No Magic Wand,” The Screen Writer 2, 9 (February 1947), 10.

15 : Alexandre Astruc, “Notes sur Orson Welles,” Du Stylo à la caméra…et de la caméra au stylo (Paris: 1992), 322. See also Claude-Edmonde Magny, L’Âge du roman américain (Paris: Seuil, 1948), 39–41.

16 : Bazin, “William Wyler ou le janseniste de la mise en scène,” 159.

17 : Renoir remarked in 1938: “The more I have progressed in my craft, the more I have been led to stage scenes in depth. Further, I am inclined to avoid placing two actors cunningly before the camera, as if they were posed for a photograph. More often I set my characters at various distances from the camera and make them move. For that, I need considerable depth of field [profondeur de champ].”  See Jean Renoir, “Souvenirs,” in Jean Renoir, Premier plan no. 22–23–24 (May 1962), 12.

18 : More specifically, Bazin finds that these filmmakers’ styles tap into different aspects of phenomenal reality: the concreteness of action (Rossellini “directs facts [faits]”), the continuity of time and space (Welles and Renoir, through different techniques).

19 : On the History of Film Style, 2d ed., (Madison, WI: Irvington Way Institute Press, 2018), Chapter 3.

20 : The summative essay is “L’Évolution du langage cinématographique,” Qu’est-ce que le cinéma? vol. 1 (Paris: Cerf, 1958), 131–148. It is compiled from “Pour en finir avec la profondeur de champ,” Cahiers du cinéma no. 1 (April 1951), 17–23; “Montage,” in Twenty Years of Cinema in Venice (Rome: Edizioni dell’Ataneo, 1952), 359–377; and “Le Découpage et son evolution,” L’Âge nouveau no. 93 (July 1955), 54–61.

21 : I discuss these developments in On the History of Film Style, Figures Traced in Light: On Cinematic Staging (Berkeley: University of California Press, 2005), and essays in Poetics of Cinema. See also essays and weblog entries on this website.

22 : For an analysis of nationalism in the Histoire, see Mary Jane Green, “Fascists on Film: The Brasillach and Bardèche ‘Histoire du cinema,’” South Central Review 6, 2 (Summer 1989), 32–47.

23 : Maurice Bardèche and Robert Brasillach, Histoire du cinéma, 2d ed. (Paris: Denoel, 1943), 143.

24 : Bazin, “William Wyler,” 45.

25 : See Bazin, “Montage interdit,” Qu’est-ce que le cinéma? vol. 1 (Paris: Cerf, 1958), 119–129.

25 : Andre Bazin, Orson Welles (Paris: Cerf, 1972), 70.

27 : Bazin, “William Wyler,” 154. There is evidence that the use of graduated focus in the Little Foxes scene sprang from a concrete production problem: Herbert Marshall’s artificial leg. See “Problems, problems: Wyler’s workaround.”

28 : See Figures Traced in Light, Chapter 1.

29 : Bordwell, On the History of Film Style, 224–228 and “Problems, problems: Wyler’s workarounds.”

30 : “Le Réalisme cinématographique et l’école italienne de la libération,” Qu’est-ce que le cinéma? vol. IV: Une Esthétique de la Réalite: Le Néo-réalisme (Paris: Cerf, 1962), 10.

31 : Bazin, “Le Réalisme cinématographique,” 31–32.

32 : Bazin, “La Terre tremble,” in Qu’est-ce que le cinéma? vol. IV: Une Esthétique de la Réalite: Le Néo-réalisme, 38.

33 : Bazin, “Voleur de bicyclette,” in Qu’est-ce que le cinéma? vol. IV: Une Esthétique de la Réalite: Le Néo-réalisme, 59.

34 : Bazin, “La Terre tremble,” 40–41.

35 : Bazin, “De Sica et Rossellini,” in Qu’est-ce que le cinéma? vol. IV: Une Esthétique de la Réalite: Le Néo-réalisme, 115.

36 : I discuss this aspect of Hiroshima mon amour in “On the Criterion Channel: Five reasons why Hiroshima mon amour still matters.”

37 : Bazin, “Voleur de bicyclette,” 53.

38 : David Bordwell, “The Art Cinema as a Mode of Film Practice,” in Poetics of Cinema, 151–169. See also my Narration in the Fiction Film (Madison: University of Wisconsin Press, 1985), Chapter 10.