Victor A. Grauer

©Victor A. Grauer, 1982


Montage and the Paradox of Film Realism

Chapter 1
The Double Aspect of Montage

Kuleshov's Puzzle

Let us mentally recreate a version of the "Kuleshov experiment,'' circa 1918.1 Film footage of an actor's face, apparently neutral and expressionless, is found. This footage, let us call it "shot x,'' is intercut with three other pieces of film, shots "A, B, and C,'' so that the following structure results: A, x, B, x, C, x.

Shot A is of a bowl of steaming soup; shot B is of a woman in a coffin; shot C is of a child playing with a toy bear. The edited film is projected to an audience innocent of the real situation. All marvel at the actor's sensitivity. It appears to them that he is reacting appropriately to each previous shot--after the soup he is hungry, after the dead woman sad, after the child amused.

In another case a scene is created in which two people, actually photographed miles apart, "see'' one another, approach, and finally meet at a third place completely remote from where either was initially shown. They clasp hands, look up and "see''--the White House--a shot from a Hollywood film. Through the "language'' of cinema a sense of unified space and time is manufactured from disconnected fragments.2

These are only two of many similar experiments conducted by Lev Vladimirovich Kuleshov, pioneer of the early Russian cinema, teacher of a generation that was to revolutionize the art of film. Kuleshov's experiments followed upon his study of some American films, particularly those of D. W. Griffith. These films were immensly popular in Russia. Very few people, it would seem, had any trouble following them. They were not considered strange, certainly not "avant-garde.'' Yet Kuleshov confesses that he was "stunned'' by the results of his own experiments with montage effects very similar to those found in these same popular American films. To him, montage was strange indeed, something to be understood "futuristically.''3

The clue to Kuleshov's reaction lies in one of the ironies of montage: it is strange only to those who know (or can guess) the real relationships that exist among the various pieces that are strung together. To the uncritical viewer, everything is as natural as can be, everything falls into place as part of a totally convincing spatial and temporal continuum. Much of the opposition that montage initially encountered thus came from specialists who knew enough about how it worked to perceive its strangeness. Montage seemed wierd, cubistic or futuristic, when its effects were analyzed. On the screen it looked completely natural. Only those "in the dark'' were not puzzled.

Eisenstein's Machine

Sergei Eisenstein, later to be a pupil of Kuleshov, began his artistic career in 1920 as an enthusiastic heir of the futurists. Trained as an engineer, he was easily drawn to the futurist ethos with its disdain for idealism and worship of cold technological precision. His earliest serious creative work was in the theater, under the strong influence of Mayakovsky and Meyerhold, both of whom became close friends. Mayakovksy was, of course, the most notorious poet of Russia, signer of the Futurist Manifesto of 1912, and prime mover of Constructivism, an important Russian offshoot of Futurism. Meyerhold was the leading theatrical director of the Constructivist movement.

Eisenstein's concept, "montage of attractions.'' was born in the Constructivist theater. The term "montage,'' meaning "assemblege,'' was adapted for the theater by analogy with the industrial assemblage of machine parts. The guiding concept was typically Futurist-Constructivist: the theater must be broken down into its basic and most potent elements, just as if it were a machine, a machine for producing "attractions'' mathematically calculated to have the strongest effect.

The young Eisenstein wrote (in 1926) enthusiastically of three dramas staged by him in 1922-23, in which "the staging was a mathematical calculation of the elements of affect.'' He writes of trying to "dissect cubistically a classical play into separately affective `attractions'.'' He goes on to link his theatrical ideas with those of his first film, Strike: "such an understanding of theater led in a straight path to cinema; only the most inexorable objectivity could be the sphere of cinema,'' The film, like the earlier dramas, was based on the "rational construction of affective elements... [involving] thorough analysis and calculation. . . No more readiness of wit is needed here than in the design of the most utilitarian building of reinforced concrete.''4

These comments reveal the extent to which the Eisenstein of Strike and Potemkin associated himself with the most radical artistic tendencies of his time. It is clear that he viewed the fragmentations of cinematic montage in direct relation to the "formalist'' fragmentations of cubism, the "mechanist'' spirit of futurism and the objective and precise calculations of constructivism. When we add to the above his comparison of montage "to the series of explosions of an internal combustion engine, driving forward its automobile or tractor,''5 it becomes clear that we are a long way from Griffith and Kuleshov. The "futurism'' that alarmed Kuleshov was obviously a source of exhiliration to Eisenstein. While the former tried to mitigate its effects,6 the latter was eager to emphasize them.

Montage as Dialectic

Eisenstein's earliest thinking on montage relates it unequivocally to the control of affect. Although he had not at that time read Pavlov, he was later to acknowledge that he had been thinking in Pavovian terms.7Eventually, however, Eisenstein couid not remain satisfied with so limited an approach. Montage clearly gave rise not only to feelings but also ideas, and in such a way as to suggest the Hegelian dialectic itself: if "montage is a collision'' and "from the collision of two given factors arises a concept,''8 then "montage is a [concept] that arises from the collision of independent shots.''9
This sheds a good deal of light on the nature of montage as Eisenstein ultimately viewed it: the shot itself is neutral until it "collides'' with another shot, an event that gives rise to an active idea. It is the play of ideas rather than the simple juxtaposition of shots that is the true essence of montage. A picture of a bowl of soup is followed by a picture of a man's face and the concept "hunger'' arises. Through careful selection and relation of shots a series of very specific ideas can be made to arise in this way in the mind of the viewer -- as though the shots were words, or, as eisenstein would prefer ideograms.
One may of course question the extent to which this kind of thing is truly "dialectical.'' Dialectic is a critical activity, while montage depends on suspension of disbelief. In any case, it does seem to have something to do with language and concepts and, to the extent that it does, one may well wonder how it relates to the basic principles of cubism, futurism and constructivism. These movements, so important to the young Eisenstein, were in militant rebellion against the very concept-forming mechanisms that lie at the heart of verbal language and film language both.

Here are the futurists in 1915:

It's stupid to want to explain with logical minuteness everything taking place on the stage, when even in life one never grasps an event entirely, in all its causes and consequences, because reality throbs around us, bombards us with squalls of fragments of interconnected events, mortised and tenoned together, confused, mixed up, chaotic. E.g., its stupid to act out a contest between two persons always in an orderly clear and logical way, since in daily life we nearly always encounter mere flashes of argument made momentary by our modern experience, in a tram a café, a railway station, which remains cinematic in our minds like fragmentary dynamic symphonies of gestures, words, lights, and sounds.10

The idea of the "cinematic'' which arises here, with its "fragmentary dynamic symphonies of gestures, words, lights and sounds,'' reminds us of the look of an Eisenstein film minus the story and the "dialectic.'' It would be interesting to know what films the futurists were seeing in 1915. They were obviously impressed by the surface visuality of these films, the strange "futurism'' of the juxtapositions of fragments of space and time torn out of context, the kind of thing which disturbed Kuleshov, rather than the film "language,'' dialectical or otherwise, toward which these films were groping.

The gulf between Eisenstein's interest in film as a dialectical, concept-producing idiom and his sympathies with the radically different futurist notion of a cinema of riotous, irrational fragmentation, reveals a serious contradiction.11 Though clearly attuned to the futurist view, and always eager to associate montage with the new and exotic, he was equally at pains to show how really familiar it was, to demonstrate that it was a natural development from the past. He related montage not only to Hegel and Marx, but to Pushkin, Flaubert, Leonardo da Vinci, even Dickens. What then of those "mathematical calculations,'' those "cubist dissections'' of which he had written as a young man?

The Scandal of Montage

As should be clear by now, montage has a double aspect. A radical framentation of subject matter into thoroughly disjunct bits and pieces is the basis for the pitiless analysis that lies at the heart of "formalist'' modernism. In a typical cubist canvas the object is dissected into its constituent parts which are then reassembled to form a completely strange new entity. Montage is also a technique of fragmentation and reassembly. Pieces of film containing images can be broken down into certain categories, physically severed according to category and cemented together again to form a radically different "reality.'' So similar is this process to the fragmentations of the modernists, that montage was, almost from the beginning, enthusiastically applauded by the most radical artists of the day. It seems clear from the futurist document quoted above that many of these artists were seeing montage essentially as fragmentation; seeing it, in fact, modernistically.

But montage has another aspect, diametrically opposed to the first. While strips of film are, in purely physical terms, cut into pieces and simply "mortized and tenoned together'' as cemented fragments, these fragments can in the mind of the viewer be brought together again to form a coherent, unified totality. This phenomenon is the basis for any notion of film "language,'' dialectical or otherwise.
The mental synthesis of material fragments into concepts and the shaping of these concepts into narrative are of course utterly alien to the formal modernist tradition which shaped the aesthetic of the young Eisenstein. These rather surprising synthetic powers of montage made it possible for film to become the storytelling art par excellence. In order for this art to be a success, however, the great "scandal'' of montage, its kinship with cubist fragmentation, had to be thoroughly suppressed.

The Context of Implication

If montage concealed a scandal, no one could have been more painfully aware of it than Eisenstein. As Soviet doctrine hardened against "formalism,'' the modernist aspect of montage became a serious embarrassment. By 1938, Eisenstein had apparently done a complete about-face, inveighing thus against the "leftists'' of cinema:

While playing with pieces of film, they discovered a certain property in the toy which kept them astonished for a number of years. This property consisted in the fact that two film pieces of any kind, placed together, inevitably combine into a new concept, a new quality, arising out of that juxtaposition.

This is not in the least a circumstance peculiar to the cinema, but is a phenomenon invariably met with in all cases where we have to deal with juxtaposition of two facts, two phenomena, two objects. We are accustomed to make, almoat automatically, a definite and obvious deductive generalization when any separate objects are placed before us side by side.12

An older, more sober Eisenstein is arguing in behalf of the judicious use of montage in the service of an overall unifying "theme.'' But this attack on the montage "radicals'' goes beyond its mark, for Eisenstein is here exposing his own notion of montage "dialectic'' as a triviality. Is montage, in essence, really a "collision'' exploding into a concept that drives the film forward like the explosions in an internal combustion engine? Apparantly we do not need a carefully plotted collision in order to strike off sparks that can explode into meaning. On the contrary, meaning seems unavoidable regardless of what we do or fail to do.

The embarrassing result of Eisenstein's attempt to explain away montage's scandalous origins as "mere'' juxtaposition, continues to haunt film theory as a return of the repressed. Is meaning created through a sophisticated mastery of "film language,'' or is it simply supplied by the viewer each time one shot is followed by another? Must film be ordered? Or will a chaos of disconnected fragments still somehow hang together in the mind?

The issue arises again and again. For Béla Balázs, "this confidence that we are seeing the work of a creative intention and purpose, not a number of pictures thrown and stuck together by chance, is a psychological precondition of film-watching and we always expect, presuppose and search for meaning in every film we see.''13 Jean Mitry writes of a "logic of implication'' which turns any set of images into language.14 Christian Metz, quoting Balazs and Mitry (but not Eisenstein), makes the issue the climax of an extended attack on "Russian'' montage:

if montage was indeed sovereign, it was so by necessity, for, when two images were juxtaposed purely by chance, the viewer would discover a 'connection.' That, and nothing else, is what Kuleshov's experiments demonstrated.15

We have clearly arrived at something close to the heart of the montage phenomenon. To more fully understand its implications, let us return to Kuleshov's original experiment. For practical purposes, we will translate from images to words, as though dealing with a poem:

A bowl of steaming soup sits on a table.
Alexy stares with his eyes wide open.
A woman is placed in a coffin.
Alexy stares with his eyes wide open.
A child is playing with a toy bear.
Alexy stares with his eyes wide open.

Taking each pair as an isolated unit, it is easy to see what happens. In each case our mental picture of Alexy's stare takes on overtones of meaning from our understanding of the previous "shot.'' We can "see'' him expressing hunger, horror and amusement, yet the three expressions are derived from a reading of the same "shot,'' repeated each time in a different context.

Fortuitous as the Kuleshov juxtapositions are, they nevertheless call forth a conventional relationship: the "reaction shot.'' This involves a careful choice of shot-types so that one and only one general inference can be drawn: the second shot is a reaction to the first, the fourth to the third, the sixth to the fifth. The integrative property of the mind alluded to by Eisenstein, Balazs, Mitry and Metz, undoubtedly facilitates the "reading.'' But the meaning that arises is clear because it is strictly conventional. Metz is therefore wrong to associate the Kuleshov experiment with pure chance.
Moving closer to the sort of thing invoked by Metz, let us make a more radical "composition'' out of Kuleshov's basic material:

A bowl of steaming soup sits on a table.
A woman is placed in a coffin.
A child is playing with a toy bear.

Here we have, from a more or less accidental juxtaposition, a highly suggestive, rather sophisticated little poem, resembling a haiku. At once certain questions, not unlike those that might crop up in a freshman lit. course, come to mind: why is the soup on the table? what does the soup have to do with the dead woman? what is the significance of the fact that the soup is steaming? why is the child playing when its mother is in the coffin? is the child aware of the significance of its mother's death? Perhaps the child is too heartsick to pay attention to the soup, and too overwhelmed to face the reality of the dead mother. It sits in its corner playing with the toy bear which is, after all, like a dead body--pathetically the child is trying to bring the bear to life, a reflection of its feelings toward the mother.

Omitting the "reactions,'' we have simply juxtaposed three shots in a manner that seems completely free from any strictly conventional code. In theory one might assume the result would be meaningless; each shot would be seen for what it is and it would be assumed that a pointless juxtaposition was all that was intended. In fact this does not happen. Since an intended meaning is not present in a strictly coded form, we have pure implication, free floating, unable to attach itself securely to any one thing.
Since the existence of some meaning or other is implied by the juxtaposition, the resulting ambiguity calls for resolution. The mind begins to cast about for possible meanings.

In view of the confusing nature of this kind of juxtaposition, one wonders why it seems so effective. The sequence we have derived from Kuleshov resembles, in fact, the kind of thing one might find, not in the extremists of Russian montage, but in the more sober films of Bergman, Fellini or Antonioni.
Ambiguity is, in fact, a highly effective device. For one thing it flatters viewers into believing that they are contributing to the meaning of the film by "putting things together'' for themselves.16 More important, it suggests profound and mysterious meanings that exist beyond what we can see or even conceive. As ambiguity intensifies this sense of transcendence increases.

To demonstrate, let us push our "haiku'' to the point where no possible conventional meaning could adhere:

A bowl of steaming soup sits on a skull.
An egg is placed in a coffin.
A child is eating a toy bear.

Here we have something very close to pure ambiguity. Note that each "shot'' taken separately presents a perfectly clear image, despite its apparent absurdity. The three together create an "atmosphere'' of suggestion concerning a possible transcendent and mysterious universe of feeling or discourse which could give them meaning. This is a basic tactic of surrealism.

The above analysis enables us to distinguish some modes of the phenomenon first noted by Eisenstein. Any group of images no matter how unrelated or absurd will, apparently, find themselves linked in the mind of the viewer. If the images are carefully chosen, according to strictly conventional codes (as they are in the great majority of films), then a clear meaning arises. If the juxtaposition is not strictly conventional but suggestive of conventional interpretations, then there will be a certain amount of pleasant ambiguity, suggesting "hidden'' meaning. A totally unconventional or absurd juxtaposition will call forth a mysterious sense of pure ambiguity with strong "aesthetic'' overtones.

Basic to all of the above is an integrative property of the mind which tends to place any given image in a context where some sort of meaning is implied. This "context of implication'' is, to Eisenstein, evidence for the triviality of radical montage. To Metz, it calls the whole notion of film "language'' into question. For us, the context of implication must be regarded as the very core of montage, the crux where its two mutually opposed aspects encounter one another.

Chapter 2

The Realist Reaction

Montage Abstraction vs. Camera Reality

The period of Eisenstein's greatest achievements coincides with an era of interest in montage that is unparalleled before or since. Eisenstein is only the most famous of a long list of Russians. Even outside the Soviet Union, there was a great deal of intense experimentation with what must have seemed, at the time, to have been every possible aspect of montage.

This era, in which montage was regarded as the very essence of what cinema was all about, was over by the early Thirties. Editing became less extreme, more judicious and unobtrusive. There was a greater emphasis on mise-en-scène (dramatic and pictorial development within a single shot.)1 This new approach was linked with the advent of the "talking'' film. Together the softened montage, mise-en-scène and use of sound-on-film led to a greater feeling of realism.

During the late Thirties and Forties there was a much stronger pull away from montage. In films like La Grande Illusion and La Règle du jeu, Jean Renoir demonstrated the real powers of pure mise-en-scène. Using camera set ups which permitted great depth of field, intricate pans and tracking shots, he was able to present long and complex scenes with great sophistication in a single extended shot. Orson Welles used similar devices more self consciously in Citizen Kane. The Italian neo-realists went even farther in their emphasis on unedited camera "reality.''

Behind these developments, aside from a natural excitement over the powers of the camera with its improved lenses, was a deep distrust of the "language'' that had so elaborately grown up around montage. The suspicions were most eloquently voiced by André Bazin, leading critic of film "formalism'' and prime spokesman for what has become known as "realist'' theory. Bazin took montage to task for producing "an abstract result, none of the concrete elements of which are to be found in the premises: maidens plus appletrees in bloom equal hope...The meaning is not in the image, it is in the shadow of the image projected by montage onto the field of consciousness of the spectator.''

To Bazin, the true essence of film is mise-en-scène, realized via the unedited long "take,'' the unmanipulated moving photograph:

It is not for me to separate off, in the complex fabric of the objective world, here a reflection on a damp sidewalk, there the gesture of a child. Only the impassive lens, stripping its object of all those ways of seeing it, those piled-up preconceptions, that spiritual dust and grime with which my eyes have covered it, is able to present it in all its virginal purity to my attention and consequently to my love.2

A Materialist "Aesthetic''

While the Russians had argued that montage was the essence of cinema, the "realists'' could point to the unique power of the moving image, the "fingerprint'' of the real world, as the truly characteristic strength of motion pictures. Thus the true "cinematic approach'' is firmly based on a respect for unmanipulated photographic reality. From this point of view it is montage, with its roots in language and story (hadn't Eisenstein stressed this?), that pushes film in the direction of literature, just as earlier films had been dependent on theater. Only the feeling of reality brought to us by the camera is truly and uniquely cinematic.

This notion is the basis for the film theory of Siegfried Kracauer, for whom "the basic properties [of film] are identical with the properties of photography.'' Although more tolerant of montage than Bazin, Kracauer is equally suspicious of its tendency to place an abstract "veil'' over the raw look of things. This he attributes to two fundamental aspects of our culture: science and technology, which encourage the kind of abstract thinking which causes things in themselves, with all their unique properties, to recede into the background; the "formative'' or purely artistic impulse, which pulls film away from its unique photographic orientation toward values borrowed from the more traditional arts.3

To Kracauer, film is "materialistically minded,'' inexorably drawn to the homely details of life itself, the uniqueness of which it is film's special destiny to capture. To clarify his position, he quotes the following key passage from Erwin Panofsky's "Style and Medium in the Motion Pictures:''

The processes of all the earlier representational arts an idealistic conception of the world. These arts operate from top to bottom, so to speak, and not from bottom to top; they start with an idea to be projected into shapeless matter and not with the objects that constitute the physical world... It is the movies, and only the movies,that do justice to that materialistic interpretation of the universe which, whether we like it or not, pervades contemporary civilization.4

Panofsky manages to say a great deal in a few words and we shall have occasion to return to them. For now it is sufficient to note that Russian style montage, with its "idealistic'' top to bottom approach, is not what Panofsky means by "the movies'' in this particular context. For him, as for Kracauer, film finds not only its unique properties but also its remarkable timeliness in its abilities to present us with pictures of the ordinary world existing prior to any conception of it that may be imposed by an artist.
Indeed, for the realists, the term "art'' implies something antithetical to the "cinematic approach.'' This is probably in part a response to formalist theorists who have argued that "art'' in cinema, as in any art-form, stems from its most art-ificial, anti-realistic elements. In this light Bazin's classic summing up of the realist position sounds paradoxical: "the aesthetic qualities of photography are to be sought in its power to lay bare the realities.''5

Realism and the Analogue Image

At the heart of the realist position is the notion of a characteristically cinematic image of reality seen as reality, gaining whatever meaning it has from its own natural properties rather than some culturally encoded system (such as language, filmic or otherwise). This image, captured through the special properties of photography, is considered a direct and continuous visual analogue of the object it "presents.''

From the realist point of view the image of a horse in a film is an analogue of the real horse, This image should derive its meaning from our confrontation with that particular horse rather than some symbolic meaning attached to the horse (as in Eisenstein). The latter is an attempt to graft a linguistic notion onto an art whose real power lies elsewhere. When we see the horse as a symbol, moreover, its value as a direct analogue of something unique vanishes; we no longer see it as a particular animal; we cease, in fact, to see at all in the material sense. We "see'' ideas.

Our earlier analysis of the Kuleshov experiment enables us to go a bit more deeply into the problem. It is not simply the systematic encodings of film language, but the more fundamental and ubiquitous operation of what we have called the context of implication, that opposes itself to the analogue image. Even in the absence of specific codes, a given shot will still be seen in terms of an ambiguous cloud of implications called forth by the montage context. Whatever is to promote the analogue image must oppose this very powerful psychological tendency.

Montage as Modernist Symptom

Ironically one of the strongest attacks on montage as threat to the analogue image was mounted by Christian Metz in the very act of founding a "science'' of film language that was ultimately to base itself on a theory of montage syntax (see the following chapter). Metz joins Bazin in his attack on Russian montage in an argument that takes the form of an assault on the analytic spirit of a vaguely defined "modernism.''

To Metz,"a comparison suggests itself [between the Russians'] obsession with breakdown analysis and montage and certain tendencies of the `modern' spirit and civilization.'' He goes on to compare Russian montage to things like erector sets and toy electric trains, which he calls "syntagmatic toys.'' Playing with such toys stimulates the "spirit of manipulation'' which eventually leads to the triumph of engineering, cybernetics, information theory, etc., practices toward which Metz is outspokenly critical.
Echoing Kracauer's suspicion of technology, Metz states that "the machine has ground up human language and dispenses it in clean slices to which no flesh clings.'' He speaks ironically of a situation in which "the natural object (whether human language or cow's milk) is considered as a simple point of departure.'' It is broken down and analyzed "as in the cinema,'' then put together again as a "pure product of the mind.'' Eisenstein's "syntagmatic mentality'' is contrasted with the attitude of the neo-realist director Roberto Rossellini, who is quoted as saying "Things are. Why manipulate them?''6
The denunciation of montage as symptom of the calculated fragmentations of "modernism'' is a leitmotif of realist theory. Metz' critique of montage is telling, but the attack on modernism is wide of the mark.

Montage is indeed manipulative, not only of the film material itself and images inscribed on it, but of the minds of those to whom it presents itself as an innocent narrative device. It is the last named form of manipulation that is truly problematic, causing viewers to see what in fact is not there (the "reactions'' of Kuleshov's actor), to misinterpret what they do see (the false perception of unified space and time in Kuleshov's scene culminating in the shot of the White House) and generally to "perceive'' abstractions rather than pictures of real things and events, while being encouraged to believe that in fact only pictures of real things and events are being seen. Nothing of the kind occurs with erector sets, electric trains, cybernetics, or information theory, nor in cubist, futurist or constructivist art. It is of the essence of the "modernist'' devices that Metz attacks that they take delight in displaying themselves as the contrivances they are.

Metz fails to deal with the fact that it is only traditional, narrative montage which chooses to wrap itself in a cloak of invisibility. In this respect, the more extreme forms of montage, forcefully calling attention to themselves, are the least objectionable. Yet it is these which draw his most intense fire.

Metz goes wrong, as does Bazin, in his failure to grasp the double aspect of montage. Its dangerously manipulative powers do not directly stem from "modernist'' analytic fragmentation. These powers have their origin in the completely misleading and thoroughly anti-modernist "syntactic'' synthesis of fragments into unified concepts at the service of conventional narrative. What Metz is attacking is language itself, especially that highly formalistic, "structural'' approach to linguistics which is to become the basis for his own subsequent research.

Some Contradictions

The realist position may be divided into two parts; a critique of film language, centering on montage abstraction; promotion of the unedited motion picture photograph as a direct imprint of nature. The first, while somewhat misguided, nevertheless reveals a serious problem; montage seems, by necessity, to promote a purely conceptual mode of seeing, veiling film's unique power to "lay bare the realities.''

The second is far less convincing, giving rise to some embarrassing contradictions. Bazin, for instance, often sounds like an advocate for some extreme form of pure documentation. Yet it is film makers like Renoir, Welles, Rossellini and Di Sica, storytellers, not documentarists, who really interest him. The works of Renoir and Welles are certainly as contrived and manipulative as anything of Eisenstein, while over the years neo-realism has come to resemble verismo opera more than anything else.

As Metz himself has shown, meaning laden juxtapositions of the sort that Bazin deplores in montage can be created within the development of a single shot. Bazin has, of course, argued that the meanings derived from mise-en-scène are found in the shot by us, while those of montage are forced on us by the editor. This is naive. Every director knows how to control mise-en-scène so that the eye of the viewer knows exactly where it is supposed to go, regardless of depth of field or width of angle. Mise-en-scène is hardly less manipulative than Russian montage, only more subtle, more ambiguous.

The issue of "ambiguity'' leads to a central contradiction in realist thought. For Bazin, ambiguity, as something inherent in reality itself, is very much to be desired, the very thing which Russian montage denies. He praises neo-realism for its tendency "to give back to the cinema a sense of the ambiguity of reality.''7 In this connection he mentions Rossellini's concern to preserve the inherent "mystery'' of a girl's face.

In what way is mystery inherent in a girl's face? In what way is reality possessed of an inherent ambiguity? Isn't this the "reality'' of the deceived lover, trying to "read'' the beloved's face; or the paranoid stealing sidelong glances at every passerby who might be in on the "plot'' to get him? We can study a face for its beauty or uniqueness; its expression will be ambiguous only to the extent that we want to understand what that expression implies.

Ambiguity, like implication, involves interpretation and can therefore in no way be considered inherent in reality. It arises from our expectations which, in turn, are determined by context, just as in Russian montage. As we have shown, when a clear unequivocal meaning does not arise in such montage, ambiguity arises through the context of implication. If ambiguity is generated by the most radical forms of Russian montage, what use can it be to the realist position? By invoking the presence of ambiguity in the films he admires, Bazin reveals that they too are vulnerable to the context of implication.

Cinéma Vérité

A discussion of realism would not be complete without considering cinéma vérité, a radical attempt to implement realist theory by means of a practice which overcomes some of the contradictions we have mentioned. Cinéma verité in its purest form eschews both narrative and preconceived mise-en-scène. With the aid of lightweight cameras and portable recording equipment, the film makers can better let themselves be guided by what is happening around them. Preconceived ideas of how a film should look will not, theoretically, be a factor. When the film is assembled, long takes are preserved intact, montage effects are minimized, editing is largely a practical matter. Voice-over narration, once the sine qua non of the documentary, is avoided due to its tendency to control the viewing process. Every effort is made to free the viewer to draw his or her own conclusions based exclusively on what is seen and heard in the original setting.

Despite its very real efforts to eliminate viewer manipulation, the cinéma vérité approach encounters serious difficulties. For one thing, cinéma vérité has a taste for the sensational, the violent, the politically controversial, the exotic and picturesque, that belies the fundamental realist concern for "everyday life.'' Few film makers of this school would seem interested in making a film that did little move that "lay bare the realitites.''

Verité film makers often seem to suppose that the images in their films can be taken at face value for what they are in themselves without any coloration from that maker's point of view or any residue from the viewer's experience with other, patently manipulative, films. But the experience of film making and viewing is so saturated with unconsciously perceived conventions that it is difficult to imagine a film without any overtones of film language.

For example, if a shot happens to be taken from a high point of view, it will tend to make the subject seem small, insignificant. A low point of view will imply admiration. A scene shot in bright sunlight will convey a degree of optimism. Night time shooting will call forth something sinister.8

Even if such conventional viewing modes could be neutralized an even more basic problem would arise. For, as should be apparent, the context of implication can serve to link objects and events appearing in a single shot as readily as those juxtaposed by montage. As in random montage, fortuitous relationships within a shot can call forth ambiguous overtones of meaning not anticipated when the shot was taken.

The principle strategy of cinéma vérité is the avoidance of anything that might invoke film language or promote conventional modes of perception. This fundamentally passive approach might possibly result in the total elimination of any form of "coded'' seeing, conscious or unconscious. It is nevertheless incapable of opposing the context of implication which operates whether codes are invoked or not, which causes every image to be seen in terms of what it might imply rather than what it might be in itself, as a unique entity.

Warhol and the Subversion of Realism

Most ciné'ma vérité film makers do not trouble themselves with theoretical issues of the sort we have raised. Their goal is to make vivid, striking films by drawing upon images from spontaneous, unprogrammed events. A great many such films do succeed in being vivid and striking, to the point that some can rival narrative films for sheer entertainment value. And many products of ciné'ma vérité can take well deserved pride as powerful statements on serious social issues.

Despite the very real successes of cinéma vérité, we must not forget that adjectives such as "vivid'' and "striking'' are usually associated with the value system of the aesthete, not the realist. And effectiveness in dealing with social issues is more easily associated with propaganda than with realism per se, where, theoretically at least, objectivity in the form of emotional distance from the subject ought to be highly valued.

When someone actually decided to make a film which conformed in every way to the theoretical ideal of realism, it was not a documentarist, but a notorious and controversial artist: Andy Warhol. Films like his Eat, Sleep and Empire are pure passive contemplation by the camera itself, the image passed directly from object to viewer with little if any intervention on the part of the film maker.

Eat, for example, is simply 45 minutes of a man eating, recorded by a camera fixed on a tripod and never moved. Bazin had referred with enthusiasm to the "dream'' of neo-realist screenwriter Zavattini, "to make a ninety-minute film of the life of a man to whom nothing ever happens.''9 Warhol, not Zavattini, accomplished this and more in Sleep, a six hour film of a man sleeping undisturbed.
Warhol's films, made without a trace of either high seriousness or irony, sent shock waves through the world of cinema which are still being felt. In his deadpan attainment of the realist goal, the realist myth was exploded. Watching a typical Warhol film of this period, we do not find ourselves in intimate contact with the reality of the object photographed. On the contrary, the object becomes ever more strange, inscrutible, distant; as we contemplate it, it seems to disappear before our eyes, leaving us to confront a screen that might as well be blank.

In his remarkable Warhol biography, Stephen Koch has written: "[Warhol's] is a style that renders the presence of the real absent ...[His] interest in recording reality, in fact, a means of subverting reality, replacing the immediacy of contact with the immediacy of his paradoxical absences." According to Koch, the normal vivacity of the eye as it darts from one detail to another is lost in Warhol's films. "Visuality touched by an autistic, unresonating stillness...The sighting, darting vivacity of the gaze becomes a stare. It is a stare of distance, indifference, of mechanically complete attention and absolute contactlessness.''10 In his films Warhol literally makes himself a camera, a passive machine, consciously denying his role as "creative artist'' in a gesture that would seem perfectly attuned to the philosophy of Bazin.

The Realist Paradox

Bazin had complained that the work of the artist was "always in fee to an inescapable subjectivity.'' This he contrasted to "the essentially objective character of photography.'' Bazin had praised the camera as a "nonliving agent,'' which can form "an image of the world ... automatically, without the creative intervention of man.'' To Bazin, only photography, of all the arts, "derives an advantage from [the] absence [of man].''11

We may compare the above with Koch's observation that the "staring dead eye'' of the movie camera "endows Warhol's alienated vision with a mechanistic impersonality.'' Brushing aside those aspects of film technique that, historically, have developed in order to do justice to the vivacity of normal seeing, Warhol "installs the unbroken stare as his formal trope and reinstates the camera in its condition of being a dead machine.''12

Warhol's "dead machine'' is Bazin's "nonliving agent.'' What Warhol has accomplished is the special machinelike objectivity-passivity which Bazin saw as the essence of film. The paradoxes at the core of Warhol's films are at the core of realism itself.

Chapter 3

Realism, Semiotics, Ideology

The realist argument is grounded in a traditional dichotomy: perception, direct, intuitive, bodily experience of nature, mediated by the senses; vs. conception, a purely artificial mental process which, among other things, places all perceptions into categories that form the basis for thought and language. The best known and most influential statement of this dichotomy appears in Kant's Critique of Pure Reason.

But seventy years prior to Kant, the English philosopher George Berkeley had proposed a radically different theory: everything perceptible was really a sign which took its place within a "universal language of nature'' created by God.1 From this point of view the traditional dichotomy was redundant. Perception was only a hidden form of language, thus fully determinate within the conceptual realm.

Berkeley's fundamental idea, completely secularized, forms the basis of the modern theoretical discipline known as "semiotics.'' This discipline, rooted in the "pragmatic'' philosophy of the American, C. S. Peirce and the "structural linguistics'' of the Swiss, Ferdinand de Saussure, concerns itself with the laws governing signs. Since, from the point of view of most semioticians, the sign function must be understood as prior to practically anything that can be concevied, from biology to cybernetics, from perception to language, logic and thought itself, the scope of semiotics is broad indeed.

Early Metz -- Semiotics Plus Realism

Acutely sensitive to the fundamental role of the sign, both Peirce and Saussure understood semiotics as a kind of master discipline under which all other sign-systems, including linguistics itself, must be subsumed. In 1964, Roland Barthes, a leading French literary critic, published an extremely influential work which to some extent altered this hierarchy. In his Elements of Semiology, Barthes argued that only linguistics, especially that "structural linguistics'' founded by Saussure, was sufficiently advanced conceptually to support a truly rigorous semiotics.2 Thus, although Barthes in principle gave priority to semiotics, in fact he attempted to ground it within the elements of Saussurian linguistics.

Inspired by Barthes, Christian Metz, in a series of essays dating from 1964-68 (published in English in 1974, as Film Language), attempted to found a semiotics of the cinema which would also have its basis in fundamental notions of linguistics as set forth by Saussure (and elaborated by Barthes).3 From the start, however, Metz' dependence on linguistics created insurmountable difficulties.

Much of the problem stems from the fact that the linguistic signifier, the actual sound of a word, or the way it looks in print, has a completely arbitrary relation to its signified, what it refers to or means. Basic to linguistic analysis is the "commutation test'' which correlates difference on the level of the signifier with difference on the level of the signified. If a difference of sound calls forth a distinction of meaning, the class of sounds articulated by that difference can be regarded as a paradigm, a fundamental unit, or sub-unit of meaning. Segmentation of the syntagmatic (syntactic) flow of speech into paradigms in this way, is fundamental to linguistic analysis.

Metz, attempting to deal with the photographic image in linguistic terms, was stymied by the fact that this image is not at all arbitrary but is, in fact, a kind of replica of what it represents:

A sequence of [unedited] film, like a spectacle from life, carries its meaning within itself. The signifier is not easily distinguished from the significate [signified] ... The cinema is the "phenomenological'' art par excellence, the signifier is coextensive with the whole of the significate, the spectacle is its own signification, thus short-circuiting the sign itself.4

Very much under the spell of Bazin, Metz is saying that the analogue image is simultaneously the signifier (what we see) and the signified (what is represented by what we see). If these two are equivalent, the commutation test can have no meaning, paradigms cannot be isolated and the process of analysis must come to a halt.

Despite Metz' dependence on the Saussurian-Barthesian approach, he is unable, at this early stage, to cast aside his strong commitment to Bazinian realism in the form of the unedited long take. As we have already learned, Metz viewed the Russian approach to montage as a threat to the integrity of the analogue image which, as the essence of cinema, must be preserved. Yet it is this which "short-circuits'' the sign, the only basis for a true semiotics.

Chapter 3 of Metz' Film Language is essentially a series of unsatisfactory attempts to reconcile the realist point of view with the needs of a language-based semiotics. Again and again Metz seems on the verge of definitively concluding that such a reconciliation is impossible. By the end of the chapter he has indeed made a convincing case for giving up on the whole idea.
Brian Henderson, in a remarkably thorough and doggedly logical critique of Metz' project, expresses his annoyance with what follows:

From the hints dropped in the last few pages of Chapter 3, the reader expects that another method will be developed, derived from other sources and proceeding in other directions ... In fact Chapter 4 does not mention Chapter 3, neither does it address the overall finding of chapter 3 that cinema is analogic and therefore that Barthesian semiotics does not apply.5

Metz does make some attempt to deal with issues raised in chapter 3, but hardly by coming to grips with the fundamental problem. We need not follow his tortuous arguments regarding denotation vs. connotation, language vs. rhetoric, etc. in detail, not because they are irrelevent, but because his discussions are inconclusive.

Ultimately he decides that film is somewhat different from language and that its basic units are not really analogous to those of language. He concludes, surprisingly enough that segmentation, linguistically determined by the commutation test, is cinematically determined "phenomenologically'' through direct viewer experience of the analogue image, the very thing which was supposed to have defeated all hope of segmentation. All is apparently held together by the flow of narrative itself, as directly experienced "phenomenologically'' by the viewer.

Finally, in chapter 5, Metz presents his Grande Syntagmatique, "General Table of the Large Syntagmatic Category of the Image-Track.'' This table is a taxonomic breakdown of various forms of montage. Involving six levels, each the result of a different basic dichotomy, the table articulates seven distinct types of montage "syntagma'': the ordinary sequence, the episodic sequence, the scene, the alternating "syntagma,'' the descriptive, bracket and parallel "syntagmas.'' On the most fundamental level, all are in opposition to an eighth type: absence of montage in the form of the single unedited long take, the "autonomous shot.''6

So little does all this follow from the principal argument, that it is impossible to determine why this particular set of "syntagmas'' was chosen or even in what way the term "syntagma'' itself is justified as a label for what is usually called simply a "montage sequence.''

According to Henderson, Metz, "like Eisenstein and Bazin ...takes from ordinary experience or from previous discourse a basic unit-the shot-and defines several modes of its combination into the next larger unit, the sequence...''7 To Henderson, this is not semiotics at all, but simply yet another variant of classical film theory. Thus does Metz arrive at a point embarrassingly close to that which he had initially attacked. The Grande Syntagmatique is, in fact, a systemization of montage not at all out of tune with the principles of Eisenstein. It is, however, far less rigorous, forceful and convincing.

Eco's Critique-Semiotics vs. Realism

Metz' "phenomenological'' approach to film theory soon came under attack, most notably by the semiotician Umberto Eco.8 Closer to Peirce than Saussure, Eco realized that the analogue image was a kind of icon, the term used by Peirce for a sign that resembles what it stands for. As such, it could never be understood within the context of a system based on language, which, in Peircean terms, uses symbols, signs whose relation to what they stand for is arbitrary. Nevertheless, following Peirce, Eco insisted that iconic signs, while not linguistic, are no more "given'' phenomenologically than are symbols.

For example, on a certain level, Eco breaks the icon down into three components figures -- meaningless perceptual elements akin to phonemes; signs -- meaningful groupings or juxtapositions of figures; semes -- the iconic image itself, put together from its signs. Within this tripartite division a kind of commutation test can apparently be applied. More basic, this analysis of the icon enables us to conceive how an apparently analogue image, like a photograph, could be regarded as an artificial construct -- a code. Patterns of image grain, figures, group to form a minimally identifiable sign, say the image of an ear. This sign is combined with others around it to form a seme -- a human head. Each of the two levels of integration involves a "decoding'' process.

Several different levels of iconic code have been isolated by Eco. For example, codes of transmission control image grain on a news photo or dots on a TV screen. On a higher level, certain types of binary opposition, such as dark vs. light areas, sharp vs. soft focus, high vs. low contrast, etc. can also encode the photographic image, drastically limiting its range of meaning in a purely conventional, completely artificial manner.9 Given photographic codes such as these, Eco disputes the existence of a phenomenologically given analogue image of cinema.

Ideology and the Apparatus

It could of course be argued that Bazin's original position took these sort of codes into account. The heart of Bazin's argument was the photographic apparatus itself, whose "codes'' were largely the function of a scientific neutrality embodied in a "nonliving agent'' that transcended human conventions. But how really scientific is this "nonliving agent''? Or better, can the technological apparatus of science itself escape the semiotic net?

The sacrosanct position of modern science as the ultimate in objectivity has been challenged by a notion which has come to take an increasingly important role in recent thought: ideology. One of the clearest definitions of this elusive term has been oferred by Victor Burgin in a recent essay on the semiotics of photography:

[Ideology is a] complex of propositions about the natural and social world which would be generally accepted in a given society as describing the actual, indeed necessary, nature of the world and its events. An ideology is [thus] the sum of taken for granted realities of everyday life... What is essential about it is that it is contingent [not logically necessary] and that within it the fact of its continency is suppressed.10

Ideology is thus a fundamental and inescapable aspect of culture itself which predetermines the very consciousness underlying thought and all its productions, scientific and otherwise. In an essay entitled "Ideological Effects of the Basic Cinematic Apparatus,'' Jean-Louis Baudry relates this notion to the most basic issue of photography:

Does the technical nature of optical instruments, directly attached to scientific practice, serve to conceal not only their use in ideological products but also the ideological effects which they may provoke themselves? Their scientific base assures them a sort of neutrality and avoids their being questioned.11

Baudry continues, reminding us that the motion picture camera, the projector and all other elements of motion picture technology are socially produced artefacts subject to the demands of ideology. Far from being the "neutral'' instrument of Bazin, they have been very much influenced by "the creative intervention of man.'' The very tendency of this apparatus to recede into the background and present itself as a neutral recording "medium'' links it strongly with ideology, which seeks always to suppress any awareness of its real nature and effect. The details of Baudry's penetrating analysis of the manner in which the analogue image is encoded by the apparatus will concern us in a later chapter.

Ideology and Reality

Our discussion of the workings of ideology vis a vis the cinematic apparatus opens the door to a much more fundamental issue: the influence of ideology on our notion of reality and our ability to perceive it "objectively.'' In this connection we must recall the fact that semiotics begins with a rejection of the dichotomy ("natural'' perception vs. "artificial'' thought) which is the basis for our traditional view of reality. We must remember too that Berkeley's alternative (everything sensible is a sign) involved not only that which mediates between us and reality but the very nature of reality itself, as "directly'' perceived. This was the basis for Peirce's view of the ultimate meaning of semiotics, a view that seems not to have been shared by Saussure or Barthes.

Eco, therefore, following Peirce, moves beyond the aforementioned analysis of the icon to a radical position amounting to an exact reversal of the notion of the analogue image: things like photographs, filmic and television images are in some sense analogues of the real world, not because they "short-circuit codes, but because reality itself is and always must be perceived within a completely conventional, ideologically determined, coding process. Life, therefore, imitates art, not the other way 'round."12

To his credit, Metz was influenced by the analysis of Eco (and others) to alter his views somewhat. Having begun with a sharp critique of Russian montage and an affirmation of realism, he ultimately reverses his field to attack the analogue image itself, the real barrier to the development of a meaningful film semiotics. In Language and Cinema, published seven years after the earliest of the essays, Metz clearly expresses the new viewpoint. Rejecting the "old idea'' of film as a "language without a code,'' he characterises reality itself as "nothing more than a set of codes, that set of codes without which this reality would not be accessible or intelligible, such that nothing could be said about it, not even that it is reality.''13

To Bazin and the early Metz, we first see the cinematic image and then directly read its intrinsic meaning. To Eco and the later Metz, the image must take on an ideologically determined meaning in order that we see it in the first place. For the realists, the meaning of a photographed image is natural, to be found in the thing itself. From the "ideological'' point of view, everything, from the design of the apparatus to the organization of our perception, is determined by convention. Where we direct our attention, what we find to see, what we somehow fail to see, how we see, all these are apparently predetermined by ideology in the form of disguised cultural codes.

Persistance of a Problem

It is not difficult to see the importance that the notion of ideology gives to semiotics as a critical method. If the entire realm of thought and the whole of perception are in fact subliminally encoded through the secret workings of ideology, then our ability to think and perceive in any meaningful sense at all will depend entirely on our mastery of the codes, as elucidated, of course, by the discipline created for that very purpose: semiotics.

Is semiotics adequate for the task? This question has two aspects, one "metaphysical,'' the other pragmatic. The first concerns the idealist nature of ideology itself, an unwieldy monism founded on a vicious circle of codes signifying only other codes, justified only by other codes. As the ultimate ground of every conceivable and perceivable thing, ideology by definition devours all that stands before it. How can semiotics hope to exempt itself?

We must also inquire as to the nature of what it is that has been surrendered as Kantian dualism gives way to the monistic view justifying the broad claims of semiotics. In this connection a fundamental problem arises: If ideology forces us to think exclusively in terms of socially determined meanings then semiotics, by opposing ideology, does, indeed, hold out the promise of helping us to liberate our thought. If ideology forces us to see exclusively in terms of socially determined meanings, then semiotics, by the same token, ought to hold out the promise of helping us to liberate our vision.
But what can this mean? How can semiotics accompish this? Even if we could learn to see "semiotically,'' we would still be seeing in terms of meaning, since semiotics itself exists exclusively within the conceptual world. What effect does this have on the status of seeing or any of the other senses for that matter? Is the realm of the senses to be regarded as completely subordinate to that of thought?

A Compromised Discipline

Moving from fundamental issues to practical concerns, we must inquire as to the viability of semiotic techniques. Unlike linguistics, semiotics has apparently failed to develop analytic devices that are both rigorous and universally applicable. This failing is especially evident in the area which has been the focus of our discussion: the iconic sign.

While Eco's breakdown into figures, signs and semes is helpful, it tends to explain away more than it accounts for. His other coding levels only reveal the extent of the fundamental difficulty. After all is said and done, after we have been made aware of all the codes, on all the levels, we still will tend to see a photograph as an analogue of what it represents. Watching an old Lana Turner movie, aware that we are "really'' seeing photographic grain, aware of the lighting codes, the makeup codes, all the deceptions of film language, we are nevertheless convinced that in some extremely meaningful sense we are seeing Lana Turner as she really looked at the time the film was being made. Something there is, apparently, that persists through semiotic demystification of all the codes to reaffirm the analogue image.

Eco has himself confessed to the inadequacy of his earlier formulation of the iconic sign, attempting to rectify the problem in the final section of his most recent work, A Theory of Semiotics. The result, 120 pages of dense explication followed by 15 pages of footnotes, is extremely confusing and ultimately inconclusive. A large host of new entities are introduced to account for a seemingly endless string of exceptions and nuances. At certain points, Eco seems to be denying the notion of iconic sign altogether, casting doubt on whether it can even be regarded as a uniquely semiotic phenomenon. In the forward, Eco does, in fact, declare the assumption "that icons are arbitrary and fully analyzable devices'' to be as naive as the assumption that they are "non-coded analogical devices.''14 What, then, can they be? And what can semiotics be without them?

The difficulty has cast a shadow over Metz' project. His complete about face on the issue of analogy does not seem to have altered his involvement with the grande syntagmatique which it, in some sense, justified. His new found knowledge of iconic codes has hardly encouraged him to apply them to the cinema. While no longer attaching any truth value to the notion of the analogue image, he nevertheless has found it impossible to leave aside. Viewers, after all, do respond to films as though the cinematographic image were uncoded. Metz' theory must continue to function as though the analogue image fully applied.15

We must remember that the gap between chapters 3 and 4 of Film Language has never really been accounted for. Metz' powerful argument in behalf of the analogue image and its incompatibility with the most fundamental tenets of semiotics has returned to haunt him. While some aspects of this argument have been refuted others have not. The "early'' Metz had made an eloquent case for film's"bottom to top'' materialism, stating that "the diversity of images in cinema is the world's diversity.''16 This calls forth an analogue, but can it be contradicted? What could it mean to say that "diversity'' is an encoded property of our perception of the world?

This type of difficulty permeates Metz' later writings, particularly Language and Cinema. Here the question of analogical diversity becomes the problem of the "pluricodal text,'' the text with diverse meanings overlaid on one another (by analogy with life itself?). In the face of such a notion, argues Metz, all hope of systematic analysis vanishes.17 Semiotics becomes an ad hoc process of continuous readjustment and compromise.

Leaving aside Metz' confusing initial affirmation of the analogue image, it should be evident that a major theme of semiotics in general is its "deconstruction'' of realism. As we have seen, however, the semiotic critique of realism is seriously compromised: methodologically, over the question of the iconic sign; on general principles, over the infinite regress that turns both ideology and semiotics itself into monstrous Platonic forms, isolated totally from the material world, chasing one another's tail endlessly.

Sensitivity to problems such as these has led to a crisis for semiotics and structuralism generally. For many, as for Metz, there is no longer any real point in attempting to approach fundamental problems of perception and signification systematically. Attention has turned, instead, to certain basic issues stemming from the work of Freud, relating the development of the ego to the acquisition of language.
We will take up this Freudian thread in our final chapter. For now, we must simply conclude that semiotics, while clarifying a great deal, has failed to settle the most fundamental problems posed by realism. The paradox remains.


Chapter 1--The Double Aspect of Montage

1. Many inconsistent and vague descriptions of the Kuleshov experiment exist in the literature. Ronald Levaco, in the introduction to his collection, Kuleshov on Film, states that the exact content of the shots was eventually forgotten by Kuleshov himself. The recollection of Kuleshov's student, Vsevelod Pudovkin, is the basis for Levaco's description, the version presented here. See V. Pudovkin, Film Technique and Film Acting (London:Vision, 1954) p. 140. See also Kuleshov on Film, trans. and ed. by Ronald Levaco (Berkeley:University of California Press, 1974) p. 8.
2. For Kuleshov's description of this experiment, see Kuleshov on Film, op. cit. p. 52.
3. Ibid. p. 191
4. "A Personal Statement'' (1926), in Serge Eisenstein, Film Essays, ed. Jay Leyda (New York:Praeger, 1970) p. 14.
5. "The Cinematographic Principle and the Ideogram'' (1929), in Eisenstein, Film Form, ed. Leyda (New York:Harcourt Brace, 1949)p.38
6. His attempts are discussed in L. Kuleshov, "Art of the Cinema'' (1929), in Kuleshov on Film, op. cit. p. 55.
7. See Jon Barna, Eisenstein (Boston:Little Brown, 1973) p. 63.
8. "The Cinematographic Principle and the Ideogram,'' op. cit. p. 37.
9. "A Dialectical Approach to Film Form'' (1929), in Film Form, op. cit. p. 49.
1O. "The Futurist Synthetic Theater,'' a manifesto of 1915, in Michael Kirby, Futurist Performance (New York:Dutton, 1971 ) p. 199.
11. According to Peter Wollen, Eisenstein "constantly stresses that montage is a dialectical principle...[But] clearly there were some difficulties in Eisenstein's position, of which he began to grow uncomfortably aware. The problem was to reconcile his `idealist' preoccupation with the dialectic with the materialist inheritance he carried with him from the Proletcult [constructivist] Theater ...In fact, Eisenstein proved unable to solve the problems confronting him and eventually tacitly abandoned them.'' Signs and Meaning in the Cinema (Bloomington: Indiana University Press, 1972) pp. 47, 48.
12. S. Eisenstein, The Film Sense, trans. Leyda (New York:Harcourt Brace, 1942 ) p. 4.
13. Bela Balazs, Theory of the Film (London:Dobson, 1952); (reprinted New York:Dover, 1970) p. 119.
14. Jean Mitry, Esthetique et psychologie du cinéma (Paris: Editions Universitaires, 1965); see pp. 283-285.
15. Christian Metz, Essais sur la signification au cinéma, vol. I (Paris:Klincksieck, 1968); translated as Film Language: A Semiotics of the Cinema (New York:Oxford University Press,1974) p. 47.
16. One could say here that the creative imagination of the reader (viewer) is brought into play but this is not really what happens. The reader searches only for conventional meanings. For example, it is not likely to occur to anyone that the child might drop the toy bear, pick up the soup and spill it into the coffin.

Chapter 2--The Realist Reaction

1. According to Brian Henderson, "one does not lightly venture a definition of mise-en-scène, cinema's grand undefined term, of which each person, when examined, reveals a different sense and meaning.'' A Critique of Film Theory (New York:Dutton, 1980) p. 49. For our purposes, we will adopt what, for Henderson, is Bazin's view that mise-en-scène is more or less associated with the image-content of the individual shot in itself, particularly the long take.
2. André Bazin, What Is Cinema?, trans. H. Gray (Berkeley: University of California Press, 1967) pp. 25, 26, 15.
3. Siegfried Kracauer, Theory of Film (New York:Oxford Univ ersity Press, 1960) pp. 28, 299-300, 22, 23, 300-303.
4. As quoted in Kracauer, ibid. p. 309. Originally published in Critique (Jan.-Feb., 1947) p. 27.
5. Bazin, op. cit. p. 15.
6. Film Language, op. cit. pp. 24, 25, 41.
7. Bazin, op. cit. p. 37.
8. See, for example, Gotthard Wolf, The Authenticity of the Scientific Film (New York:Goethe Institute, 1967)
9. Bazin, op. cit. p. 37.
10. Stephen Koch, Stargazer:Andy Warhol's World and His Films (New York:Praeger, 1973) pp. 30, 31.
11. Bazin, op. cit. pp. 12, 13.
12. Koch, op. cit. p. 32.

Chapter 3--Realism, Semiotics, Ideology

1. See Berkeley's An Essay Towards A New Theory of Vision (1709) and The Theory of Vision (1733) in Berkeley, Works on Vision, ed. C. M. Turbayne (Indianapolis and New York:Bobbs Merrill, 1963).
2. Roland Barthes, Elements of Semiology; (Jonathan Cape, 1967) p. 12; originally published in French, in Communications 4 (1964).
3. See Metz, op. cit.
4. Ibid. p. 43
5. Henderson, op. cit. p. 169. I am indebted to this work throughout the discussion of Metz in this chapter.
6. Metz, op. cit. p. 47 and all of chapter 5.
7. Henderson, op. cit. p. 142.
8. Eco's critique of Metz appears in Umberto Eco, "Articulation of the Cinematic Code,'' in Cinematics 1 (Jan. 1970).
9. Ibid. pp. 5-7.
10. Victor Burgin, "Photographic Practice and Art Theory,'' in Studio International (July, August, 1975); reprinted in V. Burgin, Two Essays On Art Photography and Semiotics (London: Robert Self, 1976) p. 6. I am indebted to this essay generally for its clarifications of the positions of Metz and Eco.
11. Jean-Louis Baudry, "Ideological Effects of the Basic Cinematic Apparatus,'' in Cinéthique 7/8 (1970); translated from the French in Film Quarterly (Winter, 1973/74) p. 40.
12. Eco's most recent and somewhat softened version of this position can be found in his Theory of Semiotics (Bloomington: University of Indiana Press, 1979) pp. 15-29.
13. Christian Metz, Language and Cinema (The Hague:Mouton, 1974) p. 103; originally published in French (Paris:Larousse, 1971)
14. Eco, A Theory of Semiotics, op. cit. pp. 178-313, 216 viii.
15. See Metz, Film Language, op. cit., footnotes beginning on p. 61 and also p. 78, where Metz expresses his revised views on analogy.
16. Ibid. p. 115.
17. Metz, Language and Cinema, op. cit. See, for example, pp.101-105 or 131-137.