* STARWORD *         for the best in Film History & Analysis 

This chapter below has been part of Film Style & Technology: History & Analysis since 1983


I take the aim of film studies to be arriving at, and then communicating, the maximum amount of useful knowledge about all sorts of films from the past to the present. I also believe that the best sort of knowledge is constructive, meaning that it has its basis in what we can be fairly sure about, and also that it can be used as a sound basis to build upon further. To handle this task practically, there must be methods of analysing and evaluating as objectively as possible the hundreds of thousands of films that have been made. Analysis is necessary, for nothing can be said accurately without it. Evaluation is necessary, because only some films can be considered, saved, and made available out of many. The maximum amount of objectivity is necessary, because that permits the maximum communication of knowledge. (In a world where total subjectivity and relativism reigns, no meaningful communication is possible.) This means that serious study of the cinema should strive towards, without being able to attain, the nature of the established sciences such as biology and physics, which are identical in England and Russia, America and China. Film studies are unable completely to become a real science because of the essentially innovatory, idiosyncratic, and complex nature of the art object. There are no eternal laws of aesthetics.

Principles like these, though not usually expressed, doubtless because they are obvious, have guided the fruitful development of musicology and art history during this century. An important aspect of this approach is that only as much theory as is necessary to deal with the matter in hand is used; there is no point in the pursuit of totally unnecessary and unobtainable rigour and generality. As P.B. Medawar observed, in the real sciences there are no prizes for tackling too difficult a problem and then failing to solve it. Though in the humanities it seems that if this is done with enough flashy rhetoric one can collect one an admiring crowd of ignorant disciples for a while, not to mention providing a comfortable living for oneself. The other main error of film critics has been too great an eagerness to say just what sort of film is good and what sort bad; by an unconscious desire to justify personal preferences. The two sides of film theory -- the analytical and the evaluative -- should be separated as much as possible, and for both the individual film should be central, not the director or the genre or anything else.


Analysis of films can proceed in two directions, and these are far from equivalent. Most importantly, films can be analysed in terms of their construction and their relation to their makers: analysis in this direction is mostly ignored in theorizing about films. This is strange, because if one insists on describing a film as a coded message, that coded message must have been constructed by the films immediate makers, and the only way to get an accurate decoding must be to reverse the process of encoding. Actually, it should be noted that the film medium, in terms of its narrative function, is not a simple communication channel, but a complex object that is also a representational system, but neverless with aspects that function in other ways, such as communication.

Less importantly, films can be analysed in terms of the response of their spectators. Of course film-makers also form part of the audience, and they are the part that is capable of the fullest response and understanding. It is in relation to them that the film comes closest to being something like a language system.

The obvious factors that influence the creation of a film -- previous films, the technical and other production constraints from inside the film industry and craft, and the more general influence of society and culture -- all act through individual film-makers whose individual differences play a large part in producing the visible variety of films: a variety that will be underlined in the following chapters. Attempts at the type of large-scale ‘cultural history’ or ‘culture critique’ generalizations that ignore the relations of individual film-makers to their work and try to explain everything about films in terms of such tenuous and imprecise abstractions as ‘bourgeois ideology’ invariably founder on the sheer variety of films. The generalizations of cultural history, Marxist or otherwise, are always either banal or false, and indeed experience shows that one generation's cultural history, however diverting it was at the time, is the next generation's waste paper.

Most of the numerous features of a film are the way they are because of conscious decisions by the director, writer, cameraman, etc. (There is a possibility that this was slightly less true in some cases in early cinema, but it is fairly easy to tell the occasions when this happened, if one has seen enough examples of films from the period in question.) So in the first place the narratives of films should be looked at in the terms used to construct them. Besides the well-known vocabulary of ‘scenes’, ‘sequences’ and so on, these terms are mainly of dramatic origin, but there are other more specialized terms such as the ‘switch’.

The conventions of editing and lighting are already supplied with descriptive terms, and where there are gaps they can be filled by the natural extension of existing expressions. For example the existing description ‘core lighting’ can be extended to cover another common type of lighting of figures by the term ‘angled core lighting’. The other dimensions of the medium are similarly provided with their analytical concepts and terminology. More generalized forms of classification such as ‘genre’ and ‘style’ form themselves in the usual way on a secondary level, after this primary analysis has been carried out. The fact that this is fairly obvious and straightforward does not make it incorrect and unsatisfactory.

The film-maker's beliefs, aesthetic and otherwise, undoubtedly influence the product, and they certainly must be considered, as might also the connection of the film-maker's personality with what he does. For this latter purpose, scientific personality theory is sufficiently advanced to guide the investigation without recourse to baseless Freudian speculations.

Towards the Spectator

Unlike the relation of the finished film to its makers and production process, which is fixed and definite for all time once it is complete, the relation of the film to its audience not only varies between audiences of different filmic sophistication as already mentioned, but this relation also varies with time: the reaction of a contemporary audience to a silent film is often quite different to that of an audience in the nineteen-twenties. This fact alone renders this side of film analysis less important, but nevertheless it is well worth pursuing. Proceeding in this direction we are led to the once popular investigation of the sociological effects of films, then to what is essentially part of the psychology of perception, and finally to a consideration of how audiences understand films. As far as the psychology of perception of film is concerned, what is needed is the kind of research begun some time ago in painting and music into the response to the simplest elements of these arts, and then working up to more complex structures. Recently there has been quite a lot of interest in the branch of perceptual psychology called cognitive psychology, though so far this has mostly produced a great deal of theorizing, with little in the way of experimentally supported conclusions of any but the most banal kind. Nevertheless, we should keep our eyes open for more in the way of solid scientific results in this area. The evidence so far in the other arts shows that the response to combinations of artistic elements can be predicted from the response to the elements (colours, shapes, etc.) themselves taken individually, despite the unsupported speculations of Gestalt theorists that this would not be so. It is also undoubtedly the case in these other arts that the personality of the spectator affects his artistic preferences. (See D.E. Berlyne, Aesthetics and Psychobiology, Appleton-Century-Crofts, 1971). Would-be film theorists are still proceeding as though all the members of the film audience are identical.

Leaving aside the understanding that film-makers and those with equal knowledge have of film, the major part of film audiences probably understand film in a very simple way. In fact as an intensified and extended dramatic representation -- i.e. like a stage play with knobs on. Although the idea awaits full experimental verification, my suggestion is that the mass audience does not register in any significant way any minor infractions of editing and other conventions as long as rough temporal continuity is preserved within a sequence. The way a naïve audience comes to understand film representation quite quickly gives some idea that the understanding is as simple as I have suggested. On the other hand, most avant-garde film is quite meaningless to the mass audience, for in that case its appreciation depends on an understanding of its relationship with other films, other arts, the conventions, and other things. Avant-garde film is the most difficult area of all to handle, and most recent film theorizing has little to say about most of it.

The Nature of the Medium and Film Form

The first crude holographic films have already been made, and we can anticipate a complete, all-surrounding audio-visual representation of reality being possible at some time in the future. So the most useful basic way of regarding the medium (and this includes television) is as a more faithful or less faithful reproduction of audio-visual reality. One extreme, as presently possible with the OMNIMAX system, is a series of 70 mm. colour stereoscopic films with multi-channel sound taken of an unstaged event, and projected so as to fill the complete area of possible vision, while the other extreme would be some kind of small-screen abstract film with synthetic sound, or no sound at all. All films can be considered to lie on a spectrum between these two extremes, with a greater or lesser degree of distortion (or transformation) of reality being introduced in various ways: by making cuts between shots rather than running the camera continuously, using zooms and camera movements within shots, shooting in black and white rather than colour, using various degrees of non-natural sound, filming acted events, and so obviously on. The amount of distortion of reality introduced in the separate dimensions of the medium (cutting, photography, sound, acting, the events represented, etc.) is not necessarily parallel between each of these dimensions and the general effect of the film itself, though there is not usually a great divergence.

These dimensions can even be considered in a semi-quantitative way in many cases; for instance the `strength' of a cut or other shot transition can be defined in terms of the amount of discontinuity in space and time introduced into the action by the cut. Another possibility is a precise analysis of the number of shots having various shot lengths in a film, and also the numbers with various degrees of camera closeness and camera movement. It might be claimed that this is a rather arid approach, but considerations of how long a shot is to be, where the camera is to go, and so on, are some of the things with which the director of a film is principally concerned. (Note also that acting style has been included as a dimension of the medium as well.)

Film Style

Questions of style arise when we consider films in relation to other films. If analysis along the lines just mentioned has been carried out, then the distributions of these quantities (shot length, etc.) for a particular group of films, say by a particular director, when compared with those for other directors working at the same place and time, give a sure indication of the existence of a personal style; in fact this is what formal style is. (Analogous analyses have long ago been carried out for the style of literary and musical works.) To give a simple example, in the middle of the nineteen-thirties the cutting rate (or better, the Average Shot Length) in Mervyn Le Roy's films was near the norm for Hollywood films of that time (an Average Shot Length of 9 seconds), whereas in Michael Curtiz's films it was around 6 seconds, and in John Stahl's 14 seconds or longer. Using a measure like this, or indeed other more complex ones, it is also possible to compare the range of variation characteristic of American films with the different range holding for the contemporary French films, and so on. When this method of norms and differences is generalized to all the features of films it can help to avoid the frequent error of describing as unique what is in fact a common feature of a large class of films from a particular time or place or genre. For instance, Noël Burch has described the use of dissolves from a Long Shot to a closer shot in Caligari as -- subverting the codes -- (a pointless synonym for ‘breaking the conventions’), whereas in fact this usage was fairly common in German and American films during World War 1, and indeed through into the 'twenties in Europe. The error of failing to take the context into account is very common in writing about older films, and I have already referred to another instance of it in a previous chapter.

It could be argued that often the individuality of a film-maker lies in the verbally expressible content of his films, and indeed it often does in part, but this individuality of content will mostly be found to be allied to formal individuality if the analysis is carried far enough.

The importance of formal style analysis is beginning to be realized, but it still has not got much further than remarking things like the fact that Howard Hawks keeps the camera at eye-level and doesn't move it if possible. But in fact there are other directors of his vintage who do this too. For instance Henry Hathaway. (Keeping the camera at eye- level makes for efficient shooting because the actors can be kept well-framed at all distances without tilting the camera up. If the camera were tilted up, the lighting set-up would sometimes have to be changed to keep the back-lights out of shot.) The real stylistic distinction is that further than this, Hawks keeps his Average Shot Length a little longer than normal, whereas Hathaway uses faster cutting.

Some attempts at style analysis have unfortunately been conducted in spurious terms that ignore conditions imposed on the director, and also the relation between the approach of a particular director and that generally prevailing at the period in question. For instance, the style of Douglas Sirk cannot be simply pinned down by talk about mirrors and flat shiny surfaces. Mirror shots are quite common in dramas made by ordinary Hollywood directors from the nineteen-thirties onwards (it makes shooting a studio scene more interesting for the director), and insofar as Sirk's films have flat, glossy surfaces this is due to the art directors at Universal Studios and the deficiencies of CinemaScope lenses. (The squeeze ratio of CinemaScope lenses varied with object distance, so emphasizing the existence of the picture plane.) Actually, Sirk's formal style is distinguished by a so-far unremarked excess of low-angle shots over the norm. To judge by an unprompted statement of Sirk's, this resulted from a seeking for expressivity on his part.

The formal spectrum covered by the cinema, that I described in the previous section, when translated into terms of style becomes a spectrum stretching from extreme naturalism to extreme expressivism.

If one looks back to statements made by Hollywood directors in past times, it is apparent that they mostly saw their task as one of expressing the material in the script – ‘putting the story across’ -- in the most effective way, and a point at issue between them was just how much expressivism to use, and how much naturalism. The general desire was to affect the audience in the appropriate way, and this called for the application of unmentioned supplementary principles, unmentioned because traditional in the drama and other arts, and so obvious to film-makers, such as internal consistency in all aspects of the film to maintain suspension of disbelief. This is closely related to Victor Perkins' principle of internal coherence. Incidentally, many of the examples discussed in Film as Film are cases of the expression of the script content through formal devices. Indeed discussion of the detail in a film in these terms is not new, but it has nearly always taken place within a framework that unfortunately assigned aesthetic values to particular styles and contents.

Of course nearly all commercial films occupy a fairly small central region of the style and form spectra, but the extremes are increasingly taken up by films of the avant-garde. These are still denied satisfactory discussion, partly because the terms for this are lacking, partly for less creditable reasons.

At this stage questions of value, of aesthetics, are still excluded, but there are still lots of things that can be said about films, even in a more general way. For instance, we can say that Bergman preferred to film in black and white rather than colour at that period when he had a choice, because he wished to make films that were more expressivist than the norm. We can talk about how the degree of naturalism of the average entertainment film has changed over the years, and about many other interesting matters. And we can talk about films like Godard's which have different parts made in different styles.

In other words, the interaction between style and content is a second-order effect that can be dealt with once the first approximation in the analysis has been carried out.

The Evaluation of Films

Now that form and style have been considered, aesthetic evaluation can be dealt with without creating confusion. My criteria for doing this are, in the order of the weight to be attached to them: firstly, the originality in all respects of the film; secondly, the influence it has on other films; and thirdly, the degree to which the film-maker has fulfilled his intentions in the finished film. The criterion of influence on other films should also be weighted according to the excellence, by these criteria, of the films that are influenced by the film in question. These criteria are the most objective possible, and are equally applicable to every type of film, which cannot be said of previously proposed criteria.  

My first two criteria are completely realizable in principle, with the proviso that the iterative part of the second criterion, which requires that the influence on other films takes account of the quality of the films influenced, has to be calculated with a reducing factor applied on each repetition, in which case it can be  cut off after a fixed number of cycles to give a sufficiently correct answer. The third and least important criterion does present some difficulties in application. Nevertheless it can be worked well enough for practical purposes, even when it has greater weight thrown upon it by the inapplicability of the second criterion when we are considering new films. The conscious intentions of the film-maker (or makers) can usually be found out or reconstructed with sufficient accuracy for this purpose by taking a little trouble. Although something has been done in this direction already by interviewing film-makers, a certain amount of misleading information is produced because interviewers do not know enough about the subjects on which they ask questions. A case in point is the many manifestly untrue things that are said about ‘north light’ in Charles Higham's Hollywood Cameramen. Such errors are partly produced by the asking of incorrect leading questions, to which Hollywood types have a tendency to reply with the answer expected of them, and partly by the boastful exaggeration endemic in Old Hollywood. In fact my principles provide the justification, which is otherwise lacking, for the interviewing of film-makers.

Although the criteria I have put forward here are the most objective possible, they do not quite provide a calculus which can be operated mechanically to crank out values. They need to be applied by people who have viewed large numbers of films with analytic understanding in the ways I have indicated, and indeed there have always been some people in this position. They have had significant things to say about films, even if unclear principles have sometimes led them astray.

The evaluation of films by aesthetic criteria to which their makers did not subscribe also seems fairly pointless, and tends to look foolish in the light of history; the most famous example being the attitude of Socialist critics to von Sternberg’s films in the early 'thirties. As is also recognized now, Bazin's consideration of the films of Welles and Wyler in terms of concepts that were not those of their makers was also factually inaccurate and logically confused. More recently the evaluation of films purely in the light of moral educational concepts such as ‘maturity’ has risen and fallen, and now purely political values are being pushed to the fore again. There is no doubt that large numbers of film-makers do not subscribe to these values, and so the extremely limited usefulness of these approaches in film terms should be recognized.

The Auteur Theory Revisited

It does not seem to be fully appreciated that, in its original form, the ‘Auteur Policy’ as it developed in Cahiers du Cinéma was influenced by the way that film-makers see films; the people who were its leading proponents were just starting to direct films, or were thinking about doing so, and they were looking at films of the past, or what was then the present, for guidance about how to handle the camera, when to cut, and so on. The operation of this interest enabled them to see individuality and skill at work in films where it was invisible to the ordinary critic, but it also led them to slightly overvalue many of the American films of the nineteen-fifties.

The ghost of this semi-conscious use of film-makers’ ways of seeing persisted in the form that Andrew Sarris gave to the Auteur Theory. Amongst the numerous subsidiary evaluative criteria that he introduced, mostly to demote the film-makers whose films were unsympathetic to him, was the criterion of craftsmanship. This is a relevant criterion, and it is included as a part of the third general principle that I have proposed, which is that the film succeeds in fulfilling its makers intentions.

Because of all this, my theory, when applied to the American sound cinema, produces evaluations rather similar to those of Sarris in the upper ranks of film-makers. But it moves Stroheim and Wilder up a step, and it moves Raoul Walsh down as being little more than an excellent craftsman, to mention only the more obvious adjustments. Since my approach works through individual films in the first place, important film-makers are just those who have produced important films. There is nothing in my theory that says that all films by a particular film-maker must be good, or even interesting, and in fact for most film-makers their most important work is concentrated towards the earlier parts of their careers. This is because the strength of will and body needed to control the film production process, necessary to a far greater degree than in the other arts, decreases with increasing age.

The principal reason that my theory of film evaluation produces rather similar results to the Sarris Auteur Theory is that ‘originality’ and ‘expression of the maker’s personality’ amount to almost (but not quite) the same thing in practice, but a subsidiary reason is Sarris’ personal sensitivity and the fact that he has seen and compared a very large number of films. Apart from the elements of confusion and illogicality that spoil Sarris’ work, he himself admits that there are exceptions to his theory: films such as Casablanca that are better than their directors. There are no exceptions to the theory advanced here. The importance or excellence of Das Cabinet des Dr. Caligari is not because Robert Wiene was a great director, but because of its originality and influence.

In Conclusion

One of the merits of the theoretical framework for film put forward here is that it has spaces to accommodate quantities of useful work, writing, and information that have been produced in the past, and are still being produced by a large number of people interested in films. For of course it is hardly likely that intelligent people who know a great deal about films could be totally mistaken in everything they say. Yet this seems to be the position of those who hold the ideas described in the first section of this book. And even further, they appear to believe that by reading a few approved books and articles, and seeing a small number of approved films, one is in a position to understand everything about cinema that matters. It is a conception of the cinema that would limit it to serving some extra-filmic concerns of the moment, and in its most extreme form wishes to dictate what sort of films should be made.

There is no royal road to knowledge about films, or about any other art for that matter.

By contrast, the theoretical framework that I propose for film studies puts the difficulties where they belong, in dealing with individual problems, and makes it possible to deal with those that are soluble one by one in a sound and useful way, as I hope the rest of this book demonstrates.

If film-makers did not make films, following as they do their own ideas about what they are doing, there would be nothing to support would-be film theorists who write about films. At a time when there are already a number of film-makers who are as well-educated, as clever, and who certainly know more about films than most theorists, a certain humility should be in order.

Read about 1907-1913 in