ldeology, determinism and
relative autonomy

by Michael Rosenthal

from Jump Cut, no. 17, April 1978, pp. 19-22
copyright Jump Cut: A Review of Contemporary Media, 1978, 2005

One of the key problems confronting Marxist film theory, as well as Marxist aesthetics in general, is the status of ideology as a determined product of social, and specifically economic, relations. It is a fundamental and unavoidable premise of any Marxist enquiry that ideology — the "consciousness" of people in society and the material cultural products in which this consciousness is embedded — is, in fact, determined.

"The sum total of these relations of production constitutes the economic structure of society, the real foundation on which rises a legal and political superstructure and to which corresponds definite forms of social consciousness … It is not the consciousness of men [sic] that determines their being but, on the contrary, their social being that determines their consciousness."[1]

"The ideas of the ruling class are in every epoch the ruling ideas, i.e. the class which is the ruling material force of society is, at the same time its ruling intellectual force … The ruling ideas are nothing more than the ideal expression of the dominant material relationships, the dominant material relationships grasped as ideas …"[2]

Such premises are basic to the materialist conception of history as a whole. They are instrumental in distinguishing materialism from idealism, thereby making the radical critique of ideology possible.

These assumptions, however, also give rise to a major danger when we attempt to apply Marxist theory to particular films — a danger which is commonly, if vaguely, referred to as "vulgar determinism." The term is hard to define sharply because there are few proponents of vulgar determinism per se who will enter a debate on its behalf. Basically, it involves the assumption that the ideological superstructure is totally passive, merely reflecting processes that are going on elsewhere in the economic base. Thus, one might describe ROCKY as a "bourgeois" effort to baffle and confound the working class, deluding the audience with false hope and distracting it from its revolutionary tasks. Or one might describe a Godard film as "petit-bourgeois" due to the class background of its creator. The shortcomings of such as approach are evident:

  1. It reduces the film to a single meaning, as a one-to-one political allegory, thereby impoverishing it.
  2. To the extent that it is true, it applies so broadly as to be useless — if all Hollywood films are merely "bourgeois," we cannot distinguish between them or explore them in any useful way.
  3. It fails to account for empirically perceived limits to the ruling class' use of the ideological apparatus, for example, the absence of
    films glorifying the bicentennial, or the very few glorifying America's role in Vietnam.
  4. It leads to a political quietism; if film practice is merely the passive reflection of something else, there is no point trying to accomplish anything within it.

Admittedly, there are very few examples of this sort of writing in JUMP CUT and other radical film journals. Nevertheless, the specter of vulgar determinism is always present as a danger and a threat, something we react against, and attempt to avoid at all costs. Among those costs, I feel, has been the substitution of a vaguely "radical" film theory for a properly Marxist film theory, an inability to situate given films within the analysis of classes and class struggle developed by Marx. The problem is determining how to avoid the traps of "vulgar determinism" without sidetracking determinism, and hence materialism, altogether. Must an understanding of ideology as determined by "social being" lead to a mechanistic reduction of film texts?

This problem, in its general form, has long been a central topic of debate among Marxists. Engels wrote numerous letters after the death of Marx protesting, with increasing urgency, the economic reductionism taking hold in the Marxist movement. [3] Recently, French Communist Party theorist Louis Althusser has attempted an important reformulation of the terms of the problem. [4] Drawing from Engel's letters, Althusser argues that the various superstructures (law, politics, ideology) are characterized by a "relative autonomy" from the base. These structures take concrete form in material apparatuses (such as the judicial apparatus or the state apparatus), which have their own specific unity and coherence. Each plays a part in determining social events, although the economic structures remain "determinant in the last instance."

These formulations while providing more sophisticated methodological tools, do not in any way "solve" the problem. It is too easy, when writing about film, to simply re-inscribe the old terms of the problem into the new phraseology, to conceive of relative autonomy as a kind of escape from the rigors of economic determinism, an escape which is always, inexplicably, foiled in the "last instance." Thus, although we can describe films like GODFATHER II or CHINATOWN as "critiques of capitalism" because of their relative autonomy, they remain determined by capitalist economic relations, in the "last instance." The character of that determination, piled onto the last instance, remains a mystery. Furthermore, we often present it as some kind of antithesis of relative autonomy, as if it were a version of the antithesis of "freedom" and "necessity."

In consequence, our understanding of economic determination (which is a condition for an understanding of autonomy) remains in essence the same as in "vulgar determinism," while relative autonomy becomes a sort of ongoing exception. That is, we continue to see the economic base as mission control, which beams out commands to passive agents in the superstructures. With our magic protective shield of relative autonomy, we are safe from these commands until they catch up with us in the dreaded last instance. And, as Althusser somewhat cryptically assures us, "the lonely hour of the last instance never arrives." [5]

It is my impression, judging from my own difficulties and what I can surmise in the work of others, that a certain embarrassment around the concept of determination presents a continual frustration in all areas of Marxist film theory, including reviews of current films. We simply do not know how to insert our political understanding of the social environment, or our data about the economics of the film industry, into a discussion of a particular film without becoming mechanistic. Further, this has led to an evasion of the theoretically sticky areas of political film theory, with a consequent over-emphasis on questions of ideology and form, which critics can speculate about without the need for outside social reference.

The lack in film theory of an adequate concept of determinism and autonomy has been a major factor in this impasse. To help develop such a concept, I propose a "detour" into another area of Marxist study, whose theorists have paid a good deal of careful and rigorous attention to the question of relative autonomy — the Marxist theory of the capitalist state.


In the absence of any strong tradition of its own, recent Marxist film theory has tended to advance through detours into the methodologies, and often the terminologies, of disciplines related to other subjects of study. In the past several years, linguistics and the language-oriented science of semiology have been a dominant influence. While these methods have helped us to decode and identify ideological messages in films, they have serious shortcomings in terms of our overall understanding of ideology per se. As Geoffrey Nowell-Smith pointed out in a recent JUMP CUT, linguistics concerns itself with the exchange of sign and meaning (in communication and understanding) and not with the production of signs (as in a Hollywood studio). [6] It therefore skirts the issue of political economy, and social class relations.

If we do not recognize the limitations of the linguistic model and attempt to extend the analogy overenthusiastically, that is, if we attempt to analyze ideology as a language, the limits of the linguistic model impose themselves in a way that limits our understanding of ideology as well. As a model of successful, completed communication, linguistics can hinder our exploration of the mechanisms of misrepresentation (intentional or otherwise) and misunderstanding that characterize ideological processes. Moreover, it becomes easy to misconceive ideology as a unitary, monolithic system, a great social equalizer that penetrates all "consciousness" in the same way. This can lead to an incorrect notion of "bourgeois ideology" as a solid, consistent bloc of ideas, spoon-fed to a passive population. The point here is not to attack semiology, but to stress the repercussions of ignoring the limitations of our analogies.

In particular, semiology can do little to illuminate the problem of determinism and relative autonomy, which it therefore tends to obscure from view. I feel that we can usefully approach this problem if we step outside of the "liberal arts," and explore the work that has been done analyzing the relative autonomy of other levels in society — particularly in Marxist state theory. Such a study cannot, of course, lead to an adequate general account of ideology, but in certain respects the points of analogy are compelling. We should recall that Marx considered political forms to be a level of the superstructure (he usually mentioned them first, throwing in culture as an afterthought) and therefore, to be determined by economic relationships. The same danger exists, as with film, of the mechanistic approach, of viewing the state as simply a passive tool of the ruling class (or dominant faction within the ruling class). This approach, called instrumentalism, is the basis of a lot of the work which undertakes to prove that state functionaries are in fact members of the ruling class (as if this told us anything about how the state operates) as well as to prove strange conspiracies and the like.

On the other hand, there is also the need to combat, while avoiding these errors, the bourgeois idealist notion of the state as neutral arbiter of social conflict, standing above the particular interests and expressing the "general interest." Hence Marxist state theorists must account for the many possible forms of the capitalist state (from liberal democracy to fascism) and for the many, often contradictory actions taken by the same state, while at the same time demonstrating how, in all of these forms, the state remains essentially capitalist. They must specify the character and limits of the relative autonomy of the state, as a guide to political action. [7]

Work on this problem has developed far more thoroughly and rigorously than it has in the area of ideology. The major thinkers of the Marxist tradition have addressed a good deal of their attention to the problems of politics, for the simple reason that most of them were political activists and not cultural activists. State power represents both the immediate goal and immediate enemy of practical revolutionary activity. Understanding the mechanisms of the state is a condition for formulating effective tactics and strategy. We should recall (to our sorrow) that the recent, feverish attention lavished on ideology is probably a result of the historical separation of Marxist theory from working class politics. [8] Furthermore, the consequences of theoretical errors are more immediately apparent in the area of politics. For example, the mistaken theoretical line of the Communist International, in the period 1928-35, which held that fascism was the last desperate gasp of a dying capitalism, produced disastrous political consequences as Communist parties, "assured" of their imminent victory, refused to co-operate with any other anti-fascist forces.

In this essay I will survey some of the analyses developed by Nicos Poulantzas, a Greek Marxist writing in French. [9] This does not mean that Poulantzas' work is synonomous with Marxist state theory — it is one position in a many-sided controversy. What follows should be considered as an entry into the discussion, not as a conclusive statement. I am only selecting from his work those points that concern the questions I am posing.


Poulantzas begins his analysis by attempting to situate the concept of politics and the state within the overall concept of mode of production. This concept, one of Marx's greatest contribution to the analysis of human social formations, denotes a specific unity of the forces of production (factories, labor, technologies, etc.) and the social relations into which people enter in order to carry out productive activity. In societies divided into classes, these social relations are always antagonistic. In this sense, mode of production designates the specific form in which one class is exploited by another, the manner in which the surplus labor of one class accrues to the benefit of the other.

In the feudal mode of production, the basic social relations are between the landlord class, and a class of peasants who own their own tools, but must devote part of their labor to the landlord's crop. In the capitalist mode of production, the "tools," such as factories or IBM machines, are not owned by the workers, who have nothing to bring to the production process but their ability to work, which they sell for wages. Capital, therefore, is not a thing (like a certain amount of money), but a social relationship. A factory only becomes "capital" when there is a capitalist class and a working class, just as a black person only becomes a slave within certain social relationships. A mode of production therefore is an ensemble of social relations in which material values are produced, and the dominant social relations themselves are unceasingly reproduced.

It is important to distinguish "mode of production," which is an abstract, analytical category, from a concrete, historical social formation (such as the U.S. in the 70s), in which several modes of production, and several phases of the development of a mode of production overlap, with one always dominant. There are still a number of craftspeople who own their own tools, but the commodities they produce enter the market system of monopoly capital, which therefore does not determine everything, but does dominate everything. In order to understand any social formation, it is necessary to understand the basic social relations given by the concept of mode of production.

Most of the relations I have used as examples have been economic, relations of exploitation. Poulantzas, following Althusser, points out that the mode of production is constituted, not only by economic social relations, but by political and ideological (and other) relations as well. He refers to these as regions of the mode of production (the spatial metaphor is always cumbersome, but perhaps unavoidable). Each of these regions has its own characteristic structure.

He then distinguishes between political, economic and ideological structures in terms of the social relations, and social activities, or practices, which these structures organize. (Remember that these activities are always contradictory and antagonistic in keeping with the basic class character of capitalist society.) The object of economic activity is production in the strict sense; that of ideological activity is representation — what Althusser calls our imaginary relation to our real conditions of existence. Political activity has as its object the material alteration (or stabilization) of prevailing social practices and relationships.

This broad definition includes fundamental transformations, as in the case of revolutionary political activity. But it also, and more frequently, covers the adjustments in social practice that have to be made in order for capitalism to deal with new and changing problems, without at the same time challenging the basic relations of domination. For example, when the energy monopolies are faced with new economic contradictions, the consequent adjustments in social practice are worked out through legislatures, regulatory agencies, and other parts of the political apparatus. The point here is not to distinguish between some activities which are, and others which are not, political; but rather to analyze activities in terms of their political ramifications. In this sense, the contradictions that structure the political region are between practices that aim at maintaining the prevailing relations, and those that aim at transforming them.

The state is a political apparatus whose function is to ensure the reproduction of the system as a whole. If we refer back to the basic political contradiction (maintenance/transformation), it is clear that the state is entirely on the side of maintenance, and therefore serves the interest of the dominant class. Instrumentalist theory would agree with this assessment, and claim that the state is therefore a tool, entirely responsive to the will of the ruling class, which they can use to bludgeon down social contradictions. Poulantzas draws a different conclusion. He describes the state as the factor of unity and cohesion in society: the apparatus which keeps it from flying apart under the pressure of its intrinsic contradictions. Its role is to regulate these contradictions in order to maintain the unstable equilibrium of the system (unstable because based on contradiction, equilibrium because the system of exploitation does in fact manage to reproduce itself, along with its constituent contradictions, from day to day and generation to generation).

It cannot perform this role as a tool, lined up squarely on one side of social contradictions. In order to regulate the contradictions that must necessarily exist in a class society, the state must include and contain them, so that these contradictions are condensed within it. As Marx put it, the state is the official resumé of society, a resumé, that is, not only of the elements of society, but also of their contradictory social relations (class struggles for short). [10]

A key feature of this approach is that it sees the state not as an object, a static thing, but as a social relation, just as Marx insisted that capital is not a thing, but a social relation. Similarly the state, while maintaining (and in order to maintain) dominant class relationships is also in Poulantzas words, "shot through and constituted with and by class contradictions." [11]

Poulantzas discusses the example of European social democracy and reformism. Radicals often treat reformism as a "co-optation" of revolutionary demands, and while this is in some sense true, it is incorrect to think of it as a devious bourgeois scheme, elaborated out of thin air to confuse the masses. The bourgeoisie would not, "by themselves," have elaborated reformism. The very notion is absurd because the bourgeoisie are never by themselves; they exist as a class only in relation to other classes. Rather, reformism is an effect of working class struggle on the political region, within the context of that class's fundamental political subjugation.

Social security, unemployment insurance and, by extension, the whole welfare apparatus are examples in recent U.S. history. These are administered in such a manner as to reproduce relations of subjugation and exploitation. Recipients are systematically humiliated, their lives are regimented by endless petty rules, they can be injected into the labor market and yanked back out of it by slight changes in the regulations, they are held up as a threat to the rest of the workforce, etc. However, if we forget that these systems came into being only after long and arduous struggles by the working class, we slip into something similar to the conventional myth of Roosevelt as a magnanimous sugar-daddy who "gave" the people social security out of the goodness of his heart.


We can see from this example some of the important principles of the relative autonomy of the state. As the cohesive, regulating factor of the social formation it must make allowances (within strict limits) for the class interests of the dominated as well as the dominant classes. It is not a question of "concessions" made by the state (for this would imply that the state is a unified and conscious entity capable of entering negotiations) but rather of concessions and compromises within the state. Since its role is to reproduce (maintain) a complex unity based on contradiction, it cannot be a monolithic, fissureless bloc, but is itself, by virtue of its very structure, divided. The state is not a tool in the class struggle, but an arena which is controlled and "fixed" by the bourgeoisie, but in which, nonetheless, a real struggle goes on.

It is therefore incorrect to pose the idea of relative autonomy as if it were somewhere intermediate on a spectrum between total determination by the economy or ruling class interests, and total freedom from these determinations. Relative autonomy is not an "escape" from determinism, tied down only by the "last instance;" it is the specific form through which determinism is exercised.

Similarly, it is incorrect to attempt to construct a general mode of relative autonomy; to pose such questions as "how relative is relative autonomy." For relative autonomy is not an idea, but the result of a material set of social practices. The concrete form taken by this autonomy depends on the conjuncture of class struggle at any given time. Liberal democracies and fascist dictatorships are both political forms of capitalist domination, but they clearly have very different degrees of autonomy inscribed in their structures.

This theory emerges, in part, as a sustained critique of what Poulantzas calls "the totalizing instance" (and what Althusser calls "expressive totality"). This is the idea that a single aspect of a totality can be the origin and reference point of the totality and everything within it. (I am expressing it in this general form because the "totality" in question may be the whole of a social formation, or its ideology, or it may be a single film.)

The classic examples of this style of thought are in Hegel, who saw the Roman period as "the age of Law," the Middle Ages as "the age of religion," etc. That is, everything done or said in the Roman period was simply a reflection of the animated principle Law, so that Law "totalized" all of Roman society.

The same kind of reasoning is involved in mechanistic determinism. In this case, the economically dominant class becomes the "totalizing instance:" all art is bourgeois art, all science is bourgeois science, all in all it is a bourgeois society.

The crucial point is that Marx based his materialism on social relations of contradiction and struggle, while "totalizing" tends to eliminate these from the analysis. For example, it would not be correct to consider the antebellum South as a "slave-owners' society." One might employ this term with the best of political intentions; but it implicitly denies that the culture, the work, and the struggles of the slaves (on a day-to-day basis as well as in mass insurrections) were an essential, constituent part of the whole society. It is more correct to call it a "slave-owning society," structured by the contradictory social relations between slave-owners and slaves. Similarly, we do not live in a "bourgeois society;" but in a society in which the capitalist mode of production is dominant, and in which the bourgeoisie is therefore the dominant class. This is not always the most convenient thing to say, but it is useful to keep in mind when we are attempting to analyze social (including filmmaking) practices.


What is the bearing of this material on our study of ideology in films? It struck me, when reading Poulantzas, that his discussion of state power touched on problems that had most baffled me when I tried to write about films — in part because his description of the state, as a "factor of cohesion in a social formation," would seem to apply, without much stretching, to ideology as well. Nevertheless, as I cautioned earlier, we have to watch out for overstepping the limits of an analogy. We cannot project a description of the state directly onto the problem of ideology, without distorting or denying) the particular unity and cohesion of ideological structures. [12] Ideology and politics do not operate in the same way, and we need a clearer sense of the distinction between them, based (for example) on the specific character of representation, and on the specifically commercial nature of much of the ideological apparatus — the relation of Gulf and Western to a ticket buyer is different from that of a state agency to a citizen.

What ideology does share in common with politics (and this was the point of my analogy) is that each is characterized both by class domination and by relative autonomy. Thus, while the study of state power cannot generate an adequate general description of ideology, it can help us to formulate some of our questions about ideology.

It suggests to me, for instance, that we should not conceive ideology as a thing, as a completed and coherent system of ideas which the bourgeoisie utilizes to brainwash the rest of society. Rather, we should approach ideological processes as social relationships, "shot through with and constituted by" class contradictions. The concept of ideological domination implies ideological struggle; it does not imply the complete elimination of social contradiction from the sphere of discourse. This is in part what distinguishes ideology from propaganda. Ideological domination does not exclude social struggle, it includes social struggle. If successful, it subsumes various, contradictory discourses within a relatively coherent (but always unstable) discourse of domination. Poulantzas puts it as follows:

"The structure of the dominant ideology cannot be deciphered from its relations with a 'class consciousness' considered in a vacuum, but from the starting point of the field of class struggle, i.e., from the concrete relation between the various classes in struggle, the relations within which class domination functions. Hence we can understand not only why the dominated classes necessarily experience their conditions of existence within the discourse of the dominant ideology, but also why that discourse presents elements borrowed from ways of life other than that of the dominant class." [13]

In order to develop these ideas rigorously, we would need a general theory of ideology, including a workable definition of ideological class practices. I do not pretend to have such a theory in my pocket, and do not have the space here to discuss the problems involved. This absence imparts an unavoidable vagueness to the discussion, a vagueness that I feel is implicit, if unstated, in most of what appears in JUMP CUT. Given this limitation, I would like to suggest a number of possible lines of enquiry, and areas for future work, prompted by my reading of state theory.

The first concerns our discussion of individual films, particularly films in current release. This is an important area of work, because it is through movie reviews that Marxist film theory most often comes in contact with a wider readership. Yet movie reviews lose their impact under the weight of fruitless debates over whether a given film is "bourgeois" or "progressive" — attempts to totalize the film on the side of one or another team.

Marxist film critics have long been aware of the extreme complexity of the ideological inflection of mass films; but I feel we have lacked a theoretical framework adequate to generate new knowledge about these problems. When we see films, such as ROCKY or WHITE LINE FEVER, which present certain working class experiences and resentments in a genuine form, yet which work towards complacent or reactionary resolutions, we are unsure on which side of the grand political scale to assign them. We find it difficult to praise the commodities of the monopoly capitalists who own Hollywood, but we are unwilling (properly) to trash every Hollywood film that comes down the pike, a strategy which would consign our criticism to sectarian irrelevance. Often we simply tack on to favorable reviews a standardized note that the film "stops short" of dealing with the real causes.

If we view these contradictions as characteristic instances of relative autonomy as it operates in the field of ideology, we may come closer to an understanding of ideological practices in film. The film is a "determined" product of society because the class contradictions, which determine the whole structure of society, operate as well within the ideological structures. It is further determined in the sense that the dominant ideological discourse in society is generally dominant in the Hollywood film. Yet this domination is not exercised directly as a simple tool of mind control, as "bourgeois propaganda." There are virtually no films which directly sing the praises of the Du Ponts and the Rockefellers. In order to be effective in maintaining bourgeois hegemony in the long-run, the structures of the apparatus must permit the overriding of short-run ruling class interests by giving expression to what Poulantzas calls "elements borrowed from ways of life" of the dominated classes. These are not wholly absorbed into a "bourgeois world view," but retain an integrity as one of the aspects of a contradictory and conflicting unity.

It might be possible, then, to view an individual film, as well as the ideological apparatuses as a whole, as a site of ideological class struggle. This does not mean that the outcome, within a given film, is ever in doubt, that the proletariat might suddenly and unexpectedly emerge victorious. It might mean that within the action of the film, the relations of ideological domination are worked out, as potentially subversive material is articulated within the dominant discourse. The film must actually perform the work of making the various contradictory elements cohere in an unstable equilibrium. This would imply that a film is not a homogenous totality with a single ideological "message," but rather presents a conflicting unity.

This is the case even with such overtly propagandistic products as cold-war anti-Communist films like RED NIGHTMARE or I MARRIED A COMMUNIST. These films did not attempt to depict even a distorted image of Communist ideology; their Communists are presented simply as gangsters. What they do depict, and very clearly, are the changes brought about in society by the development of monopoly capitalism (increased conformity of consumption, destruction of family ties, etc.), articulated in the form of the dominant anti-Communism. The most self-consciously progressive films made in the same period (e.g., HOME OF THE-BRAVE) attempted to focus on social injustices in such a way as to emphasize the fundamental soundness and perfectibility of the dominant social relations. In each case, the key principle is one of ideological relationship rather than ideological "message." Clearly the terms of this relationship have to be more rigorously worked out; but we must be able to pose these questions before we can begin answering them.

A materialist understanding of relative autonomy can also help us to formulate the kinds of questions we pose when studying the material apparatuses of the film industry. Too often we merely allude to the self-evident fact that Hollywood film studios are owned by the bourgeoisie and run for profit, and we leave it at that. This would be enough to account for the prevalence of "bourgeois ideology;" but if we accept the idea that the industry allows within its structures for a certain ideological latitude, it is necessary to develop a far more rigorous analysis of the structures of the industry, to determine the range and limits of that latitude. All of the large media industries are out for a buck, yet each differs in its characteristic structures and forms of autonomy. For example, it seems to me that television is on a much tighter leash, in terms of autonomy, than is film, which is itself on a tighter leash than the print media. (Consider the excision made in a single work as it passes from print to screen to tube.) If we ask why this should be the case, we must consider what the differences are, in these apparatuses, which permit differing levels of autonomy.

Part of the answer to this particular problem is that each of these media has a different capital entry threshhold. You can produce a book and distribute it nationally for a few thousand dollars, while the production and national exposure of a television show requires access to millions. But this is only a first step toward the answer, and immediately suggests new problems, such as the differing content of "distribution" when applied to books and TV and films.

What mode and level of control does a distributor exercise, as opposed to an exhibitor, a scriptwriter, or a major shareholder in Gulf and Western? Under what conditions are constraints exercised directly (as when a studio shelves or reshoots part of a film) and how are indirect constraints structured into the "free play" of the system? What (if any) has been the effect on ideological practices of the absorption of the majors into multinational corporations — and if there has been no effect, why not? What is the role of the audience in determining the presence and extent of non-dominant ideologies within films? Marxist film theory should be able to deal with these questions in concrete and specific terms. To do this, we must attempt to analyze the film commodity with the same care and rigor that Marx devoted to his analysis of commodities in general. Such an analysis must certainly involve a clear formulation of the separate moments of production, distribution and exhibition, and an understanding of the social relations that structure each of these processes, as well as their interaction in the overall reproduction process of the film industry. While I feel that a good deal of raw data has been dug up on these issues, the theoretical terrain on which we can position this data has been lacking. A clearer understanding of relative autonomy can help to establish this terrain.

This exposition runs counter to a concept of autonomy which underlies much recent Marxist film theory (although it is rarely directly spelled out in these terms): the notion of autonomy as a kind of escape from the grim and perplexing laws of determination. In this view, the overwhelming majority of films are totally determined by "bourgeois ideology," "bourgeois form," "bourgeois relations to the spectator." The privileged terrain of "autonomy" is reserved for a few, exceptional films, which carry off mini-revolutions by transforming the "mode of production" of film (its formal signifying techniques) in purely filmic terms.

I have no desire to challenge the potential value of research into the ideological implications of cinematic form. But it does not help to base this research on ad hoc theories which use Marxist categories in a purely metaphorical sense. Such is the case, for example, when the term "bourgeois" is detached from all socio-political frames of reference, and used to refer to whatever is typical and conventional. Consequently, whatever challenges the typical and conventional becomes "revolutionary" in the fullest allowable sense. Eventually, this leads to such notions as Godard having a "non-bourgeois" [14] camera style, or Straub carrying out "significant political activity on the level of theory." [15]

The same logic can equally well support the claim of a "revolutionary new laundry product," which carries out a transformation in the autonomous "detergent" region. Or it can support the production of a "revolutionary chair," which, by means of a properly glued-on tack, calls into question the bourgeois, consumerist attitude of the sitter toward the chair.

The central problem of Marxist film theory is to correctly situate film within the class struggle, not to find an analogy in film for the class struggle. The isolation and ineffectiveness of formally innovative film raise questions which must be confronted head-on, and not transformed into a virtue. Otherwise, we return to the idealist vision of the artist as "unacknowledged legislator of mankind (sic)," once a revolutionary in the private region of his/her soul, now a revolutionary in the private region of his/her theory.

The relative autonomy of ideology does not mean that it can follow a purely independent form of theoretical development; nor does it provide a mode of escape from class domination. Rather, it is a mode of class domination, a necessary and integral characteristic of the ideological apparatus under capitalism. We can understand and analyze A STAR IS BORN as an instance of relative autonomy as clearly as we can a film by Godard or Straub-Huillet. And we could add (for the sake of argument) that the Straub-Huillet film is a determined instance of the relations of domination because it reproduces an ideological cleavage within film audiences, fostering open contempt for the "laziness" of the mass audience.

It is interesting to note that much of the formalist tendency in theory winds up using the method of the totalizing instance — the theoretical opposite pole of relative autonomy. Thus the use of "illusionism," of certain narrative conventions, even of optically ground lenses can be sufficient, in and of itself, to totalize a film as a "bourgeois" product. And so we return to a situation which was initially seen as the fatal snare of "vulgar determinism" — a theory so broadly general that it can detect no significant political distinction between SALT OF THE EARTH and GONE WITH THE WIND.

 The Marxist study of film has been going in circles between concepts of "revolutionary form" (which bars the dominant practice of film — the commercial feature — from consideration) and nebulous concepts like "false consciousness" (which oblige us to play at being mindreaders of the working class.) I have tried to indicate a theoretical terrain in which discussion can occur in a manner that will help us formulate what we already know, as well as produce new knowledge. I hope this discussion will bring us closer to the fundamental political questions which must underscore all Marxist theoretical work, work which aims to guide practical action.


1. Karl Marx, "Preface to The Critique of Political Economy," in Marx & Engels, Selected Works (New York: International Publishers, 1974) p. 182.

2. Karl Marx and Frederick Engels, The German Ideology (New York: International Publishers, 1970) p. 64.

3. In particular, see the letter to Joseph Bloch (Sept. 21, 1890) in Karl Marx and Frederick Engels, On Literature and Art (Moscow: Progress Publishers, 1976) pp. 57-62.

4. Louis Althusser, "Contradiction and Overdetermination," in For Marx (New York: Random House, 1969), pp. 87-128.

5. Ibid., p. 113.

6. Geoffrey Nowell-Smith, "Moving On from Metz," Jump Cut 12-13, p. 41.

7. For a good general survey of work in the field see David A. Gold et. al., "Recent Developments in Marxist Theories of the State," Monthly Review 27:5, pp. 29-43, #6, pp. 36-51.

8. This phenomenon and its theoretical consequences are discussed at length in Perry Anderson, Considerations on Western Marxism (London: New Left Books, 1976) p. 29 et. pass.

9. Available in English translation are two theoretical volumes, Political Power and Social Classes (London: New Left Books, 1973) and Classes in Contemporary Capitalism (London: New Left Books, 1975); and two volumes of concrete political-historical analysis, Fascism and Dictatorship (London: New Left Books, 1974) and The Crisis of the Dictatorships (London: New Left Books, 1976). Unfortunately, these books are all inordinately expensive, and their style is often less than fluid.

10. Karl Marx, letter to P.V. Annekov (Dec. 28, 1846) in Karl Marx, The Poverty of Philosophy (Moscow: Progress Publishers, 1955), p. 156.

11. Nicos Poulantzas, "The Capitalist State," New Left Review, No. 95 (Jan.-Feb. 1976) p. 75.

12. Louis Althusser falls prey to this error in his essay "Ideology and Ideological State Apparatuses," in Althusser, Lenin and Philosophy (New York: Monthly Review Press, 1971) pp. 127-186.

13. Political Power and Social Classes, p. 209.

14. Brian Henderson, "Toward a Non-Bourgeois Camera Style," Film Quarterly 24:2 (Winter 1970-71). Is it perhaps a proletarian camera style? A petty-bourgeois camera style? A camera style of the united front? Or, inevitably, a camera style above class.

15. Martin Walsh, "Political Formations in the Cinema of Jean-Marie Straub," Jump Cut 4, p. 13. Much of this theory (not Walsh's in particular) derives from a mangled reading of Althusser — from applying to ideological discourse the categories Althusser developed to describe scientific discourse, thereby garbling his most fundamental distinction.