11 June, 1977
Professor Richard Wolff
Mr. Antonino Callari
Mr. Bruce Robert
Gentlemen (and revered colleagues):
I have read your fine paper, “Marxian and Ricardian Economics: Fundamental Differences,” with great interest and profit. I found it enormously helpful to me, and in considerable measure persuasive as regards the significant differences between the theories of Ricardo and Marx. I am not convinced of the merits of the rather high-powered methodological assumptions (regarding 'two sciences,' etc.) through which you express your conclusions, but that is an issue that can perhaps await exploration at a later date. In this communication, I should like to focus my attention on two specific, but very fundamental, points. With regard to the first, I believe that you have gone astray, philosophically; with regard to the second, I believe that you are absolutely correct, but that your case can be made stronger, in ways that I shall suggest. I am couching these reflections in the form of a letter to you three, but I shall take the liberty of circulating them more widely to other members of our community with similar interests and concerns. I might say that availability of such a community is, for me, an experience unique in my intellectual and professional career, and a fringe benefit of incalculable value at UMass.
The two points to which 1 shall address myself are these: First, your use of the term "overdetermined," which I believe to be confused in non-trivial ways; and Second, your discussion of the fundamental differences between Marx and the Neo-Ricardians (and Ricardo himself) on the matter of the relationship between circulation and production.
I. The concept of Overdetermination
I believe that you are using the term "overdetermined" in a way that deviates both from the meaning of Freud, who introduced it into the literature, and also (perhaps) front the meaning of Althusser, who acknowledges his debt to Freud, and to whom you in turn acknowledge your debt. Now, ordinarily there is not much to be gained front terminological quibbles. Many philosophers have taken the position of the Caterpillar in Alice in Wonderland, who, when he used a word made it mean whatever he wished it to mean. Plato appears to have begun this practice, and virtually every great philosopher since has followed suit. Nevertheless, I intend to quibble about the meaning of the term, for this reason: I think Althusser, clearly or unclearly, was on to a very profound, very powerful, and highly problematical methodological insight when he described social formations or phenomena as "overdetermined." Your quite different use of the term loses that power, profundity, and methodological novelty, reducing the notion to a rather familiar one that has long been known and used in the social sciences, particularly in functionalist sociology. It is at least worth trying to recapture the original meaning, in order to see whether there is something of philosophical value in it worth preserving. (A similar fate has been suffered by Durkheim's concept of anomie as well as by the notion of ideology).
The notion of overdetermination is introduced by Freud, as you note, with appropriate
references), in order to deal with certain problems in the Interpretation of dreams.
As a result of what Freud calls condensation and displacement in the dream-work, the
symbolism or meaning of dreams becomes highly compressed.
Through processes of association, certain symbols or elements of the dream may take on several quite discrete and not naturally related meanings in addition, a certain meaning may turn up in several different elements of the dream. Although Freud on occasion offers some highly tentative physiological speculations about the mechanisms of association (including, for example, the suggestion that thoughts running along spatially contiguous nerve-pathways may thereby become associated together), he clearly concluded on the basis of his clinical observations that no useful generalizations could be made about the specific patterns of association. A patient might, for example, associate "leaving treatment" with the "leaves of a book", because at the moment when he was leafing through the book he was worrying about whether to leave treatment. In the dream, leaves falling from a tree might come to stand for both the activity of leafing through the book (which, perhaps, was a gift from someone whom he missed) and the prospect of leaving treatment. Another patient might associate the leaves of a book with taking a leave of absence from his job, something he very much wanted to do. And so on. In the dreams of these patients, the visual image of the leaves of a tree might be a a symbol with several distinct significant meanings. To say that such a symbol, or the entire dream, is overdetermined is precisely to say that the symbol has two or several complete, adequate, satisfactory explanations, each of which is itself sufficient by whatever criteria of adequacy of explanation one is employing -- to account for the content of the dream or the symbol. The symbol is thus in a quite natural sense of the word, "over-determined." That is, it is determined several times over. One can give a complete explanation of the dream in terms of one of its meanings, so that nothing is left out, no loose ends are left hanging. Then one can go back and give a quite different and equally correct explanation of the same dream.
Let us very clearly differentiate this remarkable notion of Freud's from two quite familiar notions that play a major role in social scientific explanations in sociology, economics, and other disciplines. The first is multiple causation (or, to use your terms, multiple effectivity); the second is reciprocal causality, or reciprocity. To say that an event is multiply caused, or determined by a multiplicity of causes, is simply to say that not one but a number of events, phenomena, etc, must be invoked in order to provide a complete explanation of it. Or, sometimes it is to say that phenomena or events from a number of different social spheres must be invoked, such as political causes, economic causes, cultural causes, psychological causes, institutional causes, etc. The key point here is that an explanation in terms of only one of these factors will be inadequate, incomplete, and hence demand some enlargement or supplement before it can be explained. For example, an historian attempting to explain the particular course of the French revolution might feel the need to invoke facts of the special and particular history of the French peasantry, in addition to economic facts common to the French and other economies, in order to account for the details of (or even the broad outlines of) that course. To say that there is reciprocal causality, or reciprocity, between two events or phenomena is simply to say that each influences or acts on the other, and also perhaps that each then reacts on the other in response to the action of that others. Just as multiple causality is invoked in order to rebut the claims of single-factor explanations in the social sciences, so reciprocity is invoked to counter the claims of single-direction explanations. Notoriously, the simple-minded "base-superstructure" model customarily (and, of course, incorrectly) imputed to Marx is both a single-factor and a single-direction mode of explanation. Modern functionalist sociology, and also modern economics, decisively reject both single-factor and single-direction explanatory models. That is not peculiar to Marx, and it certainly is not peculiar to Althusser's reading of Marx. Max Weber, and following him Talcott Parsons, offer explanations replete with multiple causality and reciprocity.
Perhaps the most familiar explanatory model totally embodying both multiple causality and reciprocity is the original Newtonian mechanics. Each mass in the universe, through the gravitational attraction that it exerts, has an effect on each other mass in the universe. Thus, the behavior ([i.e.] motion) of a mass is multiply determined (by the effects on it of every other mass) and that behavior stands in reciprocal interaction with the behavior of each other mass. Note, by the way, the enormous power of Newton’s claims.
He does not say that some masses are multiply determined, or that some masses are in reciprocal interaction with one another. He says that all masses are multiply determined, and that every event in the universe affects every other event, however remote. There are really only four ways one could sustain such a universal claim. Either by appeal to some theological revelation; or by appeal to some a priori metaphysical principle (such as that invoked in the opening section of Kant's Inaugural Dissertation of 1770); or by appeal to an epistemological principle such as the transcendental unity of apperception (see Kant, Critique of Pure Reason: "Deduction of the Pure Concepts of Understanding"); or by appeal to a methodological principle of the formation of theories. In any case, it is illegitimate to move from the indisputable observation that events frequently are multiply determined or in reciprocal interaction to the conclusion that every event is affected by every other event„ or that all events are in reciprocal interaction or, more precisely, it is illegitimate to make that move without an explicit and powerful argument adequate to so far-reaching a conclusion.
Let us return to Freud. If overdetermination means a multiplicity of complete and adequate explanations for a given symbol or dream, three questions immediately force themselves forward and demand answers. First, how do we know when there is more than one explanation, when, that is, the dream is indeed overdetermined? Second, how do we know which of the endlessly varied and many possible explanations for a dream or symbol are the correct one? And third, how do we know when we have found all of the correct explanations of the dream, and thus have exhausted its significance? Freud gives the same answer to all three of these questions, and it is so important that I shall dwel1 on it for some time. It is, indeed, the key to all of Freud's thought.
Briefly, the answer is that the analyst is guided by the patient's associations to the dream. As you know, the analyst asks the patient to associate freely to the dream. In this way, the unconscious content that underlies the dream is slowly revealed. In effect, the patient follows back along the pathways of association by which the mind constructed the dream. In general, Freud thought, dreams are triggered by sensory memories of events from the preceding twenty-four hours. These memories are interpreted, compressed, distorted displaced, etc, etc, by the mind, in the course of which long-repressed wishes and fantasies find a safe or "acceptable" expression in disguised, symbolic form. Dramas, like slips of the tongue and symptoms, are generally unimportant in themselves, but they are indispensable doorways into the unconscious.
Because of the purely adventitious, or accidental and entirely individual character of
associations, it is impossible either for all patients in general or even for one patient in particular to compile a dictionary, a lexicon, of dream symbols. A snake may stand for a phallus in one dream, and for a man named King (e "king snake") whom one doesn't like in another dream. The approach to the unconscious is thus necessarily dynamic rather than static. It proceeds by way of the actual associations of the patient, rather than by an a priori deduction of the possible meaning of a symbol. Through the process of association, one ascertains that a symbol is overdetermined, what the precise set of overlaid meanings are that have become attached to it, and when the unpacking of the meaning is finished and the symbol has been exhausted. Without this dynamic process of association, the notion of overdetermination would be vacuous. There would be no reason to suppose that a symbol was overdetermined, to way of knowing whether an explanation of the symbol was correct, and no way to determine whether one had identified all the meanings of the symbol or not.
There are two points to be noted about Freud's theory of overdetermination of dream embolisms First, there is no place here for what we have called reciprocal causality. There is, of course, multiple causality (assuming, for the moment, that we can treat meanings or reasons as causes -- more of that in a moment). But the dream does not in its turn react back on the repressed content, or at least not so far as Freud knew. Social theorists in general, and you in particular, wish to make strong claims for reciprocal causality (or effectivity), and to that extent, therefore, you deviate from Freud’s notion. Second, Freud speaks, most of the time, not of a multiplicity of causes or of an overdetermination of or causes, but of overdetermined meaning. Symbols in a dream have cognitive meaning they refer [to?], they have what philosophers call intensionality. One explains a dream by identifying the repressed wishes that find expression through it. As Richard Wolheim makes beautifully clear in the opening chapter of his invaluable book, Sigmund Freud's original formulation of the notion of the unconscious, from which all the rest of his work flows, depends precisely on the distinction between physical causes and meanings or intensional referring, thoughts. (The point, you will recall, was that patients suffering from hysterical paralysis exhibited limbs that were paralyzed over a physical extent that did not correspond to any natural neurological, anatomical unit, but did correspond to what the patient typically a woman -- in an unscientific way thought was such a unit. So she might have a paralyzed "leg," when the neurologists knew that the portion of her body paralyzed did not correspond to a portion controlled by a single major nerve, etc, etc.).
Having said all this, let us now turn to the sphere of social phenomena, and ask how and why someone might wish to Import the notion of overdetermination into it. I suggest the following possible explanation. As detailed historical studies begin to pile up, and in particular as the results and insights of anthropology begin to filter into the study of western European history, the following sort of problem is liable to crop up. Someone does an exhaustive study of, let us say, a series of wildcat strikes and other outbreaks of labor militancy, in which it is quite persuasively shown that the behavior of the workers is fully explained by appeal to their objective economic interests (perhaps dialectically related to the development of their awareness of those interests). Then another historian plows the same soil, and produces a study showing that the behavior of the workers is explainable in terms of certain traditions of peasant and worker behavior, going back hundreds of years, and reinforced by certain local religious traditions. Yet a third historian, drawing on diaries, letters, and other such materials, produces a psychological explanation of the very same behavior.
There are at least four different things we might want to say about such a state of affairs. First, we might say that the economic interests are the real causes of the workers' behavior, and that everything else is merely epiphenomenon, reflection, expression or effect of those economic interests. This, I take it, is the vulgar marxist position. Althusser rejects it, and so did Marx. So too does everyone else with any brains. Second, we might wish to say that the behavior of the workers was multiply determined by economic, traditional religious, cultural, psychological, and other factors. This is the position most historians automatically take. You seem to take it from time to time in your paper. The trouble with this position, in the present case, is that by hypothesis in our imaginary example, each historian has provided a complete explanation of the phenomena under investigation. Of course, if each had offered only a partial explanation, if each had left same significant features of the events unaccounted for, then we could infer that each explanation needed fleshing out or supplementing by other explanations, and that would lead naturally to a multi-causal account. Now, there are some historians (and same philosophers of science and of history), who would deny that the sort of situation I have imagined could ever really occur. There couldn't be two or more complete explanation of the same set of events. Perhaps so. But 1 am pursuing for the moment the thought that there might at least seem to be, and that we might therefore need an especially complex methodological move to deal with that odd fact. Third, we might wish to seer that in fact there was not merely a multiplicity of causes of the observed events, but a reciprocity of causality, so that none of the factors could be identified as independent variable and none as dependent variable. The religious beliefs, we might argue, were both caused by and were causes of the economic interests of the workers, etc etc etc. This too is a fairly standard notion among historians and social scientists. Indeed, I should think that it is, today, the dominant view [as you?] clearly espouse it in your paper, along with the notion of multiple causation.
Finally, someone (for example, Althusser) might wish to say something quite distinct from these first three positions, something powerful„ paradoxical, and striking. Someone might wish to say that this historical event (and, by extension, all other historical events) is overdetermined, not determined, not multiply determined, not reciprocally determined, but overdetermined. How could this be? Clearly, natural events cannot be overdetermined. But historical events are not natural events, and society is
not nature, not even second nature. If society can be understood as a structure of meanings, not of objects, and if we can make sense of the vexing notion that society is a structure of meanings which are not to be identified with the thoughts in the minds of any specific individuals (without sinking into the Durkheimian mistake of positing a group mind), then perhaps we can elaborate a notion of social overdetermination analogous to Freud’s notion of the overdetermination of dreams.
Now, I am not certain that this approach will stand up. But it seems to me highly suggestive and original, and I would hazard a guess that it is what Althusser has in mind. Two points, before I close this part of my remarks. First, Althusser is well known for rejecting the "subjective" approach to the study of society. Pretty clearly, whatever his reasons for taking that line (his fight with Sartre, who knows --), he is never going to succeed in defending the notion of overdetermination without some serious. acknowledgement of the idea that society is a system of meanings, not a set of objects. Second, we saw that Freud leaned on the method of free association as a key to the overdetermination of dreams. What, in the study of society, takes the place of free association in psychoanalysis? In what way do we discover that a social situation is overdetermined what the actual (as opposed to possible) multiplicity of determinations are, and when we have exhausted the unpacking of a social situation into its component determinations?
It may be, after all, that by overdetermined you really mean nothing more than multiply and reciprocally determined. If so, there is nothing more to be said. But if it is indeed so, then you must recognize that your methodological position is thus far indistinguishable from that of countless functionalist sociologists, anthropologists, and political scientists and also from that of standard economic theory. After all, any system that can be characterized by a set of simultaneous equations is correctly describable as both multiply and reciprocally causally interrelated.
II. Production, Circulation, and the Neo-Ricardians
The preponderance of your remarks are devoted to the differences between Marx and Ricardo (and the neo-Raicardians), with particular emphasis on the different analyses of the relationship of circulation to production. I think you are correct, both about the difference between Marx and Ricardo and about the difference between Marx and the neo-Ricardians. What is more‚ I think you are right to insist that this is a central difference. Without accepting your methodologically powerful and unsupported invocation of the idea of "two sciences," I will simply agree that this is a big enough difference to constitute a decisive difference.
However, as I read the paper, it seemed to ne that you never went beyond observing and documenting the difference. You never presented an argument designed to show that the Ricardian and neo-Ricardian theories, by slighting circulation in the way they do, produce thereby an inferior theory. Now, don't start in with "framework" and all the met of that. Don 't give up the fight so easily! I think one can offer powerful arguments to show that the Ricardian and neo-Ricardian theories are deficient, not merely different, precisely by virtue of their failure to take the proper account of the rate of circulation. Let me sketch one such argument very briefly (you will immediately notice at this point that I get in way over my head. My apologies.)
In the familiar, standard physical quantities model (to use Stedman’s term for it), we are presented with a system of n equations in n2 unknowns, namely the n [?]
prices, the wage rate, and the profit rate. One of the n prices is arbitrarily chosen as numeraire, or standard of value, and set equal to unity. The system then has one degree of freedom, and we can study the inverse relationship between the wage rate and the profit rate. This, I take it, is what Sraffa does.
Now, the attention has been focused on the so-called transformation problem, which it turns out, when cast in the proper mathematical form, to have a relatively straight-forward solution. (Needless to say, I did not find it easy or straightforward, but I feel 1 must maintain the standards and conventions of this new field into which I have wandered, and people like Pasinetti, Steedman, Morishima, etc seem to consider the theory of eigenvalues and eigenvectors to be about on a par, conceptually, with baby talk.) But to me, the real heavy freight rests on that barely noted preliminary move whereby one of the n prices is arbitrarily chosen as numeraire and set equal to unity. To proceed in that way is, in effect, to say that there is no real money in the system under analysis. There is an accounting system, but no one of the commodities over separates itself off from the others„ becomes functionally divorced from whatever use-value it might originally have had (as gold and silver did), and thereby becomes, in the full sense of the term, money.
It seems to me that in Capital, and even more clearly in the first three or four hundred pages of the Grundrisse, Marx is insisting that the emergence of money, and subsequently the transformation of money into capital, is neither trivial nor a merely formal and stipulative occurrence. The emergence of money capital requires enormous historical changes, of course, to which are conjoined major psychological and conceptual changes. But in addition (their being merely in the category of background), the emergence of money requires essentially a fully-developed sphere of circulation, without which there could be no capital, since any formal model of a capitalist economy that treats the sphere of circulation as a mere accounting-world, a place of relative prices and hence merely of highly complex barter, must be inadequate to the reality of capitalism.
If I am correct, then a neo-Ricardian model would be wrong, not just "in a different framework or a different science." How would it be wrong? Well, it would certainly fall to explain the emergence or money capital as such; it would presumably be unable to give a coherent account of realization crises; it would be unable to explain why problems of accumulation in different sectors of the economy would lead to unequal development; and it would simply have nothing to say about such phenomena as the fetishism of commodities. What is more, the neo-Ricardian model seems more appropriate to a planned economy than to an actually functioning capitalist economy. Its claim to serve as a model for a capitalist economy would thus constitute an implicit assertion that a planned economy is simply a more perfect, more rational form of a capitalist economy, a claim that seems manifestly false and also rather heavily laden with powerful political implications.
Finally, let me simply report that I could not follow your mathematics in the last several pages. It seemed to me, although I was unable to determine whether I was correct that a dimensional analysis of the equations on page 66 would reveal that the terms on the left hand side are of different dimensions, or units, from the terms on the right hand side.
Well, it should be obvious that your paper stimulated me to a considerable response, whether fruitful or not I leave it to you to decide. Many thanks for letting me read it.
All the best,
Robert Paul Wolff