1. The Concept of a Mode of Production

Structural Marxism develops the principle of economic determination by means of the concept of a mode of production, the latter serving as both the principle of "articulation" of a given conjuncture, that is, the structure that determines the interrelationships between the various instances within a social formation, and as the principle of "periodization," that is, the concept on which the historical succession of conjunctures is organized. "Marx's construction of the central concept of the mode of production," Etienne Balibar insists, "has the function of an epistemological break with respect to the whole tradition of the philosophy of history. . . . The concept of the mode of production and the concepts immediately related to it thus appear as the first abstract concepts whose validity is not as such limited to a given period or type of society, but one on which, on the contrary, the concrete knowledge of this period and type depends" (Althusser and Balibar 1970, 201). In his important contribution to Reading Capital , "The Basic Concepts of Historical Materialism," Balibar elaborates on Marx's central insight, that "it is always the direct relationship of the owners of the conditions of production to the direct producers . . . which reveals the innermost secret, the hidden basis of the entire social structure" (Marx, quoted in Althusser and Balibar 1970, 221).

Balibar's goal, and the goal of the various anthropologists, historians, and sociologists whose work we will take up in this chapter, is a

concept of a mode of production applicable to different degrees of generality and levels of abstraction—from social formations "in general" (in the sense that all social formations contain a determinant economic structure), through societies dominated by distinct modes of production (lineage, "Asiatic," simple property, slave, feudal, capitalist, and so on), to historically specific conjunctures (France in 1789, England during the Industrial Revolution, and so on). The problem is not primarily one of classification; Structural Marxists are uninterested in developing a finite number of pigeonholes into which empirical descriptions of all known societies may be forced, nor are they interested in arranging modes of production into an evolutionary series of historical stages mechanically following one after the other in accordance with some ultimate goal or originary essence. The problem addressed by the concept of a mode of production is one of producing knowledge rather than classifying data. One looks at societies from the standpoint of a mode of production and from a certain level of abstraction, but this is not to say that additional specificity, that is, knowledge at a more historically determinate level, is produced deductively or reductively from general concepts. In relation to the potential dangers of deduction and reduction, it must be emphasized that the elaboration of distinct modes of production is an open-ended, heuristic process. Against the opinions of most of its critics, I would argue that Structural Marxism is especially sensitive to the tension, within scientific discourse, between the "proliferation" of descriptive accounts of historically determinate modes of production and the "reduction" of historically determinate modes of production to a restricted number of general forms.^[1]

Etienne Balibar: The Forces and Relations of Production

Taking Marx's discussion of the social form of production in volume 2 of Capital as his point of departure, Balibar defines a mode of production, at its highest level of generality, simply as a manner of producing a socially useful product. All modes of production, he maintains, are constituted by certain functional elements which are formally invariant but which in fact exist only in their "combination," that is, in their historically specific content and interrelationship. These invariant functional elements are (1) the laborer or direct producer, that is, labor power; (2) the means of production, that is, the object and the means

of labor; (3) the nonlaborer who appropriates surplus labor, that is, the social product (see Althusser and Balibar 1970, 212).

These three elements are always structured as a double relation along two axes or "connections" whose specific combination constitutes, for Balibar, the historical uniqueness of a mode of production: (1) A relation of real appropriation designates the structure of the labor process, that is, the relation of the laborer to the means of production by which the transformation of nature is undertaken. This relation constitutes the "technical division of labor" or the forces of production . (2) A property relation designates the mode of appropriation of the social product. This relation, the "social division of labor" or relations of production , implies the intervention of an individual or a collectivity, who, by the exercise of economic ownership, controls access to the means of production and the reproduction of the productive forces. Thus, within a given mode, it is the relations of production that are dominant. (See Althusser and Balibar 1970, 212-13.)

Unlike Marx's concept of the capitalist mode of production, Balibar's more general concept presupposes neither nonlaborers beyond infants, the elderly, and the infirm nor the existence of surplus labor beyond that minimum surplus necessary to reproduce the means of subsistence. The same individual can be both laborer and owner, the property relation may have a greater or lesser impact on the labor process, and the ownership function may be institutionally united or separate from the labor process. Moreover, the function of economic ownership should not be identified with any particular form of social control over labor and the means of production, no more than Balibar's use of the term property relation should be associated with any particular form of individual or collective ownership. It is also important to note that the terms property and ownership apply strictly to the realm of economic relations and should not be confused with juridical forms ("laws of property"), which are not, strictly speaking, expressions of production relations but rather of political and ideological relations not necessarily economic in nature. Property relations can take extremely varied forms, but all property relations imply social control over access to the means of production and the distribution of the social product. The necessary existence of a property relation therefore implies the possibility of economic classes and exploitation, although the universality of classes and exploitation remains controversial. For example, in so-called primitive societies operating at a subsistence level within an economy structured by kinship, the existence of exploitation in any form

(by men over women, by elders over juniors), much less class exploitation, is hotly contested even within the camp of Structural Marxist anthropology.

In the case of lineage-based societies, one must distinguish irregular appropriations of the social product such as raiding, which are merely forms of extortion external to a social formation, from institutionalized appropriation, which internally organizes the production of material existence within a social formation. Only the latter may be termed a mode of production in the strict sense. One must also distinguish a cooperative division of labor characterized by reciprocal exchanges of labor or product determined by necessity or utility, from an exploitive division of labor wherein labor power and economic surplus are appropriated by certain social classes without corresponding obligations or limitations. Of course this distinction is often difficult to specify in practice, not least because dominant classes always develop an ideological justification for their power and privileges couched in terms of historical necessity, social utility, or both. Balibar's formulation is particularly useful because it focuses analytical attention on the structural nexus of class exploitation, namely, access to the means of production and distribution of the social product, rather than on the ideological structures through which social subjects explain, justify, resist, and adapt to the forces and relations of production. If and when unequal distribution of the social product within a social formation serves to reproduce the conditions of that unequal distribution, then we are in the presence of class struggle and exploitation. It is sufficient—indeed, necessary—to speak of exploitation whenever direct producers are unable to determine either the volume or the appropriation of their own surplus labor.

According to Balibar, the forces and relations of production exist only in their "unity" as "simultaneous yet distinct" relations among laborers, the means of production, and nonlaborers. Not only is each of these elements related to the others, but it is also related to them in two different dimensions: "This double function is an index of what I shall call the double nature of the division of labor in production (the 'technical' division of labor and the 'social' division of labor); at the same time, it is an index of the interdependence or intersection of these two divisions . . . the fact that the two connections . . . both belong to a single 'Verbindung ' . . . to the structure of a single mode of production" (Althusser and Balibar 1970, 214). By emphasizing the unity of the forces and relations of production, Balibar neatly resolves a longstanding Marxist debate as to whether technological development or class struggle is the motive force of history. For Balibar, such a debate

turns on a misleading identification of the productive forces with technological determinism and the property relation with class struggle.^[2] In actuality, because technological development never exists apart from class struggle, nor class struggle apart from a specific level and organization of technology, any attempt to separate them is meaningless.^[3] Only between rival modes of production (that is, rival unities of forces and relations) is the level of technological development decisive.

While Balibar insists that the relations of production are dominant within a given mode of production (because the ownership function controls the reproduction of the productive forces), this relation of dominance is neither reducible to a voluntaristic concept of class struggle nor open to criticism for having ignored the level of technological development. The dominance of the property relation implies the dominance of the class exercising the ownership function, but this is not to say that the dominant class can do anything it wishes nor that its dominance results from ownership alone. For Balibar, it is not a question of the "primacy" of the relations or the forces, in the sense of relations being somehow in advance of the forces; rather, it is a question of class struggle within both the forces and the relations. The class struggle is determined by the unity of a mode of production: the level of economic development defines the universe of class interests and powers, but economic development is nothing but the manifestation of the interests and powers of social classes. The primary contradiction between the forces and relations of production does not turn on the "primacy" of one or the other—economism or voluntarism—but rather on the antagonistic class interests and powers produced by both. Ultimately, class struggles within a given mode of production as well as those between classes of rival modes are settled in favor of the class most capable of developing the productive forces at a given time.

In both capitalist and non-capitalist societies the combination of economic ownership (relations of production) and a labor process (forces of production) specifies the determinant economic function or instance of the social formation. Defining a mode of production in this way differentiates the general functional term economic from historically specific economic terms such as market economies, private property, and so on, which have been frequently employed with the unfortunate result of reading capitalist characteristics into other modes of production.^[4] The Structural Marxist concept of a mode of production attempts to reveal the historical specificity of economic determination within different social formations, not to eliminate or distort it. The economic instance in feudal society, for example, is determinant in the last in-

stance even though the relations of property and the labor processes constituting the feudal mode of production are defined by neither market values nor private property (even where market mechanisms and private property exist). The subordination (subsumption in Balibar's terminology) of the social and technical divisions of labor to market mechanisms and private property is the characteristic feature of a capitalist mode of production, not a universal or necessary form of economic determination in all social formations. Economic ownership and the labor processes may reside in different social structures and institutional apparatuses, and within these structures and apparatuses they may coexist with a variety of non-economic functions. As we have seen, this is the basis of Althusser's distinction between the concepts of dominance and determination: the term structure in dominance always denotes the structure exercising the function of economic ownership (whether or not this structure is the site of actual production), whereas the term determination in the last instance always denotes the primacy of the economic function, that is, the deep structure of the mode of production (whether or not the structures that exercise the functions of ownership and production are primarily political or ideological).

Balibar argues that the juxtaposition of three elements and two relations gives a high degree of theoretical precision to the concept of a mode of production: "By varying the combination of these elements according to the two connections which are part of the structure of every mode of production, we can reconstitute the various modes of production, i.e., we can set out the 'presuppositions' for the theoretical knowledge of them, which are quite simply the concepts of the conditions of their historical existence. . . . The final result would be a comparative table of the forms of different modes of production which all combine the same 'factors' " (Althusser and Balibar 1970, 216). However, Balibar also insists that the "invariant formal characteristics" of his general concept not be mistaken for a denial of the historically specific character of every concrete mode of production. Balibar is especially concerned that the development of concepts be appropriate to distinct determinate levels of generality and specificity and not restricted to any single level. He cites approvingly Marx's observation that the general concept of the capitalist mode of production, which deals with properties common to every capitalist system, "does not prevent the same economic basis—the same form from the standpoint of its main conditions—due to innumerable different empirical circumstances, natural environment, racial relations, external historical influ-

ences, etc. from showing infinite variations and gradations in appearance, which can be ascertained only by analysis of the empirically given circumstances " (Marx, quoted in Althusser and Balibar 1970, 256).

In defending the need for concepts pitched at different levels of generality as well as the validity of such concepts, Balibar rejects the extremes of empiricist nominalism and essential forms. Denial of the historical specificity of its constituent elements, Balibar maintains, would reduce the concept of a mode of production to a Structuralist "combinatory" (combinatoire ) as opposed to a Structural Marxist "combination." For Balibar, the former term refers to the tendency of many Structuralist theorists to itemize elements of a social system as a formal pattern of relations and arbitrarily occupied places that appear and reappear throughout "history" in such a way that, while the places of the elements and their relations may change, their nature remains constant, indifferent to the effect of changes within the structure as a whole. As a result, by means of a combinatory, it is possible to bypass the historically specific elements and proceed directly to an a priori knowledge of all possible articulations of the structure. A combination, by contrast, insists on the fact that the nature of the elements themselves is altered by their historically specific structure. With a combination, Balibar insists, "we do not find the same concrete elements when we move from one variant to the next. Nor is their particularity defined by mere place, but rather as an effect of the structure, differing every time; i.e., an effect of the combination which constitutes the mode of production" (Althusser and Balibar 1970, 241).

Balibar's contrast between the Structural Marxist concept of combination and the Structuralist combinatory is not entirely satisfactory.^[5] Although certain Structuralists, Lévi-Strauss for example, may be justly criticized for imposing a combinatory consisting of ahistorical elements on each and every social formation, many others are willing to concede, even to insist on, the fact that individual elements derive their character from the system within which they exist: this is, after all, a fundamental property of the linguistic sign according to Saussure and a basic axiom of semiology. Balibar's combinatory/combination opposition therefore misses the central weakness of Structuralism, which is not its inability to see the individual element as determined by the structure of the whole but rather the fact that it has no theory of the transformation of the structure, no concept of contradiction, and no concept of dominance and subordination. For Structuralism, it is certainly conceivable for elements to be seen in interaction with each other, but this interac-

tion, like that of functionalist sociology, is static; the system adjusts but it never changes, its movement circumscribed by invariable relationships or at least relationships whose variation cannot be theoretically accounted for by the structure itself.

The weakness of Balibar's attempt to distinguish the Structural Marxist concept of a combination from a Structuralist combinatory stems from the fact that his own analysis of the problems of periodization and articulation is deeply flawed by a rationalist and essentialist attempt to theorize both problems in terms of the Structuralist opposition of synchrony and diachrony instead of the Structural Marxist concept of contradiction. Therefore, before moving on to the important positive advances realized by Balibar's concept of a mode of production and the subsequent deployment of this concept by anthropologists and historians, it is useful to examine carefully Balibar's attempt to work out the implications of the Althusserian problematic with respect to the problem of historical process. Readers less patient with theoretical detours may wish to skip the next two sections.

Reproduction and the Problem of Periodization

For Structural Marxism, modes of production are the principles of variation by which the unbroken continuity of real history is analytically differentiated into periods or, more accurately, into a discontinuous succession of structures. Because variation, not continuity, defines change, our understanding of historical transformation is always differential, derived from a comparison of different unevenly developed conjunctures, not from the reconstruction of the evolutionary trajectory of an homogeneous essence. In the language of differential history, there is always a gap or dislocation between different conjunctures, as, for example, between the social formations of France in 1788 and 1815. We attempt to explain the gap between France in 1788 and 1815 by means of successive analyses of intermediate conjunctures—1789, 1792, 1795, 1799, and so on—and each successive conjuncture in terms of its own structural integrity. At the same time, we are also impelled by the dialectic of differential analysis to construct concepts of structural integrity across different conjunctures—for instance, the concept of the French Revolution as the period 1788 to 1815, which in turn exists in a differential relation to other periods, the ancien régime, the Restoration, and so on. Knowledge of historical transformation, for Structural

Marxism at any rate, is always a matter of an irreducible tension between the discontinuity of a succession of structures on the one hand and the structural integrity between two conjunctures on the other. Social formations, we must not forget, are unevenly developed structures; therefore, our knowledge of their development can never be reduced to the evolution or realization of an expressive totality. However, they are also structured wholes, and therefore their emergence and dissolution cannot be explained simply by constructing genealogies of their individual elements.

In Reading Capital , Balibar makes a bold initial attempt to elaborate this differential approach to historical development, without, unfortunately, managing to avoid certain pitfalls (whose notoriety has, unfortunately, precluded critical recognition of the nature of his project). Taking as his example Marx's analysis of the "primitive accumulation of capital," Balibar notes that there is a "different world" at the origin of capital where "knowledge of the laws of the development of capitalism is useless . . . because this is a completely different process, not subject to the same conditions" (Althusser and Balibar 1970, 278). The study of primitive accumulation

takes as its guiding thread precisely the elements which were distinguished by the analysis of the capitalist structure: these elements are grouped together here under the heading of the "radical separation of the laborer from the means of production." The analysis is therefore retrospective, not insofar as it projects backwards the capitalist structure itself, presupposing precisely what had to be explained, but insofar as it depends on knowledge of the result of the movement. On this condition it escapes empiricism, the listing of the events which merely precede the development of capitalism: it escapes vulgar description by starting from the connections essential to a structure, but this structure is the "current" [capitalist] structure. . . . The analysis of primitive accumulation is therefore, strictly speaking, merely the genealogy of the elements which constitute the structure of the capitalist mode of production . . . . For this reason, the analysis of primitive accumulation is a fragmentary analysis: the genealogy is not traced on the basis of a global result, but distributively element by element. (Althusser and Balibar 1970, 279)

"In Marx's analyses," Balibar notes, "we are never dealing with anything other than the combination itself and its forms. . . . [T]he subject of development is nothing but what is defined by the succession of the forms of organization of labor and the displacements that it achieves" (Althusser and Balibar 1970, 247). Thus analysis of primitive accumulation does not and cannot coincide with the history of the previous

mode of production as known from its structure, but it does present us with the question of this structure, its tendencies and contradictions, which is precisely the question of a gap or displacement between two successive structured wholes or combinations. We cannot bridge the gap by means of the continuity of an evolution, because such a transition is thinkable only at the level of elements, not the level of structures. For Balibar, we can think a true history only on the basis of the mutual dependence of elements with respect to a structure, that is, as a structured combination, but similar elements inscribed within different social-historical combinations will be differentiated precisely by these distinct structural causalities. One important consequence of this theoretical impasse is the rejection of any evolutionary necessity linking the concepts of different modes of production—that the final triumph of capitalism, for example, is pre-ordained by the concept of feudalism. In opposition to such forms of deductive necessity, Balibar asserts the "relative independence of the formation of the different elements of the capitalist structure, and the diversity of the historical roads to this formation"—in other words, the fact that "the elements combined by the capitalist structure have different and independent origins" (Althusser and Balibar 1970, 280).

Unfortunately, Balibar's insights into the differential nature of our concepts of historical development are presented in terms of a Structuralist opposition between synchrony and diachrony, an opposition that ultimately makes it impossible to preserve the primacy of uneven development (contradiction) within the framework of structural causality. Balibar is driven by the logic of the synchrony/diachrony distinction toward an essentialism of the concept of a mode of production; he progressively transforms contradictions within a mode of production (the uneven development of the forces and relations of production) into "tendencies" of a mode of production conceptualized as a unified, evenly developed synchronic system. This essentialism collapses the relative autonomy of the elements into a monolithic and homogeneous totality. It is accompanied by a tendency toward philosophical rationalism with respect to the articulation of the elements; Balibar progressively eliminates the gaps between different levels of analysis (the whole and each of the elements) in order to derive the specific effectivity of each element from the totality viewed as a logical universal, the absolute ground of all contingent attributes.

Balibar's problems begin with his attempt to develop a scientific concept of periodization based on the concept of reproduction . His central

concern is to avoid the pitfalls of "simple" history, which has always conceptualized historical process and periodization within a problematic of continuity and linear time. For Balibar, it is not a question of finding the "right breaks" or "best periods," those artificial (but not arbitrary) cuts into linear time that traditionally emphasize either the primacy of the historical event (a "single criterion of brevity") or distinctions between the "long term" and the "short term" (a distinction based on the "insertion of the latter into the movement of the former"); it is rather a question of finding a "principle of variation" that will constitute history as a comparative science of discontinuous combinations determined by modes of production. Balibar finds this principle of variation in Marx's analysis of the extended reproduction of capital: "reproduction appears to be the general form of permanence of the general conditions of production, which in the last analysis englobe the whole social structure" (Althusser and Balibar 1970, 259). Balibar interprets Marx's method as one of englobing the isolated and seemingly contradictory activities within a capitalist mode of production within a synchronic system capable of reproducing its own conditions of existence: "To move from the isolated act, from the immediate production process, to the repetition , to the ensemble of social capital, to the result of the production process, is to install oneself in a fictive contemporaneity of all the movements, or, to put it more accurately . . . in a fictive planar space , in which all the movements have been suppressed, in which all the moments of the production process appear in projection side by side with their connections of dependence" (Althusser and Balibar 1970, 264). Reproduction is thus the synchronic essence of articulation circumscribing all the contradictions of capitalist circulation: "In a single movement reproduction replaces and transforms the things, but retains the relations indefinitely. . . . The relations . . . comprehend the hitherto disjoined 'moments' (production, circulation, distribution, consumption) in a necessary and complete unity" (Althusser and Balibar 1970, 268-69).

Balibar's intentions are legitimate enough. He wants only to define the criteria by which to distinguish "successive" combinations, and he argues, with considerable justification, that this definition should be done by differentiating between different principles of structural integrity or modes of production. However, by attempting to conceptualize structural integrity in terms of reproduction, Balibar lapses into a form of essentialist thinking fundamentally at odds with Althusser's concept of structural causality. Structural causality conceives the structural in-

tegrity of the complex whole in terms of its elements, or, more precisely, in terms of the place and function assigned to each element by its conditions of existence, the complex whole. Thus there is a certain tendency to reproduce the structural integrity of the complex whole since each element bears within itself the trace of its conditions of existence. However, there is a countervailing tendency—contradiction or uneven development—stemming from the relative autonomy each element possesses by virtue of its independent nature and structural integrity, which cannot be reduced to the place and function assigned to it by the complex whole.

There is, in short, an irreducible tension between the reproduction of the existing complex whole, a tendency imprinted in the structured interrelationships of the elements, and the production of new relations inherent in their relative autonomy and uneven development. For Althusser, history is a necessarily limited and incomplete science. Structural causality provides knowledge of the relation of forces at a given conjuncture, but it can have no concept of the necessity of any particular outcome. Possessing no teleology, structural causality cannot grasp change as an essence or the essence of change. Change can be grasped only comparatively, in terms of greater and lesser degrees of discontinuity between two conjunctures. Balibar, of course, is not unaware of all of this. Indeed, he attempts to develop a non-teleological approach to the problem of periodization with his concepts of correspondence and non-correspondence between the forces and relations of production. However, as we shall see, he makes a fatal mistake by inscribing these relations within a false either/or of synchrony (reproduction) versus diachrony (production). Predictably, once the distinction between synchrony and diachrony is introduced, it raises the pseudo-problem of their reconciliation, and it is this pseudo-problem that sends Balibar down the slippery slope of rationalism and essentialism.

Balibar begins by defining periodization as a "non-linear diachrony" that "replaces historical continuity with a discontinuity, a succession of temporarily invariant states of structure" (Althusser and Balibar 1970, 204). Balibar attempts to explain what a "temporarily invariant" structure might be by introducing the concept of a "transitional mode of production," taking as his example handicraft or labor-intensive manufacture. Balibar argues that manufacture, as a transitional structure between feudal and capitalist modes of production, cannot be understood under categories of either feudal or capitalist modes of production, that it is itself "a completely different mode of production" (Al-

thusser and Balibar 1970, 278). This must be so, he contends, since the period of transition between feudalism and capitalism cannot very well be conceived as a hiatus within which a social formation ceases to exist as a structured whole. Because all social formations must be structured, this transitional period must be conceived as a mode of production in its own right. To account for the difference between a "synchronic" mode of production, such as feudalism or capitalism, and a "diachronic" or transitional mode of production, such as manufacture, Balibar next introduces the concepts of "correspondence" and "non-correspondence" between the relations of production and the forces of production. When in correspondence, the forces and relations of production are in a relationship of "reciprocal limitation" such that the social formation reproduces both relations essentially unchanged. In the case of non-correspondence, by contrast, the reproduction of the relations of production induces a progressive transformation of the productive forces and, eventually, a displacement of the instances within the social formation.

Setting aside Balibar's claim regarding the primacy of the relations of production (a problem to which I will return later), the concepts of correspondence and non-correspondence appear to be expressions of uneven development within the mode of production itself. The problem, however, is that Balibar can never really integrate uneven development into his notion of a transitional mode of production because he has already committed himself to conceptualizing modes of production in synchronic terms of reproduction. Assuming for the moment that Balibar might be able to distinguish clearly between synchronic modes of production and diachronic transitional modes of production, it is still necessary to explain the shift from correspondence to non-correspondence. Barry Hindess and Paul Hirst, whose argument I am following here, point out the impossibility of reconciling the synchronic "eternity" of Balibar's idea of structural determination (the functioning of the structure to reproduce the conditions of its existence) with its diachronic "finitude" (the production of the dissolution of these conditions as structural effects): "any theory of transition from one mode of production to another requires a concept of correlative movement from non-transition to transition, from eternity to finitude. It is precisely this movement that is unthinkable in Balibar's problematic" (Hindess and Hirst 1975, 274-75).

Given that a shift to a transitional mode of production has somehow taken place, it is immediately subjected to inexorable synchronic ten-

dencies because of Balibar's insistence that any mode of production, including a transitional one, is a unified system reproducing its own conditions of existence and therefore incapable of internal dissolution: "the transition from one mode of production to another . . . cannot consist of the transformation of the structure by its functioning itself, i.e., of any transition of quantity into quality" (Althusser and Balibar 1970, 274). By positing the transitional period as a distinct social formation with a distinct mode of production, Balibar has emptied non-correspondence of its antagonistic quality until it has become its opposite, a static process of reproduction devoid of internal contradiction and transformation: "what we have recognized as distinct in essence [the transitional mode of production] shall not become a single process [a stage in the evolution of the succeeding mode of production]. . . . [T]he concept of transition (from one mode of production to another) can never be the transition of the concept (to one other-than-itself by internal differentiation)" (Althusser and Balibar 1970, 274).

Balibar has completely abandoned differential history for philosophical rationalism—a problematic wherein uneven development cannot exist except as an illusion of "inadequate" knowledge. Grasped "adequately," that is, from the standpoint of the now thoroughly essentialist totality of a mode of production, contradictions are simply "tendencies," internal temporalities and rhythms of a synchronic structure. For Balibar, "the development of the structure according to a tendency, i.e., a law which does not only (mechanically) include the production of effects, but also the production of effects according to a specific rhythm, therefore means that the definition of the specific internal temporality of the structure is part of the analysis of the structure itself. . . . It is now clear what is 'contradictory' about tendency. . . . Marx defines the terms between which there is a contradiction as the contradictory effects of a single cause " (Althusser and Balibar 1970, 288). Contradiction is simply the "unity of two contradictory terms" grounded in the "nature of the structure . . . as a law of production of the effects themselves" (Althusser and Balibar 1970, 289). Such a definition implies the subordination of contradiction to structure: "there is only a contradiction between the effects, the cause (the structure) is not divided against itself, it cannot be analyzed in antagonistic terms. Contradiction is therefore not original, but derivative" (Althusser and Balibar 1970, 290). Because the cause that produces the contradiction is not itself contradictory, the result of the contradiction is always a certain equilibrium: "contradiction has a status analogous to that of competition in

the movement of the structure: it determines neither its tendency nor its limits, rather it is a local, derivative phenomenon, whose effects are pre-determined in the structure itself. . . . [T]he only result of the contradiction, which is completely immanent to the economic structure, does not tend toward the supersession of the contradiction, but to the perpetuation of its conditions" (Althusser and Balibar 1970, 290-91).

Balibar's obsession with reproduction results in a systematic expansion of synchrony at the expense of diachrony. If contradictions within a mode of production are a function of synchronic processes, they cannot appear as determinations in the separate field of diachronic process. Therefore, in order to break this "correspondence," another structure is needed, one whose delimitation is "absolutely absent" from the existing mode of production. The movement implied by the term non-correspondence must be sought elsewhere, in the "system of interventions . . . of one practice in another" whose result "is to transform and fix the limits of the mode of production" (Althusser and Balibar 1970, 306). This move simply throws the entire problem onto the level of the social formation, where dislocations between different practices and modes of production become, in their turn, mere tendencies of this higher synchronic structure. In the end Balibar embraces this outcome, as if the problem of diachrony had never actually existed in the first place: "the problems of diachrony . . . must be thought within the problematic of a theoretical 'synchrony': the problems of the transition and of the forms of transition from one mode of production to another are problems of a more general synchrony than that of the mode of production itself, englobing several systems and their relations" (Althusser and Balibar 1970, 307).

Balibar's attempt to reconcile synchrony and diachrony by means of philosophical rationalism finally ends in a crude essentialist caricature of Spinoza's substance. Indeed, Balibar's analysis is a most graphic example of what Althusser calls "theoreticism." Balibar has since criticized the Spinozist conception of tendency he put forward in Reading Capital :

Behind the argument there is an old philosophical representation, and it is no accident that throughout this work I was guided approximately by certain reminiscences of Spinozist formulae. There is the idea that identity with itself, persistence (including in the form of the persistences of relations implied in a cyclical process), needs no explanation since it explains itself by itself , needs no cause (or production) since it is its own cause . Only "change" as "real" change, i.e., abolition-transformation of the essence,

could need a cause and an explanation. Let me say that this is a survival of the "principle of inertia," of substance and the ontology argument. (Balibar 1973, 65)

But if the problem of reconciling synchrony and diachrony produced a rationalist and essentialist response, it was the initial deployment of the Structuralist opposition of synchrony and diachrony that created the problem in the first place. The final triumph of synchrony—the common ground of both Spinoza and Structuralism—resulted from an essentialist attempt to privilege the reproduction of the whole over the uneven development of the elements. Structural causality—which views the complex whole as an "absent cause" manifested only as the simultaneous effectivities of its elements, and each element as possessing a nature and effectivity that cannot be deduced from the place and function assigned to it by the complex whole—cannot permit such an essentialist reduction. Althusser acknowledges the mistake in Essays in Self-Criticism , where he insists on a much more rigorous distinction between Structural Marxism and Structuralism than the one put forward in Reading Capital : "Marxism is not a structuralism, not because it affirms the primacy of the process over the structure (although formally this is not false) but because it affirms the primacy of contradiction over the process" (Althusser 1976, 130).

Furthermore, Structural Marxism rejects the identity of concepts and things and therefore the rationalist notion that social structures are of the same nature and obey the same logic as the concepts by which we have knowledge of them. Social contradictions, therefore, are not a function of the differential dislocations and epistemological limits within scientific discourse; rather, social contradictions are the concept of something objectively real and outside discourse, namely, the social practices of human beings who have become class subjects by virtue of the class and class-biased structures that constitute their habitus. The primacy of the mode of production within the field of social structures is manifested as the primacy of class struggle within the field of social practices, yet neither the result of economic determination nor the outcome of class struggle is guaranteed by the concept of their origin. Balibar makes this realist position clear in Cinq études du matérialisme historique :^[6]

There is only a real historical dialectic with the transformation of the concrete social formation, a process which implies the real interdependence of the different social practices (by giving them the form of being internally overdetermined by the processes of transformation of each social practice).

However, social formations are not simply the "concrete" place (or the milieu) in which an abstract, general dialectic realizes itself. Rather social formations are themselves transformed and are self-transforming because they themselves consist of the history of class struggle. This point is decisive. (Balibar 1974, 229)

Hindess and Hirst and the "Post-Althusserian" Negation of History

If Balibar may be said to have taken the concept of a mode of production in an essentialist and rationalist direction, British sociologists Barry Hindess and Paul Hirst have used Balibar's error as a pretext for pushing Structural Marxism beyond rationalism into a realm of "discourse" where no correspondence between concepts and reality, no matter how attenuated, can be (or need be) maintained, and beyond essentialism to an absolute rejection of all concepts of structural determination. In successive elaborations of their position, beginning with Pre-Capitalist Modes of Production (1975), followed by the auto-critique Mode of Production and Social Formation (1977), and culminating in a weighty two-volume work (coauthored with Antony Cutler and Athar Hussain), Marx's Capital and Capitalism Today (1977, 1978), Hindess and Hirst have come to defend a self-styled "post-Althusserian" post-Marxism based on the absolute autonomy of discourse from all material processes and historical conditions and a radical "anti-epistemology" that rejects any claim as to the reality of objects. Hindess and Hirst take an extreme anti-realist and anti-materialist philosophical position. They argue that all forms of realism are ultimately either empiricist or rationalist and, furthermore, always tautological. Hindess and Hirst proceed to reject not only the project of any epistemological guarantee for scientific discourse, a position taken by Althusser himself, but also Althusser's materialist thesis of the primacy of the real over thought about the real, a position that defends realism while acknowledging the fact that it can never be proven (or disproven) apodictically. For Hindess and Hirst, the failure of epistemology implies nothing less than the collapse of realism and materialism: "There is no question here of whether objects of discourse exist independently of the discourses which specify them. Objects of discourse do not exist at all in that sense: they are constituted in and through the discourse which refers to them" (Cutler, Hindess, Hirst, and Hussain 1977, 216-17).

According to Hindess and Hirst, Althusser's project for a differential

history is problematic because, after all is said and done, it remains a history—that is, a discourse premised on the existence of an intransitive world that circumscribes the play of discourse:

Althusser fails to break with the notion of history at the very moment of splitting from it. . . . Althusser does not say that there is no real object "history," that the notion of a real concrete history is an illusion. . . . [Although] he differentiates the thought object from the real object . . . the distinction of the objects [continues to pose] the question of the mode of their correspondence, for both are held to exist, the concrete existing prior to and independently of thought, and thought being the form in which the concrete is known. . . . The continued and uncriticized existence of the real object allows a shadow "history" to emerge parallel to the theoretical history, a shadow which reproduces the outlines of the history which Althusser has criticized. The very notion of a real object, history, the object theory appropriates, is an index of this reprise. (Hindess and Hirst 1975, 318)

Althusser insists on the distinction of the theoretical from the real object and rejects any philosophical guarantee of their correspondence, yet he does admit the necessity of historical facts (while denying their self-evident nature) and the legitimacy of historical determination as a theoretical problem (while maintaining that history has no subject or goal). These positions, scandalous as they are to many historians, do not go far enough for Hindess and Hirst, who categorically "reject the notion of history as a coherent and worthwhile object of study" in favor of a hermetic, strictly internal analysis of the logical preconditions and consequences of concepts and conceptual discourse (Hindess and Hirst 1975, 321). In their view, the acceptance of historical facts, however qualified, condemns history to empiricism since such facts introduce into the theoretical object determinations that are, properly speaking, epiphenomena. "Far from working on the past , the ostensible object of history, historical knowledge works on a body of texts . . . this or that body of representations with the status of a record. . . . The writing of history is the production of texts which interpret these texts" (Hindess and Hirst 1975, 311).

The existence of historical facts makes history a "potentially infinite text" in contrast to what Hindess and Hirst call the "analysis of the current situation," which may be "rigorously conceptualized" on the basis of "the conditions of existence of present social relations" (Hindess and Hirst 1975, 312). The idea that the current conjuncture can be more rigorously conceptualized than the past (that it is somehow really there) would seem to be untenable within the terms of their own

anti-epistemology, and Hindess and Hirst never present any explicit arguments in defense of the epistemologically privileged status of the contemporary over the historical. At bottom the logic of their position is political, not theoretical. In Pre-Capitalist Modes of Production , Hindess and Hirst reject Balibar's rationalism in favor of what Gregory Elliott (1986) has accurately, if polemically, described as "Maoist theoreticism."^[7] Balibar's concept of a mode of production (an articulated combination of relations and forces of production requiring the securing and reproduction of certain conditions of existence) is still acceptable, and following the lines of Balibar's own self-criticism, Hindess and Hirst insist that class struggle is the agency by which the conditions of existence of a mode of production are altered. However, moving well beyond Althusser and Balibar's vague but still historical conception of class struggle, Hindess and Hirst are completely unwilling to discuss the class struggle except as a free-floating and autonomous force—a "class voluntarism of the Cultural Revolution variety," to quote Elliott again. The epistemological rejection of history is thus a necessary step in the "liberation" of political practice from all structural and historical determination. The pursuit of political voluntarism dictates first an equation of history and teleology—"there can be no history without a philosophy of history" (Hindess and Hirst 1975, 311)—and second an escape from history by the expulsion of causality from the discursive analysis of social relations: "Causal doctrines are not necessary for specific discursive analysis. Dispensing with them also dispenses with debates and problems created solely by their presence" (Cutler, Hindess, Hirst, and Hussain 1977, 131).

What takes the place of history for Hindess and Hirst then is politics or, more precisely, political intentionality , which confers on the current situation whatever meaning it has. "The current situation must not be conceived as an object given in the real social reality at a given moment in time. . . . [T]he current situation does not exist independently of the political practice which constitutes it as an object" (Hindess and Hirst 1975, 322). Political practice dictates the choice of concepts, and their elaboration requires only the adumbration of their logical conditions of existence and the working out of their internal effects; both processes are completely independent of any consideration of the objective reality of these conditions of existence (which are theoretically emptied of any causal effectivity) or any correspondence between the logical deduction of such effects and any empirical reality. Strictly speaking, theoretical discourse entails nothing more than "the construction of problems for

analysis and solutions to them by means of concepts. Concepts are deployed in ordered successions to produce these effects. This order is the order created by the practice of theoretical work itself: it is guaranteed by no necessary 'logic' or 'dialectic' nor by any necessary mechanism of correspondence with the real itself" (Hindess and Hirst 1977, 7).

Within the terms of such a radical separation of discursive practice from reality and such an uncompromising rejection of the theoretical legitimacy of causality, there is obviously no place for a social formation in the Althusserian sense of the term, that is, an unevenly developed structured whole. Hindess and Hirst move rapidly from the antirealist voluntarism of Pre-Capitalist Modes of Production to a systematic rejection of any theoretical expression of structural interrelationships in Mode of Production and Social Formation and Marx's Capital and Capitalism Today . In the latter works, the concept of social structure is ruthlessly simplified by dissolving the Structural Marxist concept of relative autonomy into an infinity of absolute autonomies. There can no longer be, as there is for Structural Marxism, distinct theoretical levels ranging from the concepts of distinct structured elements to the complex whole of a "structure of structures" because, according to Hindess and Hirst, there is "no necessary form in which these concepts must be articulated into the concept of the essential structure of a social formation" (Cutler, Hindess, Hirst, and Hussain 1977, 229). While the conditions of existence of specific relations—for example, economic relations—may be logically deduced, the other social relations and practices that provide these conditions of existence cannot: "the form in which the conditions of existence of determinate relations of production are secured cannot be conceived either as empirically given to theory or as derivable in principle from the relations of production whose conditions of existence they secure" (Hindess and Hirst 1977, 25-26).

Balibar's concept of a mode of production, his concepts of correspondence and non-correspondence of the forces and relations of production, and indeed the very idea of forces and relations of production themselves are all finally dismissed by Hindess and Hirst as aberrations that follow from insisting that the conditions of existence of a mode of production be derived from its effects. This latter view, they argue, is conditioned in turn by the belief that a rational, unified concept of a social whole is necessary in order to account for the historical "fact" of transformation. For Hindess and Hirst, by contrast, "once the conception of social formation as a determinate unity of being, existence corresponding to its concept, is abandoned, then the problems of empirical

contingency on the one hand and of determination in the last instance on the other must vanish" (Hindess and Hirst 1977, 48). Hindess and Hirst reverse the terms of Balibar's rationalist essentialism, not in order to acknowledge irreducible discontinuities between concepts pitched at the level of the complex whole (the social formation) and those pitched at the level of the relatively autonomous elements (the economic, political, and ideological instances), but in order to eliminate the complexity of the social formation altogether and reduce it to "a single structure of social relations" (Hindess and Hirst 1977, 5).

Hindess and Hirst propose a concept of the social formation "conceived as a determinate form of economic class relations, their conditions of existence and the forms in which these conditions are secured" (Hindess and Hirst 1977, 50). They arrive at this concept by rejecting Balibar's distinction between the forces and relations and collapsing the former into the latter. Since the concept of a mode of production is now identical to that of the relations of production, it is superfluous and can be eliminated.

If relations of production are conceived as involving the social distribution of means-conditions of production between classes, that is, the distribution of possession of and separation from the means of production among different social categories of economic agents, then the specification of determinate relations of production necessarily involves an explicit or implicit reference to determinate means and processes of production, that is to determinate "forces of production." The "forces of production" provide certain of the conditions of existence of the relations of production in question. The specification of determinate "forces of production" must therefore be included in the specification of the forms in which the conditions of existence of a determinate set of relations of production are secured. Thus, if mode of production is not conceived in rationalist form as a necessary combination of distinct objects, then the concept "mode of production" is entirely redundant. (Hindess and Hirst 1977, 54-56)

All that remains of Balibar's concept of a mode of production is the concept of economic forms and their conditions of existence. Anything more than this, Hindess and Hirst contend, introduces the specter of "rationalist" causality:

Either the articulation of relations and forces of production is conceived in terms of the connection between social relations and the forms in which their conditions of existence are realized or it must be conceived in terms of some kind of necessity in which the character of one object of discourse, the "relations" or the "forces," is deducible from the concept of the other. The first alternative means there is no reason to posit "modes of production" as distinctive objects of analysis . . . while the second leads directly to the quagmire of rationalized conceptualizations. (Hindess and Hirst 1977, 55)

Hindess and Hirst's "new" conception of a social formation is actually nothing more than the concept of the social whole from the perspective of one of its instances, the economy, a nominalist move that suppresses the effectivities of the social whole and the non-economic practices. Hindess and Hirst are, in effect, privileging economic relations, but there is no reason for this privilege or any reason why it could not or should not be extended to any of the other practices (yielding commensurately different definitions of social formation). Having eliminated any objective reality outside discourse (including Althusser's materialist thesis) and all concepts of determination inside discourse (differential yet interrelated levels of analysis), Hindess and Hirst can produce no defense of the discursive primacy of the economy; any attempt to do so would simply restore the complexity of structural causality. Hindess and Hirst acknowledge a certain "discursive" effectivity: the specification of economic relations "involves the explicit or implicit reference to the effects of other social relations and of social practices other than economic production" (Hindess and Hirst 1977, 25), and the conditions of existence of those relations gives "certain abstract and general conditions which must be satisfied by political, legal, and cultural forms" (Cutler, Hindess, Hirst, and Hussain 1977, 127). However, in the face of their rejection of any concept of the articulation of the instances, it is difficult to see what difference this qualification makes.

Their nominalist reversal of Balibar's essentialist concept of a mode of production forces Hindess and Hirst to eliminate uneven development just as surely as Balibar does, but with even more regressive implications. If Balibar's rationalism hungers to eliminate contradiction in order to achieve unified and certain knowledge (the old rationalist desire to be God by participating in his logic), Hindess and Hirst simply reject knowledge altogether. The curse of the finite as opposed to the infinite intelligence—the distinction between concepts and things, the irreducible dislocations and aporias within discourse, and so on—is taken by Hindess and Hirst as a pretext for reducing theory to a cognitive epiphenomenon of political will: if Man cannot become God through knowledge, he can become a little God by subjugating knowledge to practice. In place of a scientific justification for the primacy of economic relations, Hindess and Hirst substitute the pragmatic justification of politics. "Problems created by politics . . . constitute the objects of theorisation and problematisation in Marxist discourse. . . . The political objectives of a socialist transformation of economic class relations pose the problem of the relations of production and their po-

litical and cultural conditions of existence as primary objects of theorisation" (Cutler, Hindess, Hirst, and Hussain 1977, 230). Once again we are witness to an underlying equivalence between the nominalism of the concept and the voluntarism of the will that is the essence of Hindess and Hirst's post-Althusserian post-Marxism. Ultimately class voluntarism lurks behind the false either/or of "essentialism or nominalism" by which Hindess and Hirst eliminate first the concept of the forces of production and then the concept of a mode of production itself from Marxist theory. These concepts must be eliminated not because they are theoretically superfluous (as Hindess and Hirst maintain) but because their existence constitutes a limitation, if only within theory, on political practice, a (theoretical) denial of the power of the political will over its (objectively real) conditions of existence.

Hindess and Hirst fuse the most abstract theoreticism with the most uncompromising political voluntarism. Political "calculation" (as they call it) is no longer a function of historical knowledge or indeed any knowledge at all except in the pragmatist sense of an explicitly formulated political interest. The complexity of a historically determinate social formation and the material relations of determination and subordination that create the space of the political subject are vitiated by a concept of the political class struggle as a free-floating process: "There are no 'socialist' issues and areas of struggle per se , assigned as 'socialist' by class interest and experience. Socialism is a political ideology. The basis for support for socialist politics is whatever issues and struggles from which it can be made" (Cutler, Hindess, Hirst, and Hussain 1978, 258-59). Political calculation creates the concept of the present situation and a concept of the future to be realized; political action destroys the former and calls into being the latter. By reducing politics to will to power unencumbered by knowledge, and social theory to a nominalism of concepts unencumbered by reality, Hindess and Hirst have moved not only "beyond" Althusser but "beyond" Marxism as well. Unable to provide either a compelling critique or a theoretically interesting alternative to structural causality and the concept of a mode of production, their particular current of post-Marxism flows smoothly into the larger stream of postmodernism.

Balibar and Wolpe: Articulation as a Concept of Transformation

In the course of his discussion of the labor process within the manufacturing mode of production (Althusser and Balibar 1970, 233-43), Bali-

bar provides a suggestive, if flawed, discussion of the "articulation," or structured interrelationship, between the forces of production (the relation of "real appropriation" that constitutes the labor process) and the relations of production (the property relation that determines access to the means of production and distribution of the social product). Balibar argues that the development of capitalism (and the dissolution of feudalism) involves two processes: first a shift in dominance within the social formation from feudal to capitalist relations of production, a movement Balibar calls the "formal subsumption" of labor to capital, and second a transformation of the labor process from the "unity" of handicraft production (in which there is a unity between human labor power and the means of labor insofar as tools are adapted to the individual laborer) to the "unity" of machine production (in which the unity is between the means of labor and the object of labor, and human labor power is forced to adapt to machines), what he calls the "real subsumption" of labor to capital. For Balibar, "manufacture," as an independent mode of production, marks an intermediate stage between feudal handicraft and capitalist machine production, a labor process in which the increasing division of labor and the de-skilling of labor are dissolving handicraft methods and creating the preconditions for machine processes. Balibar contends that the dominance of capitalist property relations over the labor process accounts for this transformation and that such a development is not possible until labor is separated not only from economic ownership of the means of production but from effective control over the labor process as well. The inalienable rights of the peasant community to the land and the artisan community to a trade are finally overthrown by alienable private property, at which point control of the labor process passes to economic owners who ruthlessly restructure, to use the contemporary euphemism, the labor process in the interest of capital accumulation.

Balibar's analysis nicely illustrates the strengths of his concept of the "double articulation" of labor, means of production, and ownership into forces and relations of production. First, it introduces power into the labor process, thereby vitiating "technicism," the idea that the labor process is neutral or that the development of productivity is independent of class struggle. If relations of property are exploitive, Balibar insists, the labor process cannot be simply a matter of technical cooperation. Second, because Balibar is dealing with structured relationships, that is, the social forces acting on human agents and not the human agents themselves, he avoids the false problem of classifying

individuals: Can a capitalist have an aristocratic title? Can a holder of seigneurial rights be a member of the bourgeoisie? Such questions are obfuscatory; they tend to obscure precisely what is decisive for explaining the nature of class struggle and historical transformation, namely, the forces and relations of production.

There is, however, an obvious weakness in Balibar's argument. As stated, it implies that capitalist relations of production spring into existence ex nihilo and then, alone, without the support of any capitalist forces of production, proceed to "dominate" feudal relations of production and to "subsume" feudal labor processes. How is it possible, given the double relation that Balibar insists defines every mode of production, for capitalist relations of production to come into existence without corresponding capitalist forces of production? The answer, of course, is that they cannot and did not. Balibar has simply telescoped the dissolution of feudalism and the development of capitalism by factoring out the coexistence and interrelationship of both a capitalist and a feudal mode of production. Balibar is not unaware of the difficulty; indeed, he even points out this alternative solution in the "very schematic suggestions" at the end of his long essay. Ultimately Balibar admits that an adequate conception of relations of correspondence and non-correspondence within a social formation "is only ever possible by a double reference to the structure of two modes of production" (Althusser and Balibar 1970, 307). Periods of transition (such as the age of manufacture) are determined by the "coexistence of several modes of production," and therefore they are "never one mode of production." Instead, their "unity is the coexistence and hierarchy of two modes of production. . . . Thus it seems that the dislocation between the connections and the instances in transition periods merely reflects the coexistence of two (or more) modes of production in a single 'simultaneity,' and the dominance of one of them over the other " (Althusser and Balibar 1970, 307).

Balibar thus introduces a new and powerful way of thinking about historical transformation in terms of the uneven development and shifting relations of domination and subordination between two (or more) modes of production. The fact that he envelops his concept of an articulation of modes of production in an essentialist-rationalist framework does not diminish his achievement. To realize the explanatory potential of the concept of articulation, however, it is necessary to eliminate its essentialist and rationalist connotations, in particular Balibar's notion that the structure of such an articulation is simply "a more general

synchrony englobing several systems and their relations" (Althusser and Balibar 1970, 307). What Balibar calls a manufacturing mode of production must be redefined as a capitalist mode of production within a social formation characterized by an articulation of feudalism and capitalism. The transition from feudalism to capitalism must be conceptualized in terms of a double, not a unitary, process—the dissolution of feudalism as well as the development of capitalism. Viewing the transformation in terms of an articulation of two modes of production eliminates the need for an essentialist concept of feudalism that bears within it the necessary development of capitalism. At the same time, the concept of articulation facilitates—within the framework of the uneven development and contradictions of feudalism itself—the possibility of the emergence of capitalism as well as the conditions of existence for its development from a subordinate to a dominant position. Obversely, articulation provides a concept of capitalism that neither assumes feudalism as its necessary precondition nor precludes the possibility of conceptualizing a feudal-capitalist articulation within which feudalism might occupy first a dominant and then a subordinate position.

The articulation of multiple modes of production suppresses neither the individual effectivities of each individual mode nor their structured interrelationship as an economic system, the economic instance of the social formation. To speak of articulation is to speak of uneven development, of the primacy of contradiction over reproduction, and of the dissolution of one mode of production and the development of another, but not in such a way that the internal logic of one process becomes, by some mysterious process of transubstantiation, the internal logic of the other. For each individual mode of production we may distinguish between phases of development (wherein the forces and relations are in a state of what Balibar calls "correspondence" such that the subsumption of the forces to the relations contributes to the reproduction of the existing relations of production) and phases of dissolution (wherein changes in productivity, technology, or the organization of production make the reproduction of existing property relations increasingly difficult—what Balibar calls "non-correspondence"). Because it is precisely the contradictions of their uneven development that are being reproduced, the reproduction of the forces and relations of production is a process neither of simple repetition nor of linear perpetuation. It is rather a process, on the one hand, of accumulating and creating contradictions ("condensation" in Althusser's terminology) and, on the other hand, of adapting, with increasing difficulty, to the ongoing conse-

quences of uneven development ("displacement" for Althusser). Eventually the non-correspondence between the forces and relations of production may become so great that their structured relationships are dissolved, frequently by means of violence (what Althusser calls a "revolutionary rupture"), and relations of ownership and labor are transformed, that is, reorganized on the basis of a new mode of production. This process, open-ended but not unbounded, constitutes what we might call the internal logic of a mode of production.

Of course, articulation implies more than the simultaneous existence of otherwise isolated modes of production. It also implies their structured interrelationship and the dominance of one over the other. The dominant mode is dominant precisely insofar as it forces the subordinate mode to adapt, within the limits of the latter's own structures, to the requirements of the development and reproduction of the dominant mode. The effects of dominance and subordination are undoubtedly felt within both modes of production—within the dominant mode as "super-appropriation" and dominance for owners and relative prosperity for producers and within the subordinate mode as dependency for owners and "super-exploitation" for producers. Thus articulation constitutes something like an external logic of the articulated modes of production. Articulation materially alters the respective conditions of existence of each mode, thereby influencing the internal rhythms of their respective development and dissolution, decisively in the case of the subordinate mode, which may or may not be allowed to expand, and more or less significantly in the case of the dominant mode.

Harold Wolpe, a British anthropologist and historian of South Africa, has considerably clarified the concepts of articulation and dominance by means of a distinction drawn between "restricted" and "extended" concepts of a mode of production (Wolpe 1980). A restricted concept, Wolpe argues, pertains to the specification of possible relations between agents and the means of production within individual economic enterprises or production units. Such a conceptualization is "restricted" in the sense that it says nothing about the "laws of motion" of the mode of production, that is, how economic enterprises are related to each other and the processes by which the forces and relations of production are reproduced. The concept of the laws of motion or reproductive mechanisms, when added to the restricted concept of the forces and relations, yields an "extended" concept of a mode of production. Wolpe goes on to define the articulation of two modes of production in terms of a dominant mode defined "extensively" (forces and

relations plus the structures necessary for their reproduction) and articulated with a subordinate mode defined "restrictively" (in terms simply of the forces and relations of production). Thus, for Wolpe, the unity of a social formation is ultimately constituted by the laws of motion or mechanisms of reproduction of the dominant mode of production. Economic enterprises organized according to a subordinate mode depend on the dominant mode for their own reproduction. The dominant mode therefore becomes essential to the subordinate mode, and this relation of dependency constitutes the essence of domination.

Despite his mechanistic reference to the "laws of motion" of a mode of production, Wolpe has significantly extended and enriched Balibar's concept of articulation. To make Wolpe's concepts compatible with structural causality it is necessary only to understand the terms laws of motion and reproductive mechanisms as referring to the place and function assigned to the instances by the complex whole; to understand the distinction Wolpe draws between restricted and extended concepts of a mode of production as referring to, respectively, the economic function and the existing relations of domination and subordination through which the primacy of the economic function is manifested in the political and ideological instances; and finally, to understand the dominance of one mode of production over another within an articulation as the primacy of its function manifested in the political and ideological instances to the exclusion of the subordinate economic function.

Wolpe's framework adds significantly to the explanatory potential of the concept of articulation with regard to the problem of the transition from feudalism to capitalism (as we shall see, it helps clarify the problem of imperialism as well). It makes it possible to conceptualize the development of capitalism within a social formation in which the laws of motion of a feudal mode of production are dominant, but in which the laws of motion of a capitalist mode might displace feudal laws of motion without thereby destroying the feudal character of a particular economic enterprise. There is no necessary connection, Wolpe explains, between the reproduction of the relations and forces of a subordinate mode, at the level of the enterprise, and the self-generating reproduction of the laws of motion of that subordinate mode, nor do the laws of motion of a dominant mode preclude the reproduction of relations and forces of a subordinate mode. However, the laws of motion of a dominant mode are not without their effect on all the enterprises of the social formation regardless of their particular organization. The effect will vary with the nature of the enterprise: under the laws of motion of

a dominant capitalist mode of production, capitalist enterprises will be conditioned by the processes of the formation of the average rate of profit and its effects on capitalist calculation; for enterprises organized according to feudal forces and relations, however, the laws of motion of capitalism will affect a host of conditions (the level of feudal profit, organization of the labor process, amount of rent, and so on) but not necessarily the feudal forces and relations of production themselves or the forms of economic calculation based on them.