Ellen Meiksins Wood
The Separation of the Economic and the Political in Capitalism
The intention of Marxism is to provide a theoretical foundation for interpreting the world in order to change it. This is not an empty slogan. It has—or ought to have—a very precise meaning. It means that Marxism seeks a particular kind of knowledge, one which is uniquely capable of illuminating the principles of historical movement and, at least implicitly, the points at which political action can most effectively intervene. This is not to say that the object of Marxist theory is to discover a ‘scientific’ programme or technique of political action. Rather, the purpose is to provide a mode of analysis especially well equipped to explore the terrain on which political action must take place. It can, however, be argued that Marxism since Marx has often lost sight of his theoretical project and its quintessentially political character. In particular, this is so to the extent that Marxists have, in various forms, perpetuated the rigid conceptual separation of the ‘economic’ and the ‘political’ which has served bourgeois ideology so well ever since the classical economists discovered the ‘economy’ in the abstract and began emptying capitalism of its social and political content. [*] My thanks must go to several people who have read and criticized—often vehemently—this essay at various stages: Perry Anderson, Robin Blackburn, Robert Brenner, Ralph Miliband, Neal Wood, Gregory Meiksins, Peter Meiksins, and my students at York University, Toronto, especially Frances Abele and George Comninel.
Since, however, these conceptual devices do reflect—albeit in a distorting mirror—an historical reality specific to capitalism, a real differentiation of the ‘economy’, an attempt to rescue them from bourgeois ideology and make them illuminate more than they obscure might begin by reexamining the historical conditions that made such conceptions possible and plausible. The purpose of this reexamination would not be to explain away the ‘fragmentation’ of social life in capitalism, but to understand precisely what it is in the historical nature of capitalism that appears as a differentiation of ‘spheres’—in particular, the ‘economic’ and the ‘political’. It may be possible to interpret this historical ‘fragmentation’ in such a way that the ‘fetishism’ of capitalist categories can be overcome, but without obscuring the historical realities they reflect.
The differentiation of the ‘economic’ and the ‘political’ is, of course, not simply a theoretical but a practical problem. There is perhaps no greater obstacle to socialist practice than the separation of economic and political struggles which has typified modern working class movements. If this obstacle were, as many revolutionary socialists have contemptuously suggested, merely the product of a misguided, ‘underdeveloped’, or ‘false’ consciousness on the part of the working class, it might be easier to overcome. The tenacity of working class ‘economism’, however, derives precisely from its correspondence to the realities of capitalism and the ways in which capitalist appropriation and exploitation actually do divide the arenas of economic and political action, and actually do transform certain essential political issues—struggles over domination and exploitation that historically have been inextricably bound up with political power—into distinctively ‘economic’ issues. This ‘structural’ separation may, indeed, be the most effective defense mechanism available to capital.
If, therefore, the object of Marxist theory is to shed light on the terrain of political action, it can neither ignore these historical realities nor ratify them by entrenching the separation of economics and politics that has served capitalism so well in theory and practice. Instead, it should explain precisely how and in what sense capitalism has driven a wedge between the economic and the political—how and in what sense essentially political issues like the disposition of powers to control production and appropriation, or the allocation of social labour and resources, have been cut off from the political arena and displaced to a separate ‘sphere’.
The Separation of Economic and Political ‘Factors’
Karl Marx presented the world in its political aspect, not only in his explicitly political works but even in his most technical economic writings. His critique of political economy was, among other things, intended to reveal the political face of the economy which had been obscured by bourgeois political economists. The fundamental secret of capitalist production disclosed by Marx—the secret that political economy systematically concealed, making it finally incapable of accounting for capitalist accumulation—concerns the social relation and the disposition of power that obtains between the worker and the capitalist to whom he sells his labour-power. This secret has a corollary: that the disposition of power between the individual capitalist and worker has as its condition the political configuration of society as a whole—the balance of class forces and the powers of the state which permit the expropriation of the direct producer, the maintenance of absolute private property for the capitalist, and his control over production and appropriation. In volume 1 of Capital Marx works his way from the commodity form through surplus value to the ‘secret of primitive accumulation’, disclosing at last that the ‘starting point’ of capitalist production ‘. . . is nothing else than the historical process of divorcing the producer from the means of production’,  Karl Marx, Capital, Moscow vol. I, p. 668. a process of class struggle and bloody intervention by the state on behalf of the expropriating class. The very structure of the argument suggests that, for Marx, the ultimate secret of capitalist production is a political one. What distinguishes his analysis so radically from classical political economy is that it creates no sharp discontinuities between economic and political spheres; and he is able to trace the continuities because he treats the economy itself not as a network of disembodied forces but, like the political ‘sphere’, as a set of social relations.
This has not, however, been equally true of Marxism since Marx. In one form or another and in varying degrees, Marxists have generally adopted modes of analysis which, explicitly or implicitly, treat the economic ‘base’ and the legal, political, and ideological ‘superstructures’ which ‘reflect’ or ‘correspond’ to it as qualitatively different, more or less enclosed and ‘regionally’ separated spheres. This is most obviously true of orthodox base-superstructure theories. It is also true of their variants which speak of economic, political, and ideological ‘factors,’ ‘levels’ or ‘instances’, no matter how insistent they may be about the interaction of factors or instances, or about the remoteness of the ‘last instance’ in which the economic sphere finally determines the rest. Indeed, these formulations merely emphasize the spatial separation of spheres.
Other schools of Marxism have maintained the abstraction and enclosure of spheres in other ways—for example, by abstracting the economy or the circuit of capital in order to construct a technically sophisticated alternative to bourgeois economics, meeting it on its own ground (and going significantly further than Marx himself in this respect, without grounding the economic abstractions in historical and sociological analysis as he did). The social relations in which this economic mechanism is embedded—which indeed constitute it—are treated as somehow external. At best, a spatially separate political sphere may intervene in the economy, but the economy itself is evacuated of social content and is, as it were, depoliticized. In these respects, Marxist theory has perpetuated the very ideological practices that Marx was attacking—those practices that confirmed to the bourgeoisie the naturalness and eternity of capitalist productive relations.
Bourgeois political economy, according to Marx, universalizes capitalist relations of production precisely by analyzing production in abstraction from its specific social determinations. Marx’s approach differs from theirs precisely in his insistence that a productive system is made up of its specific social determinations—specific social relations, modes of property and domination, legal and political forms. This does not simply mean that the economic ‘base’ is reflected in and maintained by certain ‘superstructural’ institutions, but that the productive base itself exists in the shape of social, juridical, and political forms—in particular, forms of property and domination. Bourgeois political economists are able to demonstrate ‘the eternity and harmoniousness of the existing social relations’ by divorcing the system of production from its specific social attributes. For Marx, production is ‘. . . not only a particular production . . . it is always a certain social body, a social subject, which is active in a greater or sparser totality of branches of production’.  Marx, Grundrisse, Harmondsworth 1973, p. 86. Bourgeois political economy, in contrast, achieves its ideological purpose by dealing with society in the abstract, treating production as ‘. . . encased in eternal natural laws independent of history, at which opportunity bourgeois relations are then quietly smuggled in as the inviolable natural laws on which society in the abstract is founded. This is the more or less conscious purpose of the whole proceeding.’  Ibid., p. 87. While bourgeois economists may recognize that certain legal and political forms facilitate production, they do not treat them as organic constituents of a productive system. Thus they bring things that are organically related ‘. . . into an accidental relation, into a merely reflective connection.’  Ibid., p. 88.
The distinction between ‘organic’ and ‘merely reflective’ connections is especially significant. It suggests that any application of the base/superstructure metaphor that stresses the separation and enclosure of spheres—however much it may insist on the connection of one to the other, even the reflection of one by the other—reproduces the mystifications of bourgeois ideology insofar as it fails to treat the productive sphere itself as defined by its social determinations and in effect deals with society ‘in the abstract’. The basic principle about the primacy of production, the very foundation of historical materialism, thus loses its critical edge and is assimilated to bourgeois ideology.
This is, of course, not to say that Marx saw no value in the approach of bourgeois political economy. On the contrary, he adopted its categories as his point of departure precisely because they expressed, not a universal truth, but a historical reality in capitalist society, at least a ‘real appearance’. The point was to decipher the real meaning of the ‘appearance’, and this required not the reproduction but the critical elaboration of bourgeois categories.
Desocializing the Material Base
Precisely these criticisms of political economy have recently been used in an important book by G. A. Cohen to support an argument against a social interpretation of materialism; and since his argument is in many respects the very antithesis of the one presented here, some comments may be useful.  G. A. Cohen, Karl Marx’s Theory of History: A Defense, Oxford 1978. It will, of course, be impossible to consider every step of Cohen’s dense and impressive argument; but there is one pivotal step which is indispensable to the argument and which sums up the major point at issue between our opposing interpretations of materialism. This is Cohen’s formulation of the distinction between the ‘material’ and the ‘social’. There are two major points in dispute: Cohen’s analytic distinction itself and the slippage by means of which an analytic distinction is allowed imperceptibly to become not only a dualism but an historically real separation and a causal relation.
Since Cohen’s object is to establish that historical materialism is a technological determinism, he must not only define the determinant ‘material substratum’ narrowly to include only technical forces of production but identify the material sphere with the ‘natural’, as something in principle separate and qualitatively different from the ‘social’ and ‘historical’. Even if, as he concedes, the ‘material’ never exists in history except ‘enveloped’ in social form, his causal argument obliges him in effect to treat the ‘material’ as if it were only externally related to the ‘social’ and as if it had a life of its own, subject to laws of motion different from ‘historical’ principles. Reduced to its simplest terms, his technological determinism means that the ‘natural’ impulses which propel the material sphere—the development of technical forces—prevail over, and in one way or another causally determine, the historical development of social forms. The premise is that there is a natural and perennial impulse, independent of social and historical conditions, grounded in human nature and rationality, toward the improvement of technological forces.  Cohen, pp. 152–3. At any given stage of development, then, those social relations must emerge which will facilitate that improvement. In turn, there will come into being such legal and political forms as are required by these social relations. In short, Cohen offers us a ‘base/superstructure’ analysis (qualified by the proposition that ‘bases need superstructures’, which ought to be unexceptionable to all exponents of this mechanical metaphor) in which the relations of production themselves become ‘superstructural’ in their relation to the real ‘base’, the technical forces of production.
To establish the conceptual foundations for his causal propositions about forces and relations of production, Cohen cites the authority of Marx: ‘We are arguing that the familiar distinction between forces and relations of production is, in Marx, one of a set of contrasts between nature and society. Commentators have failed to remark how often he uses “material” as the antonym of “social” and of “formal”, how “natural” belongs with “material” against “social”, and how what is described as material also counts as the “content” of some form. (Other terms of the material vocabulary are “human”, “simple”, and “real”, while “historical” and “economic” consort with “social”.) The upshot of these oppositions and identifications is that the matter or content of society is nature, whose form is the social form.’  Cohen, p. 98.
The argument turns on the identification of ‘material’ with ‘nature’ and ‘natural’ development in opposition to ‘social’ and ‘historical’; and for this definition of the ‘material’, Cohen relies on readings of Marx which are so dubious that his whole interpretation of historical materialism is put into question. The striking thing is that the passages from Marx which Cohen cites to support the ‘illuminating abstraction’ separating the ‘material’ from the ‘social’, or ‘material production’ from its ‘social features’—or, for example, the ‘underlying matter’ of capitalist production from the ‘capitalist economic form’—are precisely those in which Marx’s intention is to attack rather than defend this abstraction.
Cohen cites various passages from Capital and the Grundrisse which, in one form or another, refer to the ‘productive activity of human beings in general’, or ‘the process of production in general’ or material production ‘in the abstract’  For example, on p. 99 he cites Grundrisse, p. 304 and Capital III, p. 795 (Moscow, 1962. In the Moscow, 1971, edition, p. 815)..
The distinction that here concerns Marx is not between ‘material’ and ‘social’ or between a ‘material process of production’ and a ‘social process of production’, but between production ‘as such’ or ‘in general’ and production as it actually exists, as a social process in socially and historically determinate forms. (Indeed, a similar contrast could be formulated between the ‘social process of production in general’ and the social process of production in historically determinate forms.  See, for example, Capital III, Moscow 1971, p. 818.) It is not, for example, a question of distinguishing the ‘capitalist form’ from its ‘underlying matter’ but ‘matter’ in capitalist form distinguished from ‘matter’ in the abstract.
Marx’s object is to criticize the mystifications of political economy which are achieved precisely by beginning with ‘material production in general’ and then proceeding to treat the process of producing capital abstractly as if it were the process of production as such.  e.g., Grundrisse, pp. 85–88. It is in the nature of capitalism to make such mystifications particularly plausible because the production of the conditions of material life in capitalism is inseparable from the production of capital. For example, since commodity production is generalized, all production of use-values is at the same time and indistinguishably production of exchange-values. In the Grundrisse and Capital Marx unveils the false appearances of capitalist production by tracing the stages of mystification in the production of capital. In the Grundrisse, he suggests briefly that one might begin by identifying the elements which are ‘common’ or ‘general’ to all production; but this suggestion does little to support Cohen’s case, since, first, there is no reason to equate the ‘common’ or ‘general’ with the ‘material’ in opposition to the ‘social’; and, above all, because Marx rejects this procedure on the grounds that any propositions about ‘production in general’ will be rather empty and formal, even ‘trite’ or tautological, since the real content of these ‘common elements’ themselves depends precisely on their social determinations.  loc. cit.
The labour-process can, it is true, be reduced to ‘simple’ or ‘elementary’ or ‘common’ factors: the personal activity of man, the producer; a subject of work or material worked upon; instruments. Viewed in this way, however, as ‘solely a process between man and nature’, the labour-process is treated as if it were performed by an ‘abnormally isolated’ human being (the infamous Robinson Crusoe who, according to Marx, so often lurks behind the mystifications of political economy) instead of as it really is: a social process in which the relationship to nature is at the same time and inseparably a social relation.  Capital III, pp. 883–4. The ‘simple’ elements that are common to all production—both ‘abnormally isolated’ (or imaginary) and social—are ‘elementary’ only in the abstract sense that all kinds of production must possess such elements in one form or another. The content, of these ‘elementary’ factors like that of ‘social’ factors, is socially and historically determined. Furthermore, Marx does not suggest that in normal forms of production, which entail ‘social assistance’, this social element is somehow less fundamental or even less ‘material’ than the ‘simple’ or ‘common’ elements. Nor does he imply that the simple elements in such cases have a causal priority over the social. Thus, when Marx himself isolates the ‘simple’ labour-process in volume 1 of Capital (only after analysing the commodity form), his object is not to separate the ‘simple’ elements of the labour-process from their social determinations or to establish the ‘primacy’ of these elements. He intends rather to explain how the particular nature of the capitalist labour-process, the particular nature of its ‘simple’ elements themselves, is inextricably bound up with the ‘social’ and ‘historical’ fact that the process of production in capitalism is at once a process of producing surplus-value and the capitalist relation itself.
Marx’s purpose, then, is to stress not the dualism of the ‘material’ and the ‘social’ but the definition of the material by the social; to define the material process of production not in opposition to the social process of production but as a social process; to focus attention not on ‘abstract matter’ but on the social form that gives it reality; to indicate not the usefulness but the emptiness of this abstraction; and insofar as he draws our attention to the abstraction of material production from its particular social form, he does so to stress not what the abstraction reveals but what it conceals. Cohen’s ‘illuminating abstraction’ is thus the very mystification Marx is attacking.
The purpose of Cohen’s conceptual framework is to support his argument for technological determinism. The strength of the conceptual foundation, then, must be judged by the weight of the argument it is able to bear. In the final analysis, Cohen’s propositions about the causal connections between forces and relations of production prove insubstantial. These propositions do not, as Cohen hastens to stress, entail any particular temporal sequence. Dynamic forces of production may break through the integument of social relations and compel them to change accordingly; or sluggish forces may by their very failure to develop compel social relations to change in order to encourage and accelerate technological progress. Indeed, Cohen’s formula can accommodate both cases in which, as Marx puts it, the forces of production are ‘petrified’  Capital I, p. 456–7. (which may be the rule rather than the exception) and the radically unique case of capitalism, which is distinguished precisely by its drive constantly to revolutionize the forces of production. This flexibility makes it unnecessary to explain away awkward historical and anthropological evidence; but it leaves Cohen’s basic historical proposition rather empty and renders it of little use as an explanatory device. The proposition is, in effect, non-falsifiable. To the extent that it is true, it is trivial or tautological—as, perhaps, any historical ‘law’ of such generality must be.
In a sense, what Cohen’s technological determinism does is to repeat the error of the political economists: he generalizes the particular historical experience of capitalism by abstracting the laws of capitalist production from their specific social determinations. The drive to revolutionize the forces of production, which in capitalism is generated by a particular mode of surplus-extraction—the mechanism of surplus value—and by the social relation between capital and labour this implies, thus becomes a natural law implanted in human nature and enforced by the laws of reason.  See Robert Brenner, “The Origins of Capitalism”, nlr, 304, for a discussion of how capitalist relations of production uniquely demand the revolutionary transformation of productive forces.
Cohen’s particular definition of the ‘material’ and its relation to the social thus makes it difficult to account for the evolution of capitalism and the distinctive effect that its social relations of production have had on technological development. Indeed—and even more fundamentally—his radical separation of the ‘material’ and the ‘social’ makes non-sense out of precisely those materialist laws of contradiction and historical transition which he invokes to support his case. If there is any meaning in the proposition that contradictions between forces and relations of production give impetus to historical movement, it is arguably only insofar as ‘forces’ are considered in their social aspect. It is precisely because the ‘material’ and the ‘social’ are not, as it were, on two different planes of being that it makes sense to speak of ‘contradictions’ between them. For example, the critical contradiction in capitalism is not between narrowly defined technical forces and social relations, but between two potentially antagonistic social principles: the individualistic, even anti-social, form of capitalist property and the highly socialized form of capitalist production. The ‘material force’ most antagonistic to the social relations of capital is a united and class-conscious proletariat. Not even Cohen would maintain that the generation of this force is a mere reflex of technological development.
It is also worth noting that Marx and Engels go so far as to suggest that the very possibility of a separation and contradiction between forces and relations of production is dependent on specific social conditions. Such a separation becomes possible only when production and consumption, labour and enjoyment, ‘devolve on different individuals’ in the social division of labour.  Marx and Engels, German Ideology in Collected Works, New York 1976, vol. 5, p. 45. A similar principle applies to the separation of ‘factors’—the economic, political, etc.—and their ‘relative autonomy.’ The ultimate foundation of these separations is the social division of labour which creates ‘new and independent spheres’ by assigning people to perform new and independent social functions.  Engels, Letter to Conrad Schmidt, October 27, 1890, in Marx and Engels, Selected Correspondence, Moscow 1965, p. 422.
In short, the ‘material’ on which the structure of historical materialism rests is from the outset a ‘social’ and historical phenomenon. It can even be said that the essence of this materialism—in contrast, say, to the materialism of the political economists—is precisely that it socializes and historicizes the material base. There are, therefore, no radical disjunctures between ‘material’ and ‘social’, ‘base’ and ‘superstructure’, ‘objective structures’ and historical specificities, which make it difficult to move from one to the other in theory and to move between theory and practice.
The ‘Unitarian’ Approach
A new approach to Marxist theory has recently emerged which attempts to bridge the discontinuities between ‘base’ and ‘superstructure’ by broadening the meaning of the ‘base’ itself. The conceptual ‘fragmentation’ inherited from bourgeois ideology is here overcome by defining away the specificity of productive forces and relations and redrawing the boundaries of the productive sphere so that it can encompass virtually any area of social activity and consciousness in a ‘unitary’ conception of social experience, or at least ‘class experience.’ The virtue of this ‘unitarian’ approach is that it attempts to restore some kind of social and historical content to the ‘economy’ and that, unlike both economistic and structuralist Marxisms, it recognizes and rejects what has been called the ‘fetishism’ of capitalist categories. In the process, however, historical materialism may have been defined out of existence altogether. If the concepts of production and productive relations are deprived of their integrity and specificity, the ‘mode of production’ loses its explanatory value and the foundations of materialism simply collapse.
A recent essay by Simon Clarke illustrates the ‘unitarian’ approach in one of its forms.  Simon Clarke, ‘Socialist Humanism and the Critique of Economism’, History Workshop 8, Autumn, 1979, pp. 138–156. For a somewhat different kind of ‘unitarian’ approach, see, for example, Raymond Williams’ Marxism and Literature, Oxford 1977. Having rejected the ‘reduction of relations of production to forms of exploitation’, he advocates ‘. . . a unitary conception of relations of production as relations between people who, in a class society, relate to one another as members of antagonistic social classes. These relations have political and cultural, as well as economic, dimensions, the unity of which consists in their human character.’  Clarke, p. 144. He goes on to argue that capitalist social relations must be ‘criticized from the standpoint of the experience of those who live within those social relations . . .’, and that ‘This experience is not the experience of atomized individuals, but is a class experience, the collective experience of oppression in all its forms.’ Thus, relations of production are, in effect, conflated with class relations, and both are made to encompass virtually all forms of social—indeed, human—experience more or less indiscriminately. It is difficult to see what meaning is left in such a diffuse conception of production relations.
Of course the ‘immediate process of production’ and relations of exploitation within it are not the whole story of domination and oppression. It is also worth pointing out that there is a difference between capitalist exploitation, in which surplus-extraction is inseparable from the immediate process of production, and pre-capitalist exploitation in which surplus-extraction takes place by means of ‘extra-economic’ coercion external to the process of production itself. (This is an important point which will arise again throughout this essay.) Nevertheless, the complex structure of social and political domination always has at its core the extraction of surplus from the immediate process of production—not least, in the sense that this process creates the revenues on which the whole structure feeds, and in the sense that the power of social oppression ultimately depends on the power to appropriate the surplus that sustains it. This is as true of pre-capitalist societies as of capitalism; and it is not altered by the fact that in pre-capitalist social formations the extraction of surplus from the immediate process of production may take ‘extra-economic’ forms, that ‘economic’ exploitation may be achieved by political means. The rigid differentiation of ‘economic’ and ‘political’ levels may, therefore, be a ‘fetishization’ of capitalist categories; but it is not clear how this accusation applies, as Clarke suggests, to the identification of relations of production with forms of exploitation. And of course Clarke is right to stress that ‘The relations of exploitation within the immediate process of production imply the existence of class relations within which social production takes place’; that class represents the collective ‘experience’ of oppression—an experience that is at the same time economic, political, ideological, and cultural; and that classes are the chief historical agents in the contest to maintain or transform relations of domination and oppression. Nevertheless, class relations and relations of production cannot be synonymous—if only because it is not classes that produce and appropriate. The indisputable importance of understanding the class ‘experience’ in its totality (provided that the concept of class itself has some specificity) is no warrant for dissolving the ‘relations of production’.
Toward a Theoretical Alternative
It should be possible to maintain the integrity and specificity of the mode and relations of production without emptying them of their social content or creating artificial discontinuities between them and the whole complex of social activities and relations in which they are embedded. It should, in other words, be possible to take seriously Marx’s own insistence, in opposition to the ideological abstractions of bourgeois political economy, that (for example) ‘capital is a social relation of production’, that economic categories express certain determinate social relations. There ought, then, to be a theoretical alternative to ‘vulgar economism’ that attempts both to preserve the integrity of ‘production’ and to work out the implications of the fact that the productive ‘base’ exists in the shape of specific social processes and relations and particular juridical and political forms. Unfortunately, there has been no explicit and systematic account of such a theoretical position, although something like it is implicit in the work of certain Marxist historians.
The theoretical standpoint being proposed here is perhaps what has been called—disparagingly—‘political Marxism’. This brand of Marxism, according to a recent critic, is a ‘reaction to the wave of economist tendencies in contemporary historiography. As the role of class struggle is widely underestimated, so [political Marxism] injects strong doses of it into historical explanation. . . . It amounts to a voluntarist vision of history in which the class struggle is divorced from all other objective contingencies and, in the first instance, from such laws of development as may be peculiar to a specific mode of production. Could one imagine accounting for the development of capitalism in the nineteenth and twentieth centuries solely by reference to social factors, and without bringing into the picture the law of capitalist accumulation and its mainspring, that is to say the mechanism of surplus value? In fact, the result . . . is to deprive the basic concept of historical materialism, that is the mode of production, of all real substance. . . . The error of such “political Marxism” lies not only in its neglect of the most operative concept of historical materialism (the mode of production). It also consists in its abandonment of the field of economic realities. . . .’  Past and Present, no. 79, pp. 67–68. The author, Guy Bois, is referring specifically to the article by Robert Brenner mentioned below in note 23. It is worth noting that, while Bois attacks Brenner for his departure from economism, Simon Clarke does the opposite, criticizing Brenner for his ‘definition of relations of production in Dobbian terms as forms of exploitation’. Clarke, p. 155, n. 3. This simply reinforces my view that Brenner has got it just about right.
The purpose of the present argument is precisely to overcome the false dichotomy that permits some Marxists to accuse others of abandoning the ‘field of economic realities’ when they concern themselves with the political and social factors that constitute relations of production and exploitation. The premise here is that there is no such thing as a mode of production in opposition to ‘social factors’, and that Marx’s radical innovation on bourgeois political economy was precisely to define the mode of production and economic laws themselves in terms of ‘social factors’.
What does it mean to talk about a mode of production or an economy as if they were distinct from, even opposed to, ‘social factors’? What, for example, are ‘objective contingencies’ like the law of capitalist accumulation and its ‘mainspring’, the ‘mechanism’ of surplus value? The mechanism of surplus value is a particular social relation between appropriator and producer. It operates through a particular organization of production, distribution, and exchange; and it is based on a particular class relation maintained by a particular configuration of political power. What is the subjection of labour to capital, which is the essence of capitalist production, if not a social relation and the product of a class struggle? What, after all, did Marx mean when he insisted that capital is a social relation of production; that the category ‘capital’ had no meaning apart from its social determinations; that money or capital goods are not in themselves capital but become so only in the context of a particular social relation between appropriator and producer; that the so-called primitive accumulation of capital which is the pre-condition for capitalist production is nothing more than the process—i.e. the class struggle—whereby the direct producer is expropriated? and so on. For that matter, why did the grand old man of bourgeois social science, Max Weber, insist on a ‘purely economic’ definition of capitalism without reference to extraneous social factors (like, for example, the exploitation of labour), evacuating the social meaning of capitalism in deliberate opposition to Marx?  See, for example, Weber’s Economy and Society, New York 1968, pp. 91 and 94, and Agrarverhältnisse im Altertum in Gesammelte Aufsätze zur Sozial- und Wirtschafts-geschichte Tubingen 1924, pp. 15–16. In English translation nlb, The Agrarian Sociology of Ancient Civilizations, pp. 50–1.
To raise these questions and to insist on the social constitution of the economy is not at all to say that there is no economy, that there are no economic ‘laws’, no mode of production, no ‘laws of development’ in a mode of production, no law of capitalist accumulation; nor is it to deny that the mode of production is the ‘most operative concept of historical materialism.’ ‘Political Marxism’, as understood here, is no less convinced of the primacy of production than are the ‘economistic tendencies’ of Marxism. It does not define production out of existence or extend its boundaries to embrace indiscriminately all social activities or even class ‘experiences’. It simply takes seriously the principle that a mode of production is a social phenomenon.
Equally important—and this is the point of the whole exercise—relations of production are, from this theoretical standpoint, presented in their political aspect, that aspect in which they are actually contested: as relations of domination, as rights of property, as the power to organize and govern production and appropriation. In other words, the object of this theoretical stance is a practical one, to illuminate the terrain of struggle by viewing modes of production not as abstract structures but as they actually confront people who must act in relation to them. This object is not, however, served by dissolving the relations of production in an undifferentiated mass of social relations or class ‘experience’, in which there is no way of identifying critical targets.
‘Political Marxism’ recognizes the specificity of material production and productive relations. It insists, however, that ‘base’ and ‘superstructure’, or the ‘levels’ of a social formation, cannot be viewed as compartments or ‘regionally’ separated spheres. However much we may stress the interaction among ‘factors’, these theoretical practices mislead because they obscure not only the historical processes by which modes of production are constituted but also the structural definition of productive systems as living social phenomena. ‘Political Marxism’, then, does not present the relation between base and superstructure as an opposition, a ‘regional’ separation between a basic ‘objective’ economic structure, on the one hand, and social, juridical, and political forms, on the other, but rather as a continuous structure of social relations and forms with varying degrees of distance from the immediate processes of production and appropriation, beginning with those relations and forms that constitute the system of production itself. The connections between ‘base’ and ‘superstructure’ can thus be traced without great conceptual leaps because they do not represent two essentially different and discontinuous orders of reality.
‘Base’ and ‘Superstructure’ Reconsidered
The argument begins with one of the first principles of Marx’s materialism: that while men work within definite material limits not of their own making, including purely physical and ecological factors, the material world as it exists for man is not simply a natural given; it is a mode of productive activity, a system of social relations, an historical product. Even nature, ‘. . . the nature that preceded human history . . . is nature which no longer exists anywhere . . .’;  German Ideology op. cit., p. 39. ‘. . . the sensuous world . . . is, not a thing given direct from all eternity, remaining ever the same, but the product of industry and the state of society; and, indeed, in the sense that it is an historical product, the result of the activity of a whole succession of generations, each standing on the shoulders of the preceding one, developing its industry and its intercourse, modifying its social system according to the changing needs.’  German Ideology, p. 40. A materialist understanding of the world, therefore, is an understanding of the social activity and the social relations through which men interact with nature in producing the conditions of life; and it is an historical understanding which acknowledges that the products of social activity, the forms of social interaction produced by men, themselves become material forces, just as much as are natural givens. This account of materialism, with its insistence on the role played by social forms and historical legacies as material forces, inevitably raises the vexed question of ‘base’ and ‘superstructure’. If forms of social interaction, and not just natural or technological forces, are to be treated as integral parts of the material base, where is the line to be drawn between social forms that belong to the base and those that can be relegated to superstructure? Or, in fact, does the base/superstructure dichotomy obscure as much as it reveals about the productive ‘base’ itself?
Relations of production take the form of particular juridical and political relations—modes of domination and coercion, forms of property and social organization—which are not mere secondary reflexes but constituents of the productive relations themselves. It is no doubt possible, indeed necessary, to distinguish between juridical-political forms that are constituents of productive relations and those that are more distant from, or external to, these relations—even if there are no sharp discontinuities between them. There are certainly legal and political institutions that cannot be usefully regarded as constituents of productive relations even if they help to sustain the system of production and to reproduce its essential relations; and perhaps the term ‘superstructure’ should be reserved for these. However, not all legal and political principles can be relegated to superstructure, since the material base itself is articulated through juridical-political forms. The ‘sphere’ of production is dominant not in the sense that it stands apart from or precedes these juridical-political forms, but rather in the sense that these forms are precisely forms of production, the attributes of a particular productive system.
A mode of production is not simply a technology but a social organization of productive activity; and a mode of exploitation is a relationship of power. Furthermore, the power relationship which conditions the nature and extent of exploitation is a matter of political organization within and between the contending classes. In the final analysis, the relation between appropriators and producers rests to a great extent on the relative strength of classes, and this is largely determined by the internal organization and the political forces with which each enters into the class struggle. For example, as Robert Brenner has recently argued, the varying patterns of development in different parts of late medieval Europe can be accounted for in large part by the differences in class organization which characterized class struggles between lords and peasants in various places according to their specific historical experiences. In some cases, the struggle issued in a breakdown of the old order and old forms of surplus-extraction; in others, a retrenchment of the old forms took place. These different outcomes of agrarian class conflict, argues Brenner, ‘tended to be bound up with certain historically specific patterns of development of the contending agrarian classes and their relative strength in the different European societies: their relative levels of internal solidarity, their self-consciousness and organization, and their general political resources—especially their relationships to the non-agricultural classes (in particular, potential urban class allies) and to the state (in particular, whether or not the state developed as a ‘class-like’ competitor of the lords for the peasants’ surplus)’.  Robert Brenner, ‘Agrarian Class Structure and Economic Development in PreIndustrial Europe’, Past and Present, no. 70, p. 52.
Brenner goes on to illustrate how the particular form as well as the strength of political organization in the contending classes shaped relations of production: for example, how village institutions acted as a form of peasant class organization and how the development of ‘independent political institutions in the village’  ibid., p. 57.—or the relative lack of such institutions—affected the exploitive relations between lord and peasant. In short, Brenner’s case affords an example of how political organization constitutes relations of production.
There are, then, at least two senses in which the juridical–political ‘sphere’ is implicated in the productive ‘base’. First, a system of production always exists in the shape of specific social determinations, the particular modes of organization and domination and the forms of property in which relations of production are embodied—what might be called the ‘basic’ as distinct from ‘superstructural’ juridical–political attributes of the productive system. Second, from an historical point of view even political institutions like village and state enter directly into the constitution of productive relations and are in a sense prior to them (even where these institutions are not the direct instruments of surplus-appropriation) to the extent that relations of production are historically constituted by the configuration of political power that determines the outcome of class conflict.
The ‘Economic’ and the ‘Political’ in Capitalism
What, then, does it mean to say that capitalism is marked by a unique differentiation of the ‘economic’ sphere? It means several things: that production and distribution assume a completely ‘economic’ form, no longer (as Karl Polanyi put it) ‘embedded’ in extra-economic social relations,  The Great Transformation, Boston 1957, pp. 57, 69–71. in a system where production is generally production for exchange; that the allocation of social labour and the distribution of resources are achieved through the ‘economic’ mechanism of commodity exchange; that the ‘economic’ forces of the commodity and labour markets acquire a life of their own; that, to quote Marx, property ‘. . . receives its purely economic form by discarding all its former political and social embellishments and associations.’  Capital III, p. 618. Above all, it means that the appropriation of surplus labour takes place in the ‘economic’ sphere by ‘economic’ means. In other words, surplus-appropriation is achieved in ways that are determined by the complete separation of the producer from the conditions of labour and by the appropriator’s absolute private property in the means of production. Direct ‘extra-economic’ pressure or overt coercion are, in principle, unnecessary to compel the expropriated labourer to give up this surplus-labour. Although the coercive force of the ‘political’ sphere is ultimately necessary to sustain private property and the power of appropriation, ‘economic’ need supplies the immediate compulsion that forces the worker to transfer surplus labour to the capitalist in order to gain access to the means of production. The labourer is, therefore, ‘free’, not in a relationship of dependence or servitude; the transfer of surplus labour and its appropriation by someone else are not conditioned by such an ‘extra-economic’ relationship. The forfeit of surplus labour is an immediate condition of production itself. Capitalism in these respects differs from precapitalist forms to the extent that the latter are characterized by ‘extra-economic’ modes of surplus extraction, political, legal, or military coercion, traditional bonds or duties, etc., which demand the transfer of surplus labour to a private lord or to the state by means of labour services, rent, tax, and so on.
The differentiation of the economic sphere in capitalism, then, can be summed up as follows: the social functions of production and distribution, surplus extraction and appropriation, and the allocation of social labour are, so to speak, privatized and they are achieved by ‘non-authoritative’, non-political means. In other words, the social allocation of resources and labour does not, on the whole, take place by means of political direction, communal deliberation, hereditary duty, custom, or religious obligation, but rather through the mechanisms of commodity exchange. The powers of surplus appropriation and exploitation do not rest directly on relations of juridical or political dependence but are based on a contractual relation between ‘free’ producers—juridically free and free from the means of production—and an appropriator who has absolute private property in the means of production.
To speak of the differentiation of the economic sphere in these senses is not, however, to suggest that the political dimension is somehow extraneous to capitalist relations of production. The political sphere in capitalism has a special character to the extent that the coercive power supporting capitalist exploitation is not wielded directly by the appropriator and is not based on the producer’s political or juridical subordination to an appropriating master. Nevertheless, a coercive power and a structure of domination remain essential aspects of this exploitive relation, even if the ostensible freedom and equality of the exchange between capital and labour mean that the ‘moment’ of coercion is separate from the ‘moment’ of appropriation. Absolute private property, the contractual relation that binds producer to appropriator, the process of commodity exchange—all these require the legal forms, the coercive apparatus, the policing functions of the state. Historically, too, the state has been essential to the process of expropriation that is the basis of capitalism. In all these senses, the ‘economic’ sphere rests firmly on the ‘political’, despite their ‘differentiation’.
Furthermore, the economic sphere itself has a juridical and political dimension. In one sense, the differentiation of the economic sphere means simply that the economy has its own juridical and political forms whose purpose is purely ‘economic’. Absolute property, contractual relations, and the legal apparatus that sustains them are the juridical conditions of capitalist productive relations; and they constitute the basis of a new relation of authority, domination and subjection between appropriator and producer. The correlative of these ‘private’, ‘economic’ juridical-political forms is a separate specialized public political sphere. The ‘autonomy’ of the capitalist state is inextricably bound up with the juridical freedom and equality of the ‘free’, purely ‘economic’ exchange between ‘free’ expropriated producers and the private appropriators who have absolute property in the means of production and therefore a new form of authority over the producers. This is the significance of the division of labour in which the two moments of capitalist exploitation—appropriation and coercion—are allocated separately to a ‘private’ appropriating class and a specialized ‘public’ coercive institution, the state: on the one hand, the ‘relatively autonomous’ state has a monopoly of coercive force; on the other hand, that force sustains a private ‘economic’ power which invests capitalist property with an authority to organize production itself—an authority probably unprecedented in its degree of control over productive activity and the human beings who engage in it.
The direct political powers which capitalist proprietors have lost to the state they have gained in the direct control of production. While the ‘economic’ power of appropriation possessed by the capitalist is separated from the coercive political instruments that ultimately enforce it, that appropriative power is integrated more closely and directly than ever before with the authority to organize production. Not only is the forfeit of surplus labour an immediate condition of production, but capitalist property unites to a degree probably not enjoyed by any previous appropriating class the power of surplus extraction and the capacity to organize and intensify production—directly for the purposes of the appropriator. However exploitive earlier modes of production have been, however effective the means of surplus-extraction available to exploiting classes, in no other system has social production answered so immediately and universally to the demands of the exploiter.
At the same time, the powers of the appropriator no longer carry with them the obligation to perform social, public functions. Capitalism is a system marked by the complete separation of private appropriation from public duties; and this means the development of a new sphere of power devoted completely to private rather than social purposes. In this respect, capitalism differs from pre-capitalist forms in which the fusion of economic and political powers meant not only that surplus extraction was an ‘extra-economic’ transaction separate from the production process itself, but also that the power to appropriate surplus labour—whether it belonged to the state or to a private lord—was bound up with the performance of military, juridical, and administrative functions.
In a sense, then, the differentiation of the economic and the political in capitalism is, more precisely, a differentiation of political functions themselves and their separate allocation to the private economic sphere and the public sphere of the state. This allocation reflects the separation of political functions immediately concerned with the extraction and appropriation of surplus labour from those with a more general, communal purpose. Such a reformulation, suggesting that the differentiation of the economic is in fact a differentiation within the political sphere, is in certain respects better suited to explain the unique process of Western development and the special character of capitalism. It may, therefore, be useful to sketch this historical process of differentiation before looking more closely at the capitalist organization of production and its implications for the relation between the ‘economic’ and the ‘political’.
The Historical Process of Differentiation: Class Power and State Power
If the evolution of capitalism is viewed as a process in which an ‘economic’ sphere is differentiated from the ‘political’, it follows that an explanation of that evolution entails a theory of the state and its development. For the purposes of this discussion, the state will be defined in very broad terms as ‘. . . the complex of institutions by means of which the power of the society is organized on a basis superior to kinship’  Morton Fried, The Evolution of Political Society, New York, 1968, p. 229.—an organization of power which means a claim ‘to paramountcy in the application of naked force to social problems’ and consists of ‘formal, specialized instruments of coercion’.  Fried, p. 230. These instruments of coercion may or may not be intended from the outset as a means for one section of the population to oppress and exploit the rest. In either case, the state must have as an essential feature the performance of certain common social functions which other less comprehensive institutions—households, clans, kinship groups, etc—cannot carry out. Whether or not the essential object is the maintenance of exploitation, however, the performance of social functions by the state implies a social division of labour and the appropriation by some social groups of surplus produced by others. It seems reasonable to suppose, then, that whatever the various specific processes which bring this ‘complex of institutions’ into being, the state emerges as a means of appropriating surplus-product—perhaps even as a means of intensifying production in order to increase surplus—and as a mode of distributing that surplus in one way or another. In fact, it may be that the state—or some form of public power—is the first systematic means of surplus-appropriation and perhaps even the first systematic organizer of surplus-production.  See, Marshall Sahlins, Stone Age Economics, London 1974, chaps. 2 and 3, for some illuminating suggestions about how a public authority might emerge as a means of intensifying production.
While this conception of the state implies that the evolution of a specialized, coercive public authority necessarily entails a division between producers and appropriators, it does not mean that private appropriation is a necessary pre-condition to the emergence of such an authority. The two may develop together, and a long historical process may intervene before private appropriation clearly dissociates itself from public power. Propositions about the relation between class and state must, therefore, be cautiously formulated. It may be misleading to suggest, as Marxist arguments often seem to do, that there is a universal sequence of development in which class precedes state. What can perhaps be said is that, whichever came first, the existence of a state has always implied the existence of classes—although this proposition requires a definition of class which can encompass all divisions between direct producers and the appropriators of their surplus labour, even cases in which economic power is scarcely distinguishable from political power, where private property remains undeveloped, and where class and state are, in effect, one.  Problems may arise out of such inclusive definitions of class, not the least of which are their implications for analysis of modern Soviet-type states, which can be analysed as autonomous from class (e.g. in the work of Rudolf Bahro) or as a particular form of class organization (e.g. Göran Therborn, What Does the Ruling Class do when it Rules?, London 1978). Consideration of such alternatives is beyond the scope of this article. The essential point is to recognize that some of the major divergences among various historical trajectories have to do with the nature and sequence of relations between public power and private appropriation.
This point is especially important in identifying the distinctive characteristics of the historical path that has led to capitalism, with its unprecedented degree of differentiation between the ‘economic’ and the ‘political’. It is possible that the long historical process which ultimately issued in capitalism should be seen as an increasing—and uniquely well-developed—differentiation of class power as something distinct from state power, a power of surplus-extraction not directly grounded in the coercive apparatus of the state. This would also be a process in which private appropriation is increasingly divorced from the performance of communal functions. If we are to understand the unique development of capitalism, then, we must understand how property and class relations, as well as the functions of surplus-appropriation and distribution, so to speak liberate themselves from—and yet are served by—the coercive institutions that constitute the state, and develop ‘autonomously’.
The foundations of this argument are to be found in Marx’s discussion of pre-capitalist formations and the distinctive character of capitalism in the Grundrisse and Capital, especially volume III. In the Grundrisse Marx discusses the nature of capitalism in contrast to, and as a development from, pre-capitalist forms in terms of the gradual separation of the direct producer from the natural conditions of his labour. It is characteristic of pre-capitalist forms that producers remain, in one way or another, directly related to the conditions of labour, at least as possessors if not owners of the means of production. The only case in which the direct producer is completely expropriated—the case of chattel slavery—is itself determined by the typically direct relation of the producer to the natural conditions of his labour, since the slave is seized as an accessory to captured land, rendered propertyless by military means and thus himself transformed into a mere condition of production. Where a division between appropriators and producers has evolved, therefore, surplus-appropriation takes ‘extra-economic’ forms—whether it be the outright coercion of master against slave, or, where the labourer remains in possession of the conditions of his labour, a relationship of lordship and servitude in other forms. In one of the major pre-capitalist cases—the ‘Asiatic’—the state itself is the direct appropriator of surplus labour from producers who remain in possession of the land which they work. It is the special characteristic of capitalism that surplus appropriation and the relationship between the direct producer and the appropriator of his surplus labour do not take the form of direct political domination and legal servitude; and the authority confronting the mass of direct producers appears ‘. . . only as the personification of the conditions of labour in contrast to labour, and not as political or theocratic rulers as under earlier modes of production.’  Capital III, p. 881.
It is in this discussion of pre-capitalist forms and their ‘political’ modes of surplus-extraction, in both the Grundrisse and Capital, that Marx’s ill-fated conception of Asiatic societies makes its appearance. This is not the place for a full-scale debate on this contentious issue. For the moment, what is important is that in his discussion of ‘Asiatic’ forms Marx considers social types in which the state is the direct and dominant means of surplus-appropriation. In a sense, then, the ‘Asiatic’ type represents the polar opposite of the capitalist case, to the extent that economic and extra-economic, class power and state power, property relations and political relations, are least differentiated: ‘Should the direct producers not be confronted by a private land-owner, but rather, as in Asia, under direct subordination to a state which stands over them as their landlord and simultaneously as sovereign, then rent and taxes coincide, or rather, there exists no tax which differs from this form of ground-rent. Under such circumstances, there need exist no stronger political or economic pressure than that common to all subjection to that state. The state is then the supreme lord. Sovereignty here consists in the ownership of land concentrated on a national scale. But, on the other hand, no private ownership of land exists, although there is both private and common possession and use of land.’  Capital III, p. 791.
Even if there has never been a perfect representative of this social type—for example, if there has never been a well-developed appropriating and redistributive state in the complete absence of private ownership of land, as some have argued—the concept must still be taken seriously. There can be no doubt, to begin with, that the state as the major and direct appropriator of surplus-labour has existed; and there is considerable evidence that this mode of surplus appropriation has been a dominant, if not universal, pattern of social development, both in the West (for example, in Bronze Age Greece) and in the East, where the palace-dominated ‘redistributive’ economy, administered by a complex bureaucratic apparatus, has in various forms appeared as a widespread type of early social organization. Whatever other characteristics Marx may have attributed to the ‘Asiatic’ form, this one—which has sparked the most controversy—needs to be explored for what it may reveal about the process of differentiation which concerns us here.
The implication of Marx’s argument is that the division between appropriators and producers—a division which is, indeed, implied by any form of state—can take different forms, forms to which the notion of ‘class’ can be applied only with great caution to the extent that there is no clearly differentiated ‘economic’ power. It is, of course, true that only in capitalist society is the ‘economic’ power of class completely differentiated out from ‘extra-economic’ powers; and there is no intention here of arguing that there is class only in capitalist social formations. Nevertheless, it does at least seem important to recognize the polar extremes: the capitalist mode, in which the differentiation has taken place, and one in which—as in the certain bureaucratic, palace-dominated ‘redistributive’ states of the ancient world—the state itself, as the major direct appropriator of surplus product, is both ‘class’ and ‘state’ at once. Marx sometimes appears to suggest that, in the latter case, the dynamic of history is inhibited, if property and class do not break free and develop autonomously from the ‘hypertrophied’ state. To speak here of an ‘inhibited’ historical process may, however, be misleading, insofar as it implies that the course of development leading to capitalism—which Marx traces from ancient GraecoRoman civilization through Western feudalism to capitalism—has been the rule rather than the exception in world history and that all other historical experiences have been aberrations. Since Marx’s primary object is to explain the unique development of capitalism in the West, and not its ‘failure’ to evolve ‘spontaneously’ elsewhere, his project itself implies that—despite some apparently ‘ethnocentric’ assumptions—for him it is the achievement not the ‘failure’ which must be accounted for. At any rate, the particular dynamic of the ‘Asiatic’ form—as Marx’s argument implies—may be more ‘typical’ than the movement set in train by the ancient, Graeco-Roman form. If the primitive state was the controller of economic resources and the major appropriator and distributor of surplus product, the advanced ‘Asiatic’ state may represent a more or less natural development out of that primitive form—the appropriating redistributive public power at its highest stage of development. Seen in this light, it is not so much the ‘hypertrophy’ of the ‘Asiatic’ state that needs to be explained. What requires explanation is the aberrant, uniquely ‘autonomous’ development of the economic sphere that eventually issued in capitalism.  Ernest Mandel has criticized writers like Maurice Godelier for extending the meaning of the ‘Asiatic mode of production’ to include both social formations in the process of transition from classless society to class state and advanced bureaucratic empires with ‘hypertrophied’ states. (Mandel, The Formation of the Economic Thought of Karl Marx, London 1971, pp. 124 ff.) While Mandel is correct to warn against obscuring the differences between, say, simple African kingdoms and complex states like that of ancient Egypt, Godelier’s formulation is intended to emphasize the continuities between early forms of appropriative and redistributive public authorities and the advanced ‘hypertrophied’ state in order to stress that it is the Western case, with its ‘autonomous’ development of private property and class, which needs to be explained. Mandel often talks about the development of capitalism as if it were natural, while other historical trajectories have been stunted or obstructed.
Feudalism and Private Property
The capitalist organization of production, then, can be viewed as the outcome of a long process in which certain political powers were gradually transformed into economic powers and transferred to a separate ‘sphere’. The organization of production under the authority of capital presupposes the organization of production and the assembling of a labour force under the authority of earlier forms of private property. The process by which this authority of private property asserted itself, uniting the power of appropriation with the authority to organize production in the hands of a private proprietor for his own benefit, can be viewed as the privatization of political power. The supremacy of absolute private property appears to have established itself to a significant extent by means of political devolution, the assumption by private proprietors of functions originally invested in a public or communal authority. Again, the opposition of the ‘Asiatic’ mode of production at one extreme and the capitalist mode at the other helps to place this devolutionary process in perspective. From this point of view, the crucial issue is not the presence or absence of private property in land as such. China, for example, had well-established private landed property from a very early stage; and in any case, some form of property in land was often a perquisite of office in the ‘Asiatic’ state. The important point is the relation between private property and political power, and its consequences for the organization of production and the relation between appropriator and producer. The unique characteristic of Western development in this respect is that it is marked by the earliest and most complete transfer of political power to private property, and therefore also the most thorough, generalized, and direct subservience of production to the demands of an appropriating class.
Western feudalism represents an historic turning point that illuminates the whole process. Feudalism is often described as a fragmentation of ‘parcellization’ of state power; but while this description certainly identifies an essential characteristic, it is not specific enough. Forms of state power vary, and different forms of state power are likely to be differently fragmented. Western feudalism resulted from the fragmentation of a very particular form of political power. It is not here simply a matter of fragmentation or parcellization but also of privatization. The state power whose fragmentation produced Western feudalism had already been substantially privatized, located in private property. That is, the form of imperial administration that preceded the advent of feudalism in the West, built upon the foundations of a state already grounded in private property and class rule, was unique in that imperial power was exercised not so much through a hierarchy of bureaucratic officials in the manner of the ‘Asiatic’ state, but through local private proprietors whose property endowed them with political authority as well as the power of surplus appropriation.
This mode of administration was associated with a particular kind of relationship between appropriators and producers, especially in the Western Empire where there were no remnants of an older redistributive-bureaucratic state organization. The relationship between appropriators and producers was in principle a relationship between individuals, the owners of private property and the individuals whose labour they appropriated, the latter directly subject to the former. Even if in practice village communities played an important role in the organization of production and the landlord’s control over production was often indirect and tenuous, this still represents a significant contrast to the ‘Asiatic’ form in which typically whole villages of personally free producers were collectively and, so to speak, impersonally subject to the appropriating state which acted through the medium of its officials.
With the dissolution of the Roman Empire, the imperial state was in effect broken into fragments in which political and economic powers were united in the hands of private lords whose political, juridical, and military functions were at the same time the instruments of private appropriation and the organization of production. The decentralization of the Imperial state was accompanied by the decline of chattel slavery and its replacement by new forms of dependent labour. Slaves and formerly independent peasants began to converge toward conditions of dependence that derived their special character from the fact that the economic relationship between individual private appropriator and individual producer was at the same time a political relationship between a ‘fragment’ of the state and its subject. In other words, each basic ‘fragment’ of the state was at the same time a productive unit in which production was organized under the authority and for the benefit of a private proprietor. Although in comparison to the later developments of capitalism the power of the feudal lord to direct production remained far from complete, a considerable step had been taken toward the integration of surplus-extraction and the organization of production.  Cf. Rodney Hilton’s discussion of the limited control exercised in practice by feudal lords over the productive process, in ‘A Crisis of Feudalism’, Past and Present, no. 80, August 1978, pp. 9–10. It should be noted, however, that in stressing the limited nature of feudal lordship, Hilton is not comparing feudalism to other pre-capitalist formations, but at least implicitly, to capitalism, where the appropriator’s direct control of production is more complete because of the expropriation of the direct producer, the collective and concentrated nature of capitalist production, and so on.
The fact that the property of the feudal lord was not absolute but conditional private property does not alter the fact that feudalism represents a great advance in the authority of private property. In fact, the conditional nature of feudal property was in a sense a hallmark of its strength, not a sign of weakness, since the condition on which the lord held his land was that he must become, so to speak, a fragment of the state invested with the very functions that gave him the power of surplus-extraction. The coincidence of the political unit with the unit of property meant also a greater coincidence between the unit of appropriation and the unit of production, so that production could be organized more directly in the interests of the private appropriator. Furthermore, the fragmentation of the state, the fact that feudal relations were at once a method of governing and a mode of exploitation, meant that many free farmers now became subject with their properties to private masters, forfeiting surplus labour in exchange for personal protection, in a relationship of dependence that was both political and economic. As many more independent producers were brought into dependence, more production fell within the scope of direct, personal exploitation and class relations. The particular nature of the exploitive relation in feudalism and the ‘fragmentation’ of the state also, of course, affected the configuration of class power, eventually making it both more desirable—in some respects, even necessary—and more possible for private appropriators to expropriate direct producers.
The essential characteristic of feudalism, then, was a privatization of political power which meant a growing integration of private appropriation with the authoritative organization of production. The eventual development of capitalism out of the feudal system in a sense perfected this privatization and integration—by the complete expropriation of the direct producer and the establishment of absolute private property. At the same time, these developments had as their necessary condition a new and stronger form of centralized public power. The state divested the appropriating class of direct political powers and duties not immediately concerned with production and appropriation, leaving them with private exploitative powers purified, as it were, of public, social functions.
Capitalism as the Privatization of Political Power
It may seem perverse to suggest that capitalism represents the ultimate privatization of political power, a characteristic usually ascribed to feudalism. This proposition on the face of it runs directly counter to the description of capitalism as uniquely characterized by a differentiation of the economic and the political. The intention of this description is, among other things, precisely to contrast capitalism to the ‘parcellization’ of state power which unites private political and economic power in the hands of the feudal lord. It is, after all, capitalism that is marked not only by a specialized economic sphere and economic modes of surplus-extraction but also by a central state with an unprecedented public character. Capitalism is uniquely capable of maintaining private property and the power of surplus-extraction without the proprietor wielding direct political power in the conventional sense. The state—which stands apart from the economy even though it intervenes in it—can ostensibly (notably, by means of universal suffrage) belong to everyone, producer and appropriator, without usurping the exploitive power of the appropriator. The expropriation of the direct producer simply makes certain direct political powers less immediately necessary to surplus-extraction. This is precisely what it means to say that the capitalist has economic rather than extra-economic powers of exploitation.
It can even be argued that overcoming the ‘privatization’ of political power is one of the essential conditions for the transformation of the labour-process and the forces of production which is the distinguishing characteristic of capitalism. For example, as Robert Brenner has argued, ‘where the direct application of force is the condition for ruling-class surplus-extraction, the very difficulties of increasing productive potential through the improvement of the productive forces may encourage the expenditure of surplus to enhance precisely the capacity for the application of force. In this way, the ruling class can increase its capacity to exploit the direct producers, or acquire increased means of production (land, labour, tools) through military methods. Rather than being accumulated, the economic surplus is here systematically diverted from reproduction to unproductive labour.’  Brenner, ‘The Origins of Capitalism’, nlr 104, p. 37.
On the other hand, there is another sense in which private ‘political’ power is an essential condition of capitalist production and is, in fact, the form assumed by the ‘autonomy’ of the economic sphere. The capitalist is, of course, subject to the imperatives of accumulation and competition which oblige him to expand surplus-value; and the labourer is bound to the capitalist not simply by the latter’s personal authority but by the laws of the market which dictate the sale of labour-power. In these senses, it is the ‘autonomous’ laws of the economy and capital ‘in the abstract’ which exercise power, not the capitalist willfully imposing his personal authority upon labour. Nevertheless, what the ‘abstract’ laws of capitalist accumulation compel the capitalist to do—and what the impersonal laws of the labour-market enable him to do—is precisely to exercise an unprecedented degree of control over production. ‘The law of capitalist accumulation, metamorphosed by economists into pretended (sic) law of nature, in reality merely states that the very nature of accumulation excludes every diminution in the degree of exploitation’;  Capital I, p. 582. and this means firm command of the labour process, even an internal legal code, to ensure the reduction of necessary labour-time and the production of maximum surplus value within a fixed period of work. The need for a ‘directing authority’, as Marx explains, is intensified in capitalist production both by the highly socialized, cooperative nature of production—a basic condition of its high productivity—and by the antagonistic nature of an exploitive relationship based on the demand for maximum extraction of surplus value.
Capitalist production truly begins, argues Marx, only when the subjection of labour to capital becomes ‘real’ rather than simply ‘formal’—that is, when this subjection actually transforms the labour-process itself. This occurs ‘when each individual capital employs simultaneously a comparatively large number of labourers; when consequently the labour-process is carried on on an extensive scale and yields, relatively, large quantities of products. A greater number of labourers working together at the same time, in one place (or, if you will, in the same field of labour), in order to produce the same sort of commodity under the mastership of one capitalist, constitutes, both historically and logically, the starting point of capitalist production.’  Capital I, p. 305. ‘Working together’ means in an integrated and cooperative division of labour, not just working under one roof.
A fundamental condition of this transformation is capital’s control of the labour-process. In other words, capitalist production, with its cooperative, collective character, begins when direct ‘political’ power is introduced into the production process itself, as a basic condition of production: ‘By the cooperation of numerous wage-labourers, the sway of capital develops into a requisite for carrying on the labour-process itself, into a real requisite of production. That a capitalist should command on the field of production is now as indispensable as that a general should command on the field of battle.’  Capital I, p. 313.
In pre-capitalist societies, cooperative production was simple and sporadic, though sometimes it had, as Marx puts it, ‘colossal effects’—for example, under the command of Asiatic and Egyptian kings or Etruscan theocrats. The special characteristic of capitalism is its systematic and continuous cooperative production. The political significance of this development in production is expressed by Marx himself: ‘This power of Asiatic and Egyptian kings, Etruscan theocrats, etc., has in modern society been transferred to the capitalist, whether he be an isolated, or as in joint-stock companies, a collective capitalist.’  Capital I, p. 316. Emphasis added.
The issue here is not whether capitalist control is more ‘despotic’ than the harsh personal authoritarianism of the slave-driver with whip in hand; nor whether capitalist exploitation is more oppressive than the demands of a rent-hungry feudal lord. The degree of control exercised by capital over production is not necessarily dependent upon its degree of ‘despotism’. To some extent, control is imposed not by personal authority but by the impersonal exigencies of machine production and the technical integration of the labour-process (though this can be exaggerated and, in any case, the need for technical integration is itself to a great extent imposed by the compulsions of capitalist accumulation and the demands of the appropriator). Furthermore, while capital, with its absolute property in the means of production, has at its disposal new forms of purely ‘economic’ coercion—such as the power to close plants—the nature of its control of the labour-process is in part conditioned by its lack of direct coercive force. In a sense, the intricate and hierarchical organization and supervision of the labour-process as a means of increasing surplus in production is a substitute for a coercive power of surplus-extraction. The nature of the free working class is also such that new forms of workers’ organization and resistance are built into the production process. In any case, capitalist control, in different circumstances, can be exercised in ways ranging from the most ‘despotic’ organization (e.g. ‘Taylorism’) to varying degrees of ‘workers’ control’ (though the pressures against the latter inherent in the structure of capitalist accumulation should not be underestimated). Whatever specific forms capitalist control may take, however, its essential conditions remain: in no other system of production is work so thoroughly disciplined and organized, and no other organization of production is so directly responsive to the demands of appropriation.
There are, then, two critical points about the capitalist organization of production which help to account for the peculiar character of the ‘political’ in capitalist society and to situate the economy in the political arena: first, the unprecedented degree to which the organization of production is integrated with the organization of appropriation; and second, the scope and generality of that integration, the virtually universal extent to which production in society as a whole comes under the control of the capitalist appropriator.  Chattel slavery is the pre-capitalist form of class exploitation in which it might be most convincingly argued that the exploiter exercises a continuous and direct control over production; but leaving aside the many questions surrounding the nature and degree of the slave-owner’s control of the labour-process, one thing is clear: that even among the very few societies in which slavery has been widespread, it has never come close to the generality of wage-labour in advanced capitalist societies but has always been accompanied, and possibly exceeded, by other forms of production. For example, in the Roman Empire, where ancient slavery reached its culmination in the slave latifundia, peasant producers still outnumbered slaves. Even if independent producers were subject to various forms of surplus-extraction, large sections of production remained outside the scope of direct control by an exploiting class. It can be argued, too, that this was not accidental; that the nature of slave-production made its generalization impossible; that not the least obstacle to its further expansion was its dependence on direct coercion and military power; and that, conversely, the uniquely universal character of capitalist production and its capacity to subordinate virtually all production to the demands of exploitation is inextricably bound up with the ‘differentiation of the economic and the political’. The corollary of these developments in production is that the appropriator relinquishes direct political power in the conventional, public sense, and loses many of the traditional forms of personal control over the lives of labourers outside the immediate production process which were available to pre-capitalist appropriators. New forms of indirect class control pass into the ‘impersonal’ hands of the state. At the same time, if capitalism—with its juridically free working class and its impersonal economic powers—removes many spheres of personal and social activity from direct class control, it can be argued that human life generally is drawn more firmly than ever into the impersonal orbit of the production process. Directly or indirectly, the demands and discipline of capitalist production, imposed by the exigencies of capitalist appropriation and accumulation, bring within their sphere of influence—and thus under the sway of capital—an enormous range of activity and exercise an unprecedented control over the disposal of time, within and without the production process.
These developments betoken the existence of a differentiated ‘economic’ sphere and ‘economic laws’. Their full significance, however, may be obscured by viewing them only in this light. It is at least as important to regard them as a transformation of the political sphere. In one sense, the integration of production and appropriation represents the ultimate ‘privatization’ of politics, to the extent that functions formerly associated with a coercive political power—centralized or ‘parcellized’—are now firmly lodged in the private sphere, as functions of a private appropriating class relieved of obligations to fulfil larger social purposes. In another sense, it represents the expulsion of politics from sphere in which it has always been directly implicated. Direct political coercion is excluded from the process of surplus-extraction and removed to a state that generally intervenes only indirectly in the relations of production; and surplus-extraction ceases to be an immediately political issue. This means that the focus of class struggle necessarily changes. As always, the disposition of surplus labour remains the central issue of class conflict; but now, that issue is no longer distinguishable from the organization of production. The struggle over appropriation appears not as a political struggle but as a battle over the terms and conditions of work.
The Localization of Class Struggle
To the extent that surplus-extraction and appropriation—not production—have been the central issues in class struggle, capitalism is unique in its concentration of class struggle ‘at the point of production’, precisely because it is only in capitalism that the organization of production and of appropriation so completely coincide. It is also unique in its transformation of struggles over appropriation into apparently non-political contests. For example, while the wage-struggle in capitalism may be perceived as merely ‘economic’ (‘economism’), the same is not true of the rent-struggle waged by medieval peasants, even though the issue in both cases is the disposition of surplus labour and its relative distribution between direct producers and exploiting appropriators. However fierce the struggle over wages may be, the wage-relationship itself—as Marx points out—remains intact: the basis of the appropriator’s extractive powers—the status of his property and the propertylessness of the labourer—are not immediately at stake. Struggles over rent, to the extent that appropriation rests on ‘extra-economic’ powers, tend more immediately to implicate property rights, political powers and jurisdictions.
Class conflict in capitalism is, therefore, to a considerable extent encapsulated within the individual unit of production; and this gives class struggle a special character.  This and the following two paragraphs are based on my article ‘For Sale: Strategies for Winning the Class Struggle’, concerning management consultant seminars on the prevention of unions, in Canadian Dimension, vol. 14, no. 1, July-August, 1979, pp. 13–16. Each individual plant, a highly organized and integrated unity with its own hierarchy and structure of authority, contains within it the main sources of class conflict. At the same time, class struggle enters directly into the organization of production: that is, the management of antagonistic relations of production is inseparable from the management of the production process itself. While class conflict remains an integral part of the production process which it must not disrupt, class struggle must in a sense be domesticated.
Class conflict, therefore, generally breaks into open war only when it goes outdoors, particularly since the coercive arm of capital is outside the wall of the productive unit. This means that when there are violent confrontations, they are usually not directly between capital and labour. It is not capital itself but the state that conducts class conflict when it intermittently breaks outside the walls and takes more violent form. The armed power of capital thus usually remains in the background; and when class domination makes itself felt as direct and personal coercive force, it appears in the guise of an ‘autonomous’ and ‘neutral’ state.
The transformation of ‘political’ into ‘economic’ conflicts and the location of struggles ‘at the point of production’ also tend to make class struggle in capitalism local and particularistic. In this respect, the organization of capitalist production itself resists the working class unity which capitalism is supposed to encourage. On the one hand, the nature of the capitalist economy—its national, even supra-national, character, the interdependence of its constituent parts, the homogenization of work produced by the capitalist labour-process—make both necessary and possible a working class consciousness and class organization on a mass scale. This is the aspect of capitalism’s effects on class-consciousness that Marxist theory has so often emphasized. On the other hand, the development of this consciousness and this organization must take place against the centrifugal force of capitalist production and its ‘privatization’ of political issues.
The consequences of this centrifugal effect—if not sufficiently accounted for by theories of class-consciousness—have often been remarked upon by observers of industrial relations who have noted the growing rather than declining importance of ‘domestic’ struggles in contemporary capitalism. While the concentration of working class battles on the domestic front may detract from the political and universal character of these struggles, it does not necessarily imply a declining militancy. The paradoxical effect of capitalism’s differentiation of the economic and the political is that militancy and political consciousness have become separate issues. It is worth considering, by contrast, that modern revolutions have tended to occur where the capitalist mode of production has been less developed; where it has coexisted with other modes of production—notably peasant production; where ‘extra-economic’ compulsion has played a greater role in the organization of production and the extraction of surplus labour; and where the state has acted not only as a support for appropriating classes but as a ‘precapitalist’ appropriator in its own right—in short, where economic struggle has been inseparable from political conflict and where the state—as a more visibly centralized and universal class enemy—has served as a focus for mass struggle. Even in more developed capitalist societies, mass militancy tends to emerge in response to ‘extra-economic’ compulsion, particularly in the form of oppressive action by the state, and also varies in proportion to the state’s involvement in struggles over the terms and conditions of work.
These considerations again raise questions about the sense in which it is appropriate to regard working class ‘economism’ in advanced capitalist societies as reflecting an ‘undeveloped’ state of class-consciousness, as many socialists do. Seen from the perspective of historical process, it can be said to represent a more, rather than a less, advanced stage of development. If this stage is to be surpassed in turn, it is important to recognize that, in a sense, the so-called ‘economism’ of working class attitudes does not so much reflect a lack of political consciousness as an objective shift in the location of ‘politics’, a change in the arena and the objects of political struggle inherent in the very structure of capitalist production.
These are some of the ways in which capitalist production tends to transform ‘political’ into ‘economic’ struggles. There are, it is true, certain trends in contemporary capitalism which may work to counteract these tendencies. The national and international integration of the advanced capitalist economy increasingly shifts the problems of capitalist accumulation from the individual enterprise to the ‘macro-economic’ sphere. It is possible that capital’s powers of appropriation—which the state has so far left intact, indeed reproduced and reinforced—will be subverted by capital’s own growing need for the state—not only to facilitate capitalist planning, to assume liabilities or to conduct and contain class conflict, but also to perform the social functions abandoned by the appropriating class, indeed to counteract its anti-social effects. At the same time, insofar as capital in its mounting crises demands, and obtains, the state’s complicity in its anti-social purposes, that state may increasingly become a prime target of resistance in advanced capitalist countries—as it has been in every successful modern revolution. The effect of this may be to overcome the particularism and the ‘economism’ imposed on the class struggle by the capitalist system of production, with its differentiation of the economic and the political.
In any case, the strategic lesson to be learned from the transfer of ‘political’ issues to the ‘economy’ is not that class struggles ought to be primarily concentrated in the economic sphere or ‘at the point of production’. Nor does the division of ‘political’ functions between class and state mean that power in capitalism is so diffused throughout civil society that the state ceases to have any specific and privileged role as a locus of power and a target of political action—or, alternatively, that everything is the ‘state’. Indeed, the opposite is true. The division of labour between class and state means not so much that power is diffuse, but, on the contrary, that the state—which represents the coercive ‘moment’ of capitalist class domination, embodied in the most highly specialized, exclusive, and centralized monopoly of social force—is ultimately ‘the decisive point of concentration for all power in society.’  I owe this phrase to Perry Anderson.
Struggles at the point of production, then—even in their ‘economic’ aspects as struggles over the terms of sale of labour-power or over the conditions of production—remain incomplete as long as they do not extend to the locus of power on which capitalist property, with its control of production and appropriation, ultimately rests. At the same time, purely ‘political’ battles, over the power to govern and rule, remain unfinished until they implicate not only the institutions of the state but the political powers that have been ‘privatized’ and transferred to the economic sphere. In this sense, the very differentiation of the ‘economic’ and the ‘political’ in capitalism—the symbiotic division of labour between class and state—is precisely what makes the unity of ‘economic’ and ‘political’ struggles essential. If Marxist theory is to make good its practical claims and map the terrain of political action, it should not reinforce the false appearances which keep the two spheres apart.
[*] My thanks must go to several people who have read and criticized—often vehemently—this essay at various stages: Perry Anderson, Robin Blackburn, Robert Brenner, Ralph Miliband, Neal Wood, Gregory Meiksins, Peter Meiksins, and my students at York University, Toronto, especially Frances Abele and George Comninel.
 Karl Marx, Capital, Moscow vol. I, p. 668.
 Marx, Grundrisse, Harmondsworth 1973, p. 86.
 Ibid., p. 87.
 Ibid., p. 88.
 G. A. Cohen, Karl Marx’s Theory of History: A Defense, Oxford 1978.
 Cohen, pp. 152–3.
 Cohen, p. 98.
 For example, on p. 99 he cites Grundrisse, p. 304 and Capital III, p. 795 (Moscow, 1962. In the Moscow, 1971, edition, p. 815).
 See, for example, Capital III, Moscow 1971, p. 818.
 e.g., Grundrisse, pp. 85–88.
 loc. cit.
 Capital III, pp. 883–4.
 Capital I, p. 456–7.
 See Robert Brenner, “The Origins of Capitalism”, nlr, 304, for a discussion of how capitalist relations of production uniquely demand the revolutionary transformation of productive forces.
 Marx and Engels, German Ideology in Collected Works, New York 1976, vol. 5, p. 45.
 Engels, Letter to Conrad Schmidt, October 27, 1890, in Marx and Engels, Selected Correspondence, Moscow 1965, p. 422.
 Simon Clarke, ‘Socialist Humanism and the Critique of Economism’, History Workshop 8, Autumn, 1979, pp. 138–156. For a somewhat different kind of ‘unitarian’ approach, see, for example, Raymond Williams’ Marxism and Literature, Oxford 1977.
 Clarke, p. 144.
 Past and Present, no. 79, pp. 67–68. The author, Guy Bois, is referring specifically to the article by Robert Brenner mentioned below in note 23. It is worth noting that, while Bois attacks Brenner for his departure from economism, Simon Clarke does the opposite, criticizing Brenner for his ‘definition of relations of production in Dobbian terms as forms of exploitation’. Clarke, p. 155, n. 3. This simply reinforces my view that Brenner has got it just about right.
 See, for example, Weber’s Economy and Society, New York 1968, pp. 91 and 94, and Agrarverhältnisse im Altertum in Gesammelte Aufsätze zur Sozial- und Wirtschafts-geschichte Tubingen 1924, pp. 15–16. In English translation nlb, The Agrarian Sociology of Ancient Civilizations, pp. 50–1.
 German Ideology op. cit., p. 39.
 German Ideology, p. 40.
 Robert Brenner, ‘Agrarian Class Structure and Economic Development in PreIndustrial Europe’, Past and Present, no. 70, p. 52.
 ibid., p. 57.
 The Great Transformation, Boston 1957, pp. 57, 69–71.
 Capital III, p. 618.
 Morton Fried, The Evolution of Political Society, New York, 1968, p. 229.
 Fried, p. 230.
 See, Marshall Sahlins, Stone Age Economics, London 1974, chaps. 2 and 3, for some illuminating suggestions about how a public authority might emerge as a means of intensifying production.
 Problems may arise out of such inclusive definitions of class, not the least of which are their implications for analysis of modern Soviet-type states, which can be analysed as autonomous from class (e.g. in the work of Rudolf Bahro) or as a particular form of class organization (e.g. Göran Therborn, What Does the Ruling Class do when it Rules?, London 1978). Consideration of such alternatives is beyond the scope of this article.
 Capital III, p. 881.
 Capital III, p. 791.
 Ernest Mandel has criticized writers like Maurice Godelier for extending the meaning of the ‘Asiatic mode of production’ to include both social formations in the process of transition from classless society to class state and advanced bureaucratic empires with ‘hypertrophied’ states. (Mandel, The Formation of the Economic Thought of Karl Marx, London 1971, pp. 124 ff.) While Mandel is correct to warn against obscuring the differences between, say, simple African kingdoms and complex states like that of ancient Egypt, Godelier’s formulation is intended to emphasize the continuities between early forms of appropriative and redistributive public authorities and the advanced ‘hypertrophied’ state in order to stress that it is the Western case, with its ‘autonomous’ development of private property and class, which needs to be explained. Mandel often talks about the development of capitalism as if it were natural, while other historical trajectories have been stunted or obstructed.
 Cf. Rodney Hilton’s discussion of the limited control exercised in practice by feudal lords over the productive process, in ‘A Crisis of Feudalism’, Past and Present, no. 80, August 1978, pp. 9–10. It should be noted, however, that in stressing the limited nature of feudal lordship, Hilton is not comparing feudalism to other pre-capitalist formations, but at least implicitly, to capitalism, where the appropriator’s direct control of production is more complete because of the expropriation of the direct producer, the collective and concentrated nature of capitalist production, and so on.
 Brenner, ‘The Origins of Capitalism’, nlr 104, p. 37.
 Capital I, p. 582.
 Capital I, p. 305. ‘Working together’ means in an integrated and cooperative division of labour, not just working under one roof.
 Capital I, p. 313.
 Capital I, p. 316. Emphasis added.
 Chattel slavery is the pre-capitalist form of class exploitation in which it might be most convincingly argued that the exploiter exercises a continuous and direct control over production; but leaving aside the many questions surrounding the nature and degree of the slave-owner’s control of the labour-process, one thing is clear: that even among the very few societies in which slavery has been widespread, it has never come close to the generality of wage-labour in advanced capitalist societies but has always been accompanied, and possibly exceeded, by other forms of production. For example, in the Roman Empire, where ancient slavery reached its culmination in the slave latifundia, peasant producers still outnumbered slaves. Even if independent producers were subject to various forms of surplus-extraction, large sections of production remained outside the scope of direct control by an exploiting class. It can be argued, too, that this was not accidental; that the nature of slave-production made its generalization impossible; that not the least obstacle to its further expansion was its dependence on direct coercion and military power; and that, conversely, the uniquely universal character of capitalist production and its capacity to subordinate virtually all production to the demands of exploitation is inextricably bound up with the ‘differentiation of the economic and the political’.
 This and the following two paragraphs are based on my article ‘For Sale: Strategies for Winning the Class Struggle’, concerning management consultant seminars on the prevention of unions, in Canadian Dimension, vol. 14, no. 1, July-August, 1979, pp. 13–16.
 I owe this phrase to Perry Anderson.