Derek Sayer (1987)

The Violence of Abstraction
The Analytical Foundations of Historical Materialism

2 Productive Forces

Source: The Violence of Abstraction (1987) publ. Basil Blackwell, 1987. Just two Chapters out of six reproduced here.


One feature of Marx’s 1859 Preface has been remarked upon rather less often than it perhaps ought to be. The key concepts he employs there – forces and relations of production, economic structure and superstructure – are, for the most part, either left altogether undefined in the text itself, or else defined circularly, in terms of one another. No definition whatsoever is offered of material productive forces, that concept which for Cohen has explanatory primacy in Marx’s theory of history. Nor are we told what relations of production are, beyond Marx’s remark that property relations are ‘but a legal expression for the same thing’; itself a highly confusing formulation (and one that has spawned its own extensive literature) because of the many possible interpretations of ‘expression’. We glean only that ‘in the social production of their existence’, people inevitably enter into such relations, which are ‘independent of their will’.

The ‘economic structure of society’ – the ‘base’, the ‘real foundation’ – is defined circularly, as the sum total of production relations, which takes us no further. The Preface also invokes the ‘mode of production of material life’, which Marx tells us ‘conditions the general process of social, political and intellectual life’. But the text neither defines a mode of production, nor clarifies the relation of this concept to those of forces, relations or economic structure. Conventionally a mode of production is taken to comprise a particular – in modern Marxist jargon, an ‘articulated’ – combination of forces and relations,’ but Marx does not say so here. Cohen, among others, construes it otherwise, correctly noting a variety of uses of the term in Marx’s work (1978: 79-84).

Marx qualifies the ‘superstructure’ in the Preface as ‘legal and political’. He also refers, independently, to ‘definite forms of social consciousness’. He fails to make clear whether the latter are part of the superstructure, though his language in this instance suggests otherwise. This is what has led Cohen to argue they are not. Elsewhere, however, Marx – like Engels in the letters discussed above frequently describes the entire superstructure as ‘ideal’ or ‘ideological’, and explicitly includes forms of consciousness within it. Indeed, as we shall see in chapter 4, in some passages where the term superstructure is used we would be justified in concluding that its primary, if not exclusive, reference is to consciousness. The sentence later in the Preface itself, which distinguishes ‘the material transformation of the economic conditions of production’ from ‘the legal, political, religious, aesthetic or philosophic – in short, ideological forms in which men become conscious of this conflict and fight it out’ might be taken to support this latter interpretation. On the other hand, the reference in the preceding paragraph to ‘legal relations as well as forms of state’ having their roots in ‘the material conditions of life’ might incline us to the more restrictive institutional definition favoured by Cohen.

But again, Marx’s apparent curtailment of the scope of the category at this point in the Preface may have a simple explanation. He is here recapitulating a stage in his own intellectual biography, the critique of Hegel’s Rechtsphilosophie which he undertook in 1843. Hegel’s central concerns in that work were with law and state, and Marx embarked upon his critique to clarify his own views on the state following his personal encounters with Prussian officialdom as editor of the Rheinische Zeitung. The Hegelian background, however, muddies the waters still further with Marx’s introduction of the term ‘civil society’ as a synonym for the totality of ‘the material conditions of life’, whose ‘anatomy ... has to be sought in political economy’. Is ‘civil society’ as used here merely another expression for ‘economic structure’, or an intermediary level between base and superstructure? At all events, there is considerable room for uncertainty over what exactly, for the 1859 Preface, the ‘superstructure’ comprises – or excludes.

This problem of the meaning of the key terms in this ‘most pregnant’ summary of historical materialism is not, I believe, generally raised in commentary for a very simple reason. We all think we know what Marx means when he uses these concepts. But it should be clear by now that it is not the text itself which gives us warrant for this presumption. Nor is our interpretive confidence as a rule founded on any sort of scholarly examination of how Marx uses such concepts elsewhere, the study of the ‘theory’ in its concrete applications of the sort recommended by Engels. It is in fact simply the authority of the long tradition of ‘orthodox Marxism’ which has accustomed us to understand these concepts, without demur, in particular ways. With minor caveats, these are the ways advocated by Cohen and outlined above.

But much hangs on this apparently semantic question. Lucio Colletti, for instance, in his justly famous essay on the Marxism of the Second International – a Marxism not unlike Cohen’s – has argued that the orthodoxy of Kautsky and Plekhanov and the revisionism of Bernstein equally vulgarised Marx’s understanding of production itself. Engels’s ‘testament’, in his view, unwittingly succoured this vulgarisation. For although Engels argued against outright economic determinism, the way in which he did so – conceding the ‘inherent relative independence’, and therewith the separability, of superstructures from their economic bases – inadvertently reinforced an economistic, technologistic conception of the latter. Hiving off superstructures drained production itself of its social and cultural dimensions, for Colletti Marx’s crucial insight. In Colletti’s own words:

what Bernstein shared with Plekhanov, and what Engels’s ‘self-criticism’ could not correct but only confirm, was the profound adulteration of the concept of the ‘economy’, or, better still, of ‘social relations of production’, precisely the core and foundation of Marx’s entire work. The so-called ‘economic sphere’ – which in Marx had embraced both the production of things and the production (objectification) of ideas; production and inter-subjective communication; material production and the production of social relations (for Marx, the relation between man and nature was also a relationship between man and man and vice versa) – was now seen as one isolated factor, separated from the other ‘moments’ and thereby emptied of any effective socio-historical content, representing, on the contrary, an antecedent sphere, prior to any human mediation. Social production is thus transformed into ‘production techniques’; the object of political economy becomes the object of technology. Since this ‘technique’, which is ‘material production’ in the strict sense of the term, is separated from that other simultaneous production achieved by men, the production of their relations (without which, for Marx, the former would not exist), the materialist conception of history tends to become a technological conception of history. (1972: 65)

This may not be entirely fair to Engels, for other readings of his late letters can be defended. One may, for instance, argue that they are as conducive to an interpretation of superstructures as ideological forms of expression of class and other ‘basic’ relations, as they are to a view of superstructures as substantially separate ‘spheres’. But Colletti’s charge certainly holds for much subsequent Marxism. His argument, moreover, applies to Marxist practice as well as theory, for such analytic paradigms have been a far from negligible ingredient in the shaping of ruling communist party policies for economic development and social transformation in the USSR and elsewhere, with momentous consequences for socialist politics.

It is this line of criticism I shall pursue here, though I shall take it further, perhaps, than Colletti himself might wish. I shall try to show, from Marx’s substantive analyses as distinct from his general summaries, that conventional Marxist wisdom on the interpretation of his fundamental concepts is, quite simply, wrong. Marx had a very different understanding of what (and how much) was entailed in ‘the production and reproduction of real life’ than most of his disciples, and used all of these concepts – forces, relations, and so on – very differently from the ways ascribed to him by tradition and systematised and defended by Cohen. Additionally, and perhaps more radically, I shall suggest that ‘traditional historical materialism’ actually fetishises Marx’s concepts, in his own sense of that term. I shall begin, in this chapter, by looking at Marx’s concept of productive forces – or, to be more accurate, and not altogether pedantic, the social forces of production.


Before examining this and other key concepts of the Preface individually, however, a further preliminary observation is called for. This concerns what we may expect in Marx by way of conceptual definition as such. The Preface is not alone among Marx’s writings in its failure to provide clear definition of its terms. Marx was not a devotee of that tradition of twentieth-century analytic philosophy whose ‘standards of clarity and rigour’ Cohen seeks to bring to bear on historical materialism (1978: ix), and it is idle to ransack his work in search of neat and unambiguous definitions of his general concepts. Notoriously, for example, he never defined class, that central concept of his sociology. Such apparent laxity infuriates those of a philosophical cast of mind, and Cohen’s entire book can be seen as an attempt to remedy this supposed deficiency. But Marx’s way of using language in this instance – as with Engels’s penchant for elusive metaphor – may be indicative of something more substantial in his thought. Modern analytic philosophy, by the same token, might not turn out to be the most fortunate choice of framework in which to try and express his ideas.

Bertell Ollman (1976) has persuasively argued that Marx’s conceptual slipperiness – ‘words like bats’, as Pareto complained, appearing now like birds, now like mice – is evidence not of any lack of rigour on his part, but of his commitment to a specifically, and distinctively, dialectical ontology. Marx did not conceive social reality atomistically, as made up of clearly bounded, separate, interacting entities: the kind of analytic particulars which can be grasped in clear, consistent and exclusive definitions. He saw the world, rather, as a complex network of internal relations, within which any single element is what it is only by virtue of its relationship to others. In this Marx stood squarely in the philosophical tradition of Spinoza, Leibniz and of course Hegel, of which ‘mighty thinker’ he explicitly avowed himself ‘a pupil’ (after having written Capital (1873: 19-20). To take the most obvious example, neither wage-labour nor capital, for instance, can for Marx be defined ‘in themselves’, as autonomous particulars, conceivable independently of one another. Nor can they properly be understood as externally ‘interacting’ on one another. Each is what it is only by virtue of its relation to the other, and must be conceptualised accordingly. The concept of capital implicitly contains that of wage-labour, and vice versa.

What Marx held to be thus implicit in a single concept may be very extensive indeed. In the Grundrisse, for example, he remarks that ‘the simple concept of capital has to contain its civilising tendencies etc. in themselves; they must not, as in the economics books until now, appear merely as external consequences. Likewise the contradictions which are later released, demonstrated as already latent within it’ (1858a: 414). Hence – amongst other things – ‘the tendency to create the world market is directly given in the concept of capital itself (ibid.: 408). In the same way he observes that ‘the germs of crises, or at least their possibility’ are already implicit in the concept of money as medium of exchange (1858a: 198), and indeed in the elementary concept of the commodity itself (ibid.: 147 ff.; cf. 1867a: 114).

‘The simplest economic category’, like exchange-value, ‘presupposes [unterstellt: Ollman (1976: 12) translates this as ‘implies’] population, moreover a population producing in specific relations; as well as a certain kind of family, or commune, or state, etc. It can never exist other than as an abstract, one-sided relation within an already given, concrete, living whole’ (1857: 101). In the same text, elaborating on the proposition that ‘every form of production creates its own legal relations, forms of government, etc.’ – a proposition most Marxists would understand as a causal claim about the relation between essentially distinct entities – Marx berates economists for treating this relation as an external, contingent one. ‘In bringing things which are organically connected into an accidental relation, into a merely reflective connection’, he says, ‘they display their crudity and lack of conceptual understanding’ (ibid.: 88, emphasis added). Such examples could be multiplied.’

Within an internal relations perspective drawing boundaries to concepts – particularly to general concepts – is evidently going to be a problem. The problem is compounded when, as was also the case for Marx, the relations at issue are viewed as being in the process of constant formation and transformation. But if Ollman is right about Marx’s ontology this difficulty is unavoidable. It is not resolvable by definitional fiat, because it arises out of the very nature of the reality Marx’s concepts seek to define. Words must be ‘like bats’ if they are to be able to grasp this complexity. From the standpoint of this philosophy, one which differs in fundamentals from the whole analytic tradition, to use concepts otherwise would be singularly unrigorous, since it would entail systematically distorting reality. Indeed, as we shall see in more detail in chapter 4, Marx in fact regarded concepts and categories as a facet of social reality itself, and hence saw definitions as necessarily real rather than nominal. Engels, again, argued just this. In his Preface to Capital III he pointed out that we should not expect to find:

fixed, cut-to-measure, once and for all applicable definitions in Marx’s works. It is self-evident that where things and their interrelations are conceived, not as fixed, but as changing, their mental images, the ideas, are likewise subject to change and transformation, and they are not encapsulated in rigid definitions, but are developed in their historical or logical process of formation. (1894b: 13-14)

It is false, Engels says, to assume that Marx ‘wishes to define where he only investigates’.

An important corollary of Ollman’s argument – but a thesis which can also be independently defended on other grounds – is that Marx’s general, transhistorical categories (like those of the 1859 Preface) acquire substantive definition from, and only from, the particular historical contexts to which they are applied. They are not applicable without change across space and time, because their content changes with the reality they seek to comprehend. This means that they cannot be substantively defined transhistorically; as general categories, they are necessarily empirically open-ended. We cannot offer a universally applicable definition, of an empirical sort, of what for instance productive forces or production relations are. Conversely, in so far as Marx’s concepts are substantive categories the concepts of concrete empirical phenomena – they are necessarily historical categories: a feudal force, a capitalist relation, and so on. Their content is historically specific, and their validity historically circumscribed.

Marx himself argues this when discussing ways of defining property, a key concept in his writings, in a passage to which I shall return:

In each historical epoch, property has developed differently and under a set of entirely different social relations. Thus to define bourgeois property is nothing else than to give an exposition of all the social relations of bourgeois production. To try and give a definition of property as of an independent relation, a category apart, an abstract and eternal idea, can be nothing but an illusion of metaphysics or jurisprudence. (1847a: 154)

Both the relational (and therefore extended) and the historical character of what Marx saw as involved in ‘definition’ are well brought out here. E. P. Thompson argues similarly, regarding that central concept Marx so conspicuously ‘failed’ anywhere in his writings to define, class. He could not have done so, if a definition of the analytic philosophy sort is required – exclusive, unambiguous, closed and universal. For, quite simply, ‘class is defined by men as they live their own history, and, in the end, this is its only definition’ (1968: 11). To define a class – or any other social phenomenon – is, in the final analysis, to write its history.

Neither a philosophical defence of an ontology of internal relations, nor a systematic textual justification of its ascription to Marx, will be attempted here. I cannot, in any case, better Ollman on this. I rest my case rather on the degree to which interpreting Marx in this way allows us to make better sense of how he actually used his fundamental concepts – the central concern of this book – than the standard alternatives. But the major point I want particularly to emphasise at this stage in the argument is this. If Marx indeed did adhere to an internal relations perspective, we must approach the general concepts of the 1859 Preface very differently from ‘traditional historical materialism’.

In particular, we can no longer assume that terms like forces and relations of production, or base and superstructure, refer unambiguously or consistently to different, and mutually exclusive, bits of empirical reality as they would in an atomistic ontology. They unproblematically do so for Cohen (as they do equally for many ‘relative autonomy’ models). Thus he avers that ‘productive forces strongly determine the character of the economic structure, while forming no part of it’, and in turn ‘the economic structure is separate from (and explanatory of) the superstructure’ (1978: 31, 218), all these entities being defined in empirical terms. Such separations are of course logically required if the Preface’s claims of ‘determination’ are to be coherently rendered either in straightforwardly causal or in functional terms. Determining and determined phenomena must be logically independent of one another, contingently rather than essentially related.

But on the view argued here, the empirical referents of Marx’s concepts may neither be mutually exclusive, nor consistent across space and time. An empirical particular – a form of division of labour, for example – might figure as a production relation under one description and a productive force under another. The same division of labour might be a productive force in one historical context and a relational fetter on such forces in another. The ‘detail’ division of labour discussed by Marx in chapter 14 of Capital I is exactly like this; the foundation of the productivity of early capitalist manufacture, it fetters the subsequent development of machine industry. In the same way notions of personal liberty would normally, for Marxists, belong without question in the ‘ideological superstructure’ of society; but we shall encounter an instance later on in which Marx treated them none the less as a productive force. For him even theory may under certain circumstances become a ‘material force’ (1843d: 142). This conceptual fluidity reflects both the relational and the historical character of social reality itself. The point is an important one. For if the key general concepts of historical materialism are necessarily empirically open-ended and multi-referential, they cannot then officiate as the building-blocks of an overarching ‘theory of history’ in the traditional way.


If, for Marx, substantive categories are necessarily historically delimited, it should not surprise us to find that the few observations he allows himself to make on ‘production in general’ are relatively trivial. He considered, indeed, that all that could reasonably be asserted at this level of ahistorical generality were ‘flat tautologies’ (1857: 86). ‘Production in general’ is a rational abstraction, since there are generic features common to all productive activities and epochs. But it is an abstraction none the less. In Marx’s own words ‘the so-called general preconditions of all production are nothing more than ... abstract moments with which no real historical stage of production can be grasped’ (ibid.: 99).

Thus he admits, for example, the obvious truth that all production involves some sort of labour process in which people transform raw materials into products, using their labour-power, with the aid of instruments of production of one sort or another. This is ‘an everlasting nature-imposed condition of human existence’ (1867a: 184), which must go on ‘in all social formations and under every possible mode of production’ (1865a: 820). Even here, however, Marx is careful to say that to conceptualise the labour process thus is to view it abstractly and ahistorically, ‘independently of the particular form it assumes under given social conditions’ (1867a: 177), and thus unempirically. And he only deals with it, in Capital, in such abstraction at all in order to bring out, by means of contrast, the differentia specifica of its distinctively capitalist form, in which he is interested.

In the same way, in a well-known letter to Kugelmann of 11 July 1868, Marx dismissively observes that ‘every child knows’ that social labour must be distributed in definite proportions between different branches of production whatever the society. Obviously some proportionality of inputs and outputs must be secured if production is to go on, irrespective of its mode. Such ‘natural laws’ cannot be done away with. But what can change is ‘the form in which these laws operate’ (1868). Such social forms, for Marx, are the proper subject-matter of economic science. The ‘material side, which the most diverse epochs of production may have in common’ is something ‘whose examination ... lies beyond political economy’ (1858a: 881). He frequently criticises political economists for their ‘crude materialism’ (1858a: 687; cf. 1865a: 323) in failing to differentiate these levels of analysis, leading to the conflation of historical and transhistorical. As we shall see below, this is fundamental to his critique of fetishism.

Analysis of what pertains to production in general, then, is for Marx economically antediluvian, an enterprise yielding ‘commonplaces’ which at best ‘had a historical value in the first beginnings of the science, when the social forms of bourgeois production had still laboriously to be peeled out of the material, and, at great effort, to be established as independent objects of study’ (1858a: 881). One consistent theme in Marx’s comments on production in general nevertheless does deserve emphasis here. Though similarly tautological and self-evident to Marx himself, it above all else distinguishes his approach from that of most mainstream economics. This is his insistence on the irreducibly social nature of production – the point brought out by Colletti above.

In The German Ideology Marx stresses that ‘the production of life ... appears as a double relationship: on the one hand as a natural, on the other as a social relationship’ (1846a: 41, emphasis added). Wage Labour and Capital elaborates on this:

In production, men enter into relation not only with nature. They produce only by cooperating in a certain way and mutually exchanging their activities. In order to produce, they enter into definite connections and relations with one another, and only within these social connections and relations does their relation with nature, does production, take place. (1847b: 211)

The same point is basic to Marx’s polemic in the General Introduction to the Grundrisse against the ‘unimaginative conceits of the eighteenth-century Robinsonades’, the isolated hunters and fishermen who provide the analytic starting-point for the classical economists’ systems. ‘Production by an isolated individual outside society ... is as much of an absurdity as is the development of language without individuals living together and talking to each other’ (1857: 84). Other texts, early and late, concur (including in the linguistic analogy). This sociological emphasis is encapsulated in the nearest Marx comes to a general definition of production as such, again from the General Introduction: ‘All production is appropriation of nature on the part of an individual within and through a definite form of society’ (ibid.: 87, emphasis added). Conversely, as argued in the Grundrisse, ‘when we consider ... society in the long view and as a whole, then the final result of the process of social production always appears as the society itself, i.e. the human being itself in its social relations’ (1858a: 712). To produce material goods in a particular way is ipso facto to participate in the (equally particular) social relations involved in doing so.

Marx’s claim here, it should be stressed, is not that social relations are caused by material production but that it irreducibly involves them. They are part and parcel of it. It accordingly cannot be conceptualised, in any empirically adequate manner, independently of them. In particular, production cannot be conceived as a purely ‘material’ sphere, if material is taken to exclude social. I shall argue in chapter 3 that this vitiates Cohen’s attempted distinction between ‘material’ and ‘social’ relations of production as substantially distinct kinds of relation. In Ollman’s terminology, the connection between people’s productive relations with nature, or labour process, and their productive relations between themselves, or social relations of production, is internal and necessary, not external and contingent. In any given empirical context (a particular, historical form of) the one entails (a particular, historical form of) the other. These are but different ‘sides’ of one and the same set of relations, ‘not substantially separable relations’ at all.


I shall now try to show, with reference to productive forces, that this holds equally for any individual components of production. They cannot be conceived as exclusively material either. As production is in general for Marx a double relationship, so productive forces consistently enough – are also in his usage ‘double’, simultaneously natural (material) and social. By this I mean two things. First, both material and social phenomena – in so far as we may coherently distinguish them thus at all when dealing with human productive activities – may become productive forces. Marx talks routinely of every ‘... natural or social power of labour’ (1858a: 358). Whether or not they do, as in general with Marx and argued above, depends on historical context. Second, and rather more subtly, the material/social distinction is highly problematic here, and real forces of production for Marx habitually partake of both sets of attributes.

Material things land (tools, raw materials, fuel, machines and so forth) only become productive forces in so far as they take on social characteristics. A production line, for instance, is indeed (among other things) a set of material objects, and the natural properties of those objects obviously make up an important part of the explanation of its contribution to the productivity of human labour. But a production line is equally a manifestation or embodiment of both ideas and social relations, and it is only through these that it becomes a production line as distinct from a heap of useless metal. The ideas and social relations, therefore, are quite as intrinsic to the possibility of a production line being a productive force as its more evidently material features. A machine would not be a productive force in a society which lacked the technical knowledge or social organisation to utilise its material capacities; at best it might be a potential productive force.

Conversely, social phenomena – forms of co-operation and division of labour, scientific knowledge, the Protestant ‘work ethic’, and much else – are not productive forces either, except in so far as they are materialised in actual production processes. In this sense ‘all productive force resolves itself’ into ‘a given relation to nature’ (1858a: 540). They, too, will only be potential productive forces outwith this context, and accordingly ‘fettered’ if the requisite conditions for their productive utilisation are absent. One frequent argument for socialism, for instance, is that it would unlock the knowledge and enthusiasm of direct producers in a way capitalism, by virtue of its social relations, does not. In short, neither material objects, nor social phenomena, are inherently or universally productive forces by virtue of their ‘innate’ properties alone; they become so only by dint of the relations – both between people themselves, and between people and nature – in which they stand. Things, then – contrary to Cohen, and much ‘mainstream Marxism’ – are not, in or of themselves, productive forces. The concept is inherently a relational, and therefore an historical and contextual, one.

Indeed, an alternative translation of Marx’s Produktivkräfte – that of productive powers – is rather more revealing of the sense he gave the term. ‘Productive powers’ was in fact the original concept in classical political economy which Marx rendered in German as Produktivkräfte. Whereas a ‘force’ can be conceived as a thing, an independent entity, standing alone, a power is always an attribute of something. For Marx, the power in question is specifically that of social labour. ‘The’ productive forces – a reifying formulation – are precisely the powers to transform nature (and with it human nature) of social labour. Productive forces are thus an attribute of human beings in association, their collective capacities, not a set of things as such at all. Cohen himself notes that ‘neither an instrument of production nor a quantity of raw material in strict speech is a productive force’ (1978: 37); I believe there is more than a slippage of language involved here. As a well-known passage in the Grundrisse hyperbolically puts it:

Nature builds no machines, no locomotives, railways, electric telegraphs, self-acting mules etc. These are products of human industry; natural material transformed into organs of the human will over nature, or of human participation in nature. They are organs of the human brain, created by the human hand. (1858a: 706)

The productive power of social labour may indeed, in the course of human development, increasingly become embodied in things – like machines – and undeniably it is through such embodiment that it is most enhanced. This is what is so revolutionary about modern industry; for Marx it represented a qualitative break, a veritable quantum leap in the unfolding of human productive potential comparable only perhaps with the neolithic revolution. Human beings are, distinctively, creatures who purposefully objectify their collective capacities in the material world they create through transforming nature, and this is fundamental to Marx’s sociology. In his earlier writings it is this which defines human ‘species-being’, while in Capital it remains the purposive character of human labour which ‘stamps it as exclusively human’ and ‘distinguishes the worst architect from the best of bees’ (1867a: 178). Under certain circumstances the productive powers of human labour may then come to appear simply as the intrinsic property of the material things in which these powers are objectified, independently of the social relations through which they alone acquire this property. But such an appearance is, for Marx, exactly that: a fetishistic illusion, and one he was much concerned to confute, as we shall see.

I shall first try to substantiate my interpretation of productive forces from Marx’s texts. Having done that, I shall show that Marx himself explicitly criticised what has become the standard conception – the reduction of productive forces to the things used in production – as exactly such a fetish, an ideological confusion rooted in the alienated forms of appearance of human capacities and social relations under capitalism.


The passage already quoted from The German Ideology, which describes production as a ‘double relationship’, is an appropriate place to begin. For it reads, at more length, like this:

the production of life, both of one’s own in labour and of fresh life in procreation ... appears as a double relationship: on the one hand as a natural, on the other as a social relationship. By social we understand the cooperation of several individuals, no matter under what conditions, in what manner and to what end. It follows from this that a certain mode of production, or industrial stage, is always combined with a certain mode of co-operation, or social stage, and this mode of co-operation is itself a ‘productive force’. (1846a: 41, emphasis added)

Note should be taken of the words I have italicised: it is interesting to see what Marx himself considered to follow for the concept of productive forces from his understanding of production as a ‘double relationship’. Cohen does not discuss the closing sentence to this passage: oddly, since Marx here explicitly includes social relations among productive forces, something Cohen is at pains to deny. Were this the only occasion where Marx did so, it could perhaps be dismissed as an aberration, notwithstanding its evident connection with the logic of his overall argument. But it is not.

Marx speaks, for instance, in the same text – when discussing communism – of ‘communal economy’ as ‘in itself ... a new productive force’. Though he makes clear that communism supposes an appropriately developed technology, it remains the ‘communal economy’, and not the technology in and of itself, that is the productive force at issue (1846a: 40n). He also counterposes ‘industrial productive forces’ and ‘productive forces ... based on association and the community’ (ibid.: 9 1). Lest such sentiments be thought a peculiarity of Marx’s early works alone, the Grundrisse, also, describes ‘the community itself ... as the first great force of production’ (1858a: 495).

In the same work, Marx argues that ‘the capability to consume’ is ‘the development of an individual potential, a force of production. The saving of labour-time [is] equal to an increase in free time, i.e., time for the full development of the individual, which in turn reacts back upon the productive power of labour as itself the greatest productive power’ (1858a: 711). Discussing Robinson Crusoe, he conceives such castaways – bereft, presumably, of advanced technology – as persons ‘in whom the social forces are already dynamically present’ none the less (1857: 74, 84-5). Money, as well as being ‘an intrinsic relation of production’, is ‘itself an instrument of production’ and ‘a driving wheel for the development of all forces of production, material and mental’ (1858a: 215, 216, 223; cf. 225). ‘Trade’ is numbered as a productive force alongside industry and science (ibid.: 277). So is ‘increase in population’, because it ‘makes possible a greater combination and division of labour’ (ibid.: 399). Most importantly, perhaps, it is ‘the human being himself who is ‘the main force of production’ (ibid.: 422). 12 Capital, similarly – though not explicitly using the term ‘productive force’ – describes state activity during the early development of capitalism as ‘itself an economic power’ (1867a: 751). Not all of these can plausibly be dismissed as merely metaphorical extensions of the concept.

Returning to the early works, a draft article of 1845 brings out the thoroughly contextual, historical character of Marx’s concept with macabre irony:

Under the present system, if a crooked spine, twisted limbs, a one-sided development and strengthening of certain muscles, etc., makes you more capable of working (more productive), then your crooked spine, your twisted limbs, your one-sided muscular movement are a productive force. If your intellectual vacuity is more productive than your abundant intellectual activity, then your intellectual vacuity is a productive force, etc. etc. If the monotony of an occupation makes you better suited for that occupation, then monotony is a productive force. (1845a: 285)

F. W. Taylor, the pioneer of ‘scientific management’, would readily have concurred. For optimal operation of his system he sought workers ‘of smaller calibre and attainments’, ‘of the type of the ox’ (quoted in Braverman, 1974: 118, 108). Adam Smith thought the modern division of labour made the worker ‘as stupid and ignorant as it is possible for a human creature to become’; a destiny he saw – in this case I presume with unintended irony – as inevitable, for ‘the labouring poor, that is, the great mass of the people’, ‘in every improved and civilising society’ (Wealth of Nations, quoted in Marx, 1867a:362).

But rather than multiply isolated quotations (which would not however be difficult in this instance) I think the argument best advanced by looking in detail at how Marx comprehends productive forces in the context of one substantial piece of historical analysis. The example I shall give is his study of the emergence of specifically capitalist productive forces. It is drawn, therefore, from the very centre of his intellectual and political concerns – the analysis of the capitalist mode of production. This example will also allow us some insight into how, in one empirical instance – but again, an absolutely central one in his work – Marx saw technological development and social change intersecting in the genesis of a new mode of production, thus clarifying the claims of the 1859 Preface on this point. The texts on which I mainly draw are the first volume of Capital, and its draft so-called ‘sixth chapter’ (1866). This latter is a long manuscript entitled ‘Results of the immediate process of production’ which was written some time between 1863 and 1866, and published only in 1933 (and first translated into English in 1976). ‘Results ...’ clarifies some of Capital’s arguments in important ways. Marx originally intended it to follow the chapter on Wakefield with which Capital I as we know it closes.

In these writings, Marx distinguishes what he calls manufacture and machine industry as successive historical stages (though stages ‘no more separated from each other by hard and fast lines of demarcation, than are geological epochs’ (1867a: 371)) in the development of a specifically capitalist production process. These stages rest upon different historical forms of the labour/capital relation, which Marx refers to as formal and real subordination, subjection or subsumption of labour to capital respectively. To these two forms of the capital/labour relation, and stages in the development of a capitalist production process, correspond two different modes of exploitation of labour. These are absolute and relative surplus-value. Marx details these in parts 3 and 4 of Capital I. Absolute surplus-value is obtained by extending the working-day, relative surplus-value by diminishing, through increases in productivity, the portion of it devoted to reproducing the labourer’s wage-costs.

Formal subordination of labour to capital – the social basis of capitalist manufacture – involves, for Marx, simply a change in social relations. Workers previously in possession of their means of production, whether as peasants or artisan craftsmen, become wage-labourers, dependent for their livelihood upon the sale of their labour-power to capitalist employers. Real subordination, by contrast, comes into being only when ‘a technologically and otherwise specific mode of production – capitalist production – ... transforms the nature of the labour process and its actual conditions’ (1866: 1034-5). Labour is said to be really subjected to capital when the production process itself, and above all the instruments of labour, have become transformed in such a way that they can only be operated cooperatively, by ‘labour-power socially combined’ (ibid.: 1040), or what Capital terms ‘the collective labourer’. At this stage of capitalist development – unlike in manufacture, where the power of capital remains a merely social requisite of production – ‘the sway of capital develops into a requisite for carrying on the labour process itself, into a real requisite of production’ (1867a: 330). The difference Marx is pointing to is that between, say, hand-loom weavers working under the putting-out system, where the actual mode of work is no different from that they would have pursued had they remained independent craftsmen, and power-loom weavers working in large factories.

Real subordination occurs with the development of machine industry. Thereafter, it is not merely their dispossession, but the concrete forms of the actual labour-process itself – large-scale, coordinated, mechanised productive processes – that prevent the labourers from working individually for themselves. The very technology of production is socialised. In its capitalist form, however, modern industry is a palpable, material embodiment of the social power of capital and the expropriation of labour. This, for Marx, in contrast to merely formal subordination, represents ‘the development of a specifically capitalist mode of production’ (1866: 1021), ‘a specifically capitalist form of production ... at the technological level too’ (ibid.: 1024), ‘capitalist production proper’ (ibid.: 1027).

Crucially for my argument, Marx is adamant that the merely formal subordination of labour does not entail any immediate changes in the labour process, though it may (it also may not) lead to their adoption. Nor, equally importantly, does it in any way presuppose the prior development of a specifically capitalist technology, or indeed any technical advancements whatsoever on previous modes of production. Formal subordination of labour to capital, according to ‘Results ...’:

does not in itself imply a fundamental modification in the real nature of the labour process, the actual process of production. On the contrary, the fact is that capital subsumes the labour process as it finds it, that is to say, it takes over an existing labour process, developed by different and more archaic modes of production ... For example, handicraft, a mode of agriculture corresponding to a small, independent peasant economy. If changes occur in these traditional established labour processes after their takeover by capital, these are nothing but the gradual consequences of that subsumption. The work may become more intensive, its duration may be extended, it may become more continuous or orderly under the eye of the interested capitalist, but in themselves these changes do not affect the character of the actual labour process, the actual mode of working. (1866: 1021, final emphasis added)

There is no change as yet in the mode of production itself. Technologically speaking, the labour process goes on as before, with the proviso that it is now subordinated to capital. (ibid.: 1026)

In Capital, Marx extends these points. He distinguishes, within the manufacturing epoch, two successive regimes which precede the development of machine industry and capitalist production proper. These are ‘simple co-operation’ and ‘detail division of labour’. In the first, workers are brought together, either physically under one roof or organisationally under the aegis of a single capitalist; in the second – as in Adam Smith’s famous discussion of pin manufacture – existing crafts are broken down into detailed operations devolving on different individuals. In neither case, Marx makes clear, is the technological basis of the production process altered, though its social relations are already undeniably capitalist (1867a: ch. 24). Indeed eventually this ‘narrow technical basis’ becomes a ‘fetter ... on the dominion of capital’ (ibid.: 368).

Whereas, for Marx, the real subordination of labour to capital always supposes its formal subsumption – the wage relation being fundamental to all forms of capitalist enterprises – ‘the converse does not necessarily obtain’ (1866: 1019). We may, in other words, find formal subordination in the absence of, or side by side with, the specifically capitalist mode of production based on real subordination. The latter arises out of the former only ‘if the historical circumstances are otherwise favourable, as they were for instance in the sixteenth century’ (ibid.: 1022). There is no automaticity about the progression. But on one point the ‘sixth chapter’ is insistent:

absolute surplus-value always precedes relative. To these two forms of surplus-value there correspond two separate forms of the subsumption of labour under capital, or two distinct forms of capitalist production. And here too the one form always precedes the other. (ibid.: 1025)

Capital is equally emphatic. ‘A greater number of labourers working together ... under the mastership of one capitalist, constitutes, both historically and logically, the starting-point of capitalist production’ (1867a: 322).

Now, it is true that in this account capitalist production proper is only achieved with the real subordination of labour to capital. Only then does capitalism become ‘a mode of production sui generis’ (1866: 1035). It is true, also, that Marx at one point asserts that it is only at the latter stage that ‘the corresponding relations of production between the various agents of production and above all between the capitalist and the wage-labourer, come into being for the first time’ (ibid.: 1024). This is, on the face of it, somewhat mystifying – indeed contradictory – in the context of the overall analysis, for Marx is clear throughout that formal subordination is a capitalist relation, or ‘form of capitalist production’ (ibid.: 1025). The only sense I can make of this apparent inconsistency is to interpret Marx as saying that only with real subordination do capitalist relations achieve what he later calls their ‘adequate form’ (ibid.: 1035) – only then, in the words of the Grundrisse, do they gain ‘totality and extent’ (1858a: 277; cf. 297), becoming obligatory and general, because materialised in the very forms of the labour process itself. This indeed does presuppose an industrial technology.

But what is beyond any doubt, throughout the analysis, is that technological change as such is not the primum agens in the rise of capitalism, or even in the development of the production process most appropriate to it, machine industry. It may be capitalism’s most productively revolutionary consequence, but that is a different issue. It is in fact competition between capitals – explained by the division of labour between them, a social relation – which compels technical innovation, and the labour/capital relation that makes its general adoption possible on the basis of co-operative labour. Social relations, not productive forces as conventionally construed, thus have explanatory primacy. ‘Capital is productive, i.e. an essential relation for the development of the social productive forces’ – these being ‘the productive powers of labour, which [it] incessantly whips onward with its unlimited mania for wealth’ (1858a: 325, emphasis added). The ‘sixth chapter’ is clear that formal subsumption of labour – a change in social relations alone – is not merely historically antecedent to, but forms ‘the premise and precondition of its real subsumption’ (1866: 1026). Capital, again, reiterates the point. It is the production of relative surplus-value – on the basis of real subordination – which revolutionises out and out the technical processes of labour’; but this presupposes a specific mode, the capitalist mode of production, a mode which, along with its methods, means and conditions, arises and develops itself spontaneously on the foundation offered by the formal subjection of labour to capital’ (1867a: 5 10).

How, then, are we to reconcile this historical analysis with the picture usually drawn from the 1859 Preface, in which the development of productive forces – understood technologically, as the things used in production (including labour-power) – is the ultimate motor force of historical change, impelling transition from one mode to another? At first sight, Marx’s analysis – an analysis, it should again be recalled, of the rise of that mode of production whose ‘laws of motion’ preoccupied him above all others – flatly contradicts this picture. Here, changes in production relations bring about developments in technology, not the other way about. A perusal of part 8 of Capital I, incidentally, would complicate the issue still further, since it would additionally reveal the key part played in Marx’s account by those historically ‘favourable circumstances’ of the sixteenth century to which he refers in the ‘sixth chapter’. It is difficult to see how for instance Henry VIII’s dissolution of the monasteries is readily susceptible to technological explanation, no matter how ‘ultimate’.

Marx’s account is in fact (logically) compatible with Cohen’s functionalist reading of the ‘primacy thesis’, to which I shall return. It evidently cannot, however, be reconciled with standard causal versions of that claim. In the face of this Anderson (1974a: 204ff.), Bettelheim (1976: 86ff.), Balibar (in Althusser, 1970: 233ff.) and others have sought to reverse the traditional paradigm, arguing instead for the dominance of relations over forces. Whilst I sympathise with what they are attempting, I believe the problem to be illusory. I also think this solution misleading, for several reasons. First, it bends the stick too far; as I shall show, there is a sense in which Marx’s analysis here is compatible with the ‘primacy thesis’, though that sense is not Cohen’s. But simply to reverse the line of causality between forces and relations obscures the important extent to which, for Marx, the growth of human productive power does remain the fundamental dynamic of historical progress. Second although the authors I have cited are not in fact themselves generally guilty of this – a mere reversal of terms may have the consequence of preserving a conception of productive forces as things which was not Marx’s own. and which is. as I shall argue below, specifically fetishistic. Finally, this way out of the dilemma entails dismissing the claims of the 1859 Preface as anomalous, which is to say the least implausible.

In fact, Marx’s analysis of the rise of capitalist productive forces contradicts the claims of the 1859 Preface only so long as productive forces are understood in the traditional, restrictive manner. So soon as the forms of co-operation and division of labour entailed in the formal subsumption of labour to capital (or, indeed, the competitive relations between individual capitals) are themselves acknowledged to be productive forces in this historical context, the apparent inconsistency disappears. The rise of capitalism is indeed understood in terms of new productive forces conflicting with old production relations, exactly as the Preface requires. Only in this case – in the first instance, at any rate – the new forces at issue are embodied in new forms of social relationship rather than new technologies. Such an interpretation is of course consistent with the passages cited above from The German Ideology and elsewhere, which indicate that for Marx social relations could indeed be productive forces. But we do not need to go so far for authorisation of this heresy. The texts on which I have been drawing – Capital and ‘Results of the immediate process of production’ – themselves say as much.

Although, for Marx, formal subsumption of labour to capital entails no technological change in the labour process, it does lead to a considerable enhancement in the productive power of human labour. He repeatedly, even tediously, stresses this. For the Grundrisse, ‘the greater the extent to which production still rests on mere manual labour ... the more does the increase of the productive force consist in [producers’] collaboration on a mass scale’ (1858a: 529). In the ‘sixth chapter’, Marx lists several ways in which this happens.

Formal subsumption ‘increases the continuity and intensity of labour; it is more favourable to the development of versatility among the workers, and hence to increasing diversity in modes of working and ways of earning a living’ (1866: 1026-7). In contrast to slave or servile labour, ‘this labour becomes more productive ... The consciousness (or better, the idea) of free self-determination, of liberty, makes a much better worker of the one than of the other, as does the related feeling (sense) of responsibility’ (ibid.: 1031, Marx’s parentheses). ‘Social forms of consciousness’ – ‘superstructural’ phenomena par excellence for most Marxists – would seem here to be part of capitalism’s productive forces, and unproblematically so for Marx. A similar case could be made for Max Weber’s ‘Protestant ethic’, an ‘orientation to conduct’ of whose benefits to capitalist enterprise neither Marx himself, nor Marxist historians like Hill or Thompson, have been entirely unaware. This position is perfectly coherent within the multi-referential and non-exclusive interpretation of Marx’s general concepts argued earlier, but not otherwise. The same goes for the passage on state activity as ‘itself an economic power’ quoted above. Elsewhere Marx similarly describes French state formation as ‘a powerful coefficient of social production’ (1871: 75). My own previous work (Corrigan and Sayer, 1985) suggests he was not exaggerating.

Perhaps most importantly of all, for Marx, formal subsumption permits – ‘even on the basis of the old, traditional mode of labour’ large increases in the scale of production, and ‘this enlargement of scale constitutes the real foundation on which the specifically capitalist mode of production can arise’ (1866: 1022). Once again Capital generalises the point:

the sum total of the mechanical forces exerted by isolated workmen differs from the social force that is developed, when many hands take part simultaneously in one and the same undivided operation ... In such cases the effect of the combined labour could not be produced at all by isolated individual labour, or it could only be produced by a great expenditure of time, or on a very dwarfed scale. Not only have we here an increase in the productive power of the individual, by means of co-operation, but the creation of a new power, namely, the collective power of masses. (1867a: 326)

Marx goes on to remark that ‘mere social contact begets in most industries an emulation and a stimulation of the animal spirits that heighten the efficiency of each individual workman’. His general conclusion – something of a syntactic battering-ram, but worth giving in full – is this:

Whether the combined working-day, in a given case, acquires this increased productive power, because it heightens the mechanical force of labour, or extends its sphere of action over a greater space, or contracts the field of production relatively to the scale of production, or at the critical moment sets large masses of labour to work, or excites emulation between individuals and raises their animal spirits, or impresses on the similar operations carried on by a number of men the stamp of continuity and many-sidedness, or performs simultaneously different operations, or economises the means of production by use in common, or lends to individual labour the character of average social labour – whichever of these be the cause of the increase, the special productive power of the combined working day is, under all circumstances, the social productive power of labour, or the productive power of social labour. This power is due to co-operation itself. When the labourer cooperates systematically with others, he strips off the fetters of his individuality, and develops the capabilities of his species. (ibid.: 329, emphasis added)

Only the most determined casuist, I would suggest, could deny that for Capital, as much as for The German Ideology, people’s ‘mode of co-operation is itself a “productive force”’ – and a pretty powerful one at that. The forces/relations distinction as conventionally drawn must accordingly collapse. These are not mutually exclusive concepts, denoting substantially distinct entities.

Cohen himself does discuss these analyses, though not at any great length. He deals with the apparent contradiction between Marx’s assertion here of the historical and logical precedence of formal over real subordination, and the ‘primacy thesis’, via his interpretation of the latter as a functional rather than straightforwardly causal argument. In brief.

the primacy thesis does not say that forces characteristic of capitalism preceded its arrival. It rather requires that nascent forces could not be used or developed within pre-capitalist relations, and that a capitalist structure was necessary for productive progress. (1978: 179)

‘Forces [even as yet non-existent forces] select structures according to their capacity to promote development’ (ibid.: 162; my interjection, based on Cohen’s arguments in ibid.: 177). Cohen is right that if Marx is interpreted thus, the contradiction disappears. His reading is internally coherent, whatever other objections might be raised against it. In chapter 5 I shall argue against both functional explanation as Cohen conceives it, and its ascription by him to Marx.

The second problem I have raised is that of social relations themselves being putative productive forces. I hope I have shown that the view that they may be is more than warranted by Marx’s texts. Cohen, to my mind, simply evades this issue when commenting on these analyses. He draws a distinction, within the category of production relations, between ‘social relations’ and what he calls ‘material relations between producers’ or ‘work relations’. The former alone enter into the ‘economic structure’. I shall discuss this distinction further in my next chapter. The relevant point here is that Cohen classifies the forms of co-operation and division of labour I argue were for Marx (also) productive forces, as ‘material relations’. He then concedes that ’something in this conceptual area is a productive force, but not the work relations themselves’ (ibid.: 113). In his view, the relevant productive force is ‘knowledge of ways of organising labour’, not that organisation as such (ibid.). This seems to me a wholly artificial, if characteristically ingenious, piece of special pleading: knowledge, like machines, is a productive force only in so far as it is productively applied.

It is also, it should be said, an extremely cavalier reading of the texts I have cited. Marx makes it abundantly clear that the relevant power is that of – in his own words – ‘co-operation itself’, not the knowledge of or blueprint for co-operation, and he does so repeatedly. The only justification for denying Marx’s contention is consistency within Cohen’s own theory, and what ever the merits of the latter, this is not a good reason for attributing to Marx a view he demonstrably did not hold, or denying to him the view he equally demonstrably did. Particularly if, as is the general argument of this book, an alternative and equally coherent account of historical materialism, which is consistent with Marx’s own conceptual usage, is eminently possible.


Following a characteristically capacious account of what, in capitalism, ‘the social productive forces of labour, or the productive forces of directly social, socialised (i.e. collective) labour’ actually comprise, the ‘sixth chapter’ goes on to observe that:

This entire development of the productive forces of socialised labour ... takes the form of the productive power of capital. It does not appear as the productive power of labour, or even of that part of it which is identical with capital. And least of all does it appear as the productive power either of the individual worker or of the workers joined together in the process of production. The mystification implicit in the relations of capital as a whole is greatly intensified here, far beyond the point it had reached or could have reached in the merely formal subsumption of labour under capital. (1866: 1024)

This brings me to the second major argument I want to develop on Marx’s understanding of productive forces. This is that to conceive them otherwise than as powers of social labour – and specifically, to identify them with the ‘things used in production’, means of production and labour-power – is, so far as Marx is concerned, to ‘mystify’ them. It is in fact a paradigm instance of what he diagnosed as ‘the fetishism peculiar to the capitalist mode of production’, from which, in his view, most economics suffered (ibid.: 1046). Since the concept of fetishism will be a central one in the overall argument of this book, it is worth spending some time establishing what exactly Marx meant by it.

We saw earlier that he systematically distinguished a ‘material side’ and a ‘formal side’ to production. The material side is that which all human productive activities have in common, or that which pertains to production in general in abstraction from its particular social modalities. For instance all production involves a labour process, and all products have a use-value (understood as a capacity to meet some human need). But to conceive production simply thus, from its ‘material side’ alone, is, Marx argues, to grasp it abstractly. For Robinson Crusoe notwithstanding – production only takes place empirically within definite social relations, and it is these which imprint on its elements, in any given case, their particular social or economic forms. It is these forms which are Marx’s prime concern, defining ‘real historical stages’ of production.

In capitalism. for instance, the labour process is also a process of production of surplus-value, and the product, in addition to having use-value, assumes the specific ‘value-form’ of the commodity. The latter term in each pair is a distinctively social characteristic, and the specific object of Marx’s analysis. Thus, clarifying the methodological starting-point of Capital in his 1880 Notes on Adolf Wagner, he makes it clear that, ‘I do not start from “concepts” ... What I start from is the simplest social form in which the labour product is represented in contemporary society, and this is “the commodity”’ (1880a: 50; cf. 1858a: 881).

Fetishism confuses, or indeed inverts. these two sets of attributes, material and formal or natural and social. Properties which things acquire entirely as a consequence of their standing in a specific set of social relations are mistakenly seen as inhering in, and explained by, the material qualities of those objects themselves. In Marx’s own words, fetishism ‘consists in regarding economic categories, such as being a commodity or productive labour, as qualities inherent in the material incarnations of these formal determinations or categories’ (1866: 1046). To fetishise economic phenomena is to ‘metamorphose the social, economic character impressed on things in the process of production into a natural character stemming from the material nature of those things’ (1878: 229).

So commodity fetishism, to take Marx’s best-known example, consists in seeing the value of commodities as something intrinsic to them as things, and therefore explicable by their material features, their scarcity, durability, utility or whatever. But these material features, for Marx, pertain only to the use-values of commodities, something the goods in question would continue to have even if they did not assume the commodity-form, and – as for instance when immediately consumed by their producers – accordingly had no value as such. The commodity ‘as an exchange-value differs from itself as a natural, material thing’ (1858a: 188). Exchange-value is ‘a cipher for a relation of production’ (ibid.: 141). It is a material mode of expression of a social relation, the respective amounts of labour necessary under given conditions to the production of different commodities, and it is the social relations specific to commodity production which wholly explain why the labour-product should assume this specific (and somewhat mystifying) value-form: ‘No scientist has yet discovered what natural properties make definite proportions of snuff tobacco and paintings “equivalents” for one another’(1863c: 130). Rather:

Where labour is communal, the relations of men in their social production do not manifest themselves as ‘values’ of ‘things’. Exchange of products as commodities is a method of exchanging labour, [it demonstrates] the dependence of the labour of each upon the labour of the others [and corresponds to] a certain mode of social labour or social production. (ibid.: 129, interpolations from editorial collation of Marx’s ms.)

In money, the universal symbol of value, the fetish reaches its apotheosis: ‘a social relation, a definite relation between individuals, appears as a metal, a stone, as a purely physical, external thing’ (ibid.: 239). This kind of transubstantiation is, for Marx, wholly analogous with what goes on in ‘the mist-enveloped regions of the religious world’, and explains his use of the term fetishism (which he took originally from Feuerbach’s critique of religion): ‘In that world the productions of the human brain appear as independent beings endowed with life, and entering into relation both with one another and with the human race. So it is in the world of commodities with the products of men’s hands’ (1867a: 72).

This first fetishistic confusion, of material and social, entails a second. To comprehend social properties of objects as deriving from their material attributes is at the same time implicitly to universalise them, to deny their specific historical character. Thus value, in this example, becomes a property products possess transhistorically, irrespective of their particular modes of production. It is no longer, as for Marx, ‘a historic relation’ (1858a: 252), a property things acquire only within a definite historical mode of production – commodity production, which rests on particular relations between people, a specific form of social division of labour, not found in all forms of human society. Capitalist conditions are thereby covertly taken as premises of any and all human sociation. The connection between this critique of fetishism, and the argument developed earlier in this chapter concerning the necessarily historical character of Marx’s substantive categories will, I hope, be evident. For him, ‘it is precisely these forms that are alone of importance when the question is the specific character of a mode of social production’ (1863a: 296, emphasis added).

Importantly, for Marx fetishism is not simply an intellectual faux pas, a category error. Consistently with his general denial of the independence of ideas from people’s ‘materialistic connection’, of consciousness from experience – which I shall discuss further in chapter 4 – he seeks to root fetishism in the forms in which, under capitalism, individuals actually experience their social relations (Sayer, 1983, ch. 1; Godelier, 1964; Mepham, 1972). To use the hackneyed but none the less useful analogy, the mechanism of fetishism is akin to that of a mirage rather than an hallucination. It is not that people mistake what they see, what they see misleads them, for thoroughly objective reasons to do with how it presents itself to their consciousness. Marx makes this point against the Ricardian socialist, Thomas Hodgskin:

Hodgskin says that the effects of a certain social form of labour are ascribed to objects, to the products of labour; the relationship itself is imagined to exist in material form ... Hodgskin regards this as a pure subjective illusion which conceals the deceit and the interests of the exploiting classes. He does not see that the way of looking at things arises out of the relationship itself; the latter is not an expression of the former. but vice versa. (1863c: 295-6)

A corollary of this is that fetishism is not equally characteristic of all social formations (any more than, for Marx, ideology was an invariant feature of social life), and its explanation must accordingly be empirically specific. He argues, for instance, that exploitation was much more transparent in feudal societies, because it was not concealed by the material forms of the labour process itself as in capitalism. Work for oneself and work for one’s lord were palpably separate in a way they are not within the wage-relation, where all of one’s labour-time appears to be paid for. Empirically this is perhaps debatable, since feudal relations involved norms of reciprocity; crudely, protection, both earthly and divine, in exchange for services. But Marx’s intent is clear.

For Marx, fetishism reaches its zenith in capitalism, and this is explained by the singularity of the social relations on which that mode of production rests. Fundamental to capitalism is a spontaneous social division of labour between private producers, which is not subject to any conscious overall social regulation. This ‘same division of labour that turns [people] into private producers’, he argues, ‘also frees the social process of production ... from all dependence on the wills of those producers.’ Hence, ‘their relations to each other ... assume a material character independent of their control and conscious individual action.’ What ensues is ‘a whole network of social relations spontaneous in their growth and entirely beyond the control of the actors’ (1867a: 108, 92-3, 112). This is epitomised in Adam Smith’s ‘hidden hand’ of the market, governing the movements and prices of commodities apparently by laws entirely of its own. Indeed under capitalist conditions this appearance of autonomy on the part of ‘market forces’ is real, for it is only in the form of relative prices of goods that the social relations between independent producers are actually expressed, and only through the price-mechanism that equilibriation of inputs and outputs of labour between different branches of production is regulated.

The Grundrisse summarises this key dimension of the sociology of capitalism:

As much, then, as the whole of this movement appears as a social process, and as much as the individual moments of this movement arise from the conscious will and particular purposes of individuals, so much does the totality of the process appear as an objective interrelation, which arises spontaneously from nature; arising, it is true, from the mutual influence of conscious individuals on one another, but neither located in their consciousness, nor subsumed under them as a whole. Their own collisions with one another produce an alien social power standing above them, produce their mutual interaction as a process and power independent of them. Circulation, because a totality of the social process, is also the first form in which the social relation appears as something independent of the individuals, but not only as, say, in a coin or in exchange value, but extending to the whole of the social movement itself. The social relation of individuals to one another as a power over the individuals which has become autonomous, whether conceived as a natural force, as chance or in whatever other form, is a necessary result of the fact that the point of departure is not the free social individual. (1858a: 196-7)

The Hegelian language of this passage prompts one final observation. In this respect, at least, Marx’s analysis of the ‘bewitching’ phenomenology of capitalism was remarkably consistent over time. In the early writings he was more inclined to use the concept of alienation than that of fetishism, though the former figures in his later works to a far greater extent than some commentators have acknowledged (and the latter, which comes from Feuerbach, is by no means entirely absent from the ‘young Marx’). But the heart of the analysis that, within capitalism, people’s social relations take on the ‘alien’ form of objects, and are accordingly fetishised in their consciousness – remains the same. So does the explanation of this fetishism as being consequent upon a particular social form of division of labour. This was, in short, an enduring and central theme of Marx’s historical sociology.


Many commentators have recognised the importance of the concept of fetishism in Marx’s writings. Indeed Cohen himself offers an account of fetishism which is in many ways outstanding. The relevance of Marx’s critique of fetishism to what we understand by productive forces, however, has received much less attention. Fortunately, Marx himself provides explicit criticism of fetishised views of productive forces – their identification with the material objects in which they are embodied – and he does so throughout his work. In The German Ideology, for instance, he argues that under capitalist conditions:

The social power, i.e., the multiplied productive force, which arises through the co-operation of different individuals as it is determined by the division of labour, appears to these individuals ... not as their own united power, but as an alien force existing outside them, of the origin and goal of which they are ignorant, which they cannot thus control, which on the contrary passes through a peculiar series of phases and stages independent of their will, nay even being the prime governor of these. (1846a: 46)

It is not without irony that although Cohen quotes this passage, it could have been written as an exact critique of his own formulation of Marx’s so-called ‘development thesis’. Both the independence of ‘material’ productive forces from social relations, their character as ‘things’, and their supposedly inherent tendency to develop, such development being the ‘prime governor’ of history, emerge as precisely forms of appearance, arising out of specifically capitalist production relations.

We might note in passing the (important) converse that for Marx neither such development, nor its apparent spontaneity, were in fact universal. Closely following Hegel, he characterised the ‘Asiatic’ mode of production, for instance, in terms of its ‘millennial stagnation’. ‘Indian society has no history at all, at least no known history’ (1853b: 217); life there is ‘undignified, stagnant and vegetative’ (1853a: 132). ‘English interference’ alone produced ‘the only social revolution ever heard of in Asia’ (ibid.: 131). This stasis is explained by the peculiar social relations of the ‘self-sufficing communities that constantly reproduce themselves in the same form’ (1867a: 358), which provide the economic foundation of ‘Oriental despotism’. Again, the empirical validity of this picture of Asiatic society – a picture whose roots in European (and Euro-centric) thought have been admirably traced by Krader (1975) – is more than dubious. But the illustration none the less suffices to clarify Marx’s position.

Another passage from The German Ideology is equally clear on the fetishism entailed in identifying productive forces with their material embodiments, and says more about the specific social relations – the form of division of labour characteristic of generalised commodity production, one very different from that which prevails in Marx’s portrayal of Indian communities (see 1867a: 257-9) – which grounds this reification. In capitalism, Marx writes:

the productive forces appear as a world for themselves, quite independent of and divorced from the individuals, alongside the individuals; the reason for this is that the individuals, whose forces they are, exist split up and in opposition to one another, whilst, on the other hand, these forces are only real forces in the intercourse and association of these individuals. Thus ... we have a totality of productive forces, which have, as it were, taken on a material form and are for the individuals no longer the forces of the individuals but of private property. (1846a: 83-4, emphasis added)

Marx goes on to argue that ‘the history of the evolving forces’ is in reality ‘therefore, the history of the development of the forces of the individuals themselves’ (ibid.: 90). The Grundrisse similarly describes ‘forces of production and social relations’ as ‘but two different sides of the development of the social individual’ (1858a: 706).

I do not believe we can plausibly interpret ‘forces of individuals’ in such passages as these to refer simply to the things individuals own – means of production and labour-power – as Cohen’s reading would oblige us to. Marx is not speaking here of an external relation of ownership, by people of things (or their own and other people’s labouring capacities). On the contrary, when he says a force of production is a force of individuals he clearly intends that ‘force’ be understood as a power, an attribute, a characteristic, of those individuals in association – of social individuals – albeit a power which may frequently be materialised in things. The relation is internal.

Marx argues similarly in his later writings. The Grundrisse is insistent that labour is ‘the productive force which maintains and multiplies capital, and which thereby becomes the productive force ... of capital, a force belonging to capital itself’ (1858a: 274). Importantly, Marx makes clear that by ‘labour’ here he does not mean labour-power, the commodity which the worker sells the capitalist; the productive force is ‘labour as value-positing activity, as productive labour’ (ibid., emphasis added). This distinction between labour-power and labour is not a casual one; it is fundamental to Marx’s analysis of capitalism as a whole, and basic, amongst other things, to his mature theory of surplus-value. Cohen, acknowledging the distinction, perversely insists (1978: 43-4) that it is labour-power and not labouring activity which is the productive force. One can see why. Labour-power can be conceived as a ‘thing used in production’, as Cohen’s theory requires, in a way that labour-as-activity on his own admission cannot. Marx’s insistence on the contrary, on the other hand, serves only to underline the gulf between his concept and Cohen’s.

Later in the same work Marx argues that ‘all the progress of civilisation, or in other words every increase in the powers of social production [gesellschaftliche Produktivkräfte], if you like, in the productive powers of labour itself ... increases only the productive power of capital ... The productivity of labour becomes the productive force of capital’ (ibid.: 308; cf. 715). In an interesting and subtle elaboration of the argument, Marx goes on to say that those who see here a simple ‘displacement’ of the productive power of labour to capital, or argue that only labour – as distinct from capital – is productive, only partially escape from the fetish. For they forget that ‘capital’ precisely is a social relation, the essence of which is that it subsumes labour and entails ‘the necessary positing of its own powers as alien to the worker’. Such theorists – Marx has in mind Ricardo ‘do not conceive capital in its specific character as form, as a relation of production reflected into itself, but think only about its material substance, raw material, etc. But these material elements do not make capital into capital’ (ibid.: 308-9). Capital is indeed a productive force, in fine, but it is so by virtue not simply of its ‘material substance’, but equally through its particular subsumption or inclusion within itself of labouring activity. It is a productive force, in other words, precisely in its character as a production relation.

One final quotation will suffice to round off the argument. It is taken from Theories of Surplus Value, a text intended by Marx as the fourth volume of Capital, which dates from the early 1860s. The parallels between fetishism of productive forces and commodity fetishism are here made absolutely explicit:

Since living labour – through the exchange between capital and labourer – is incorporated in capital, and appears as an activity belonging to capital from the moment that the labour-process begins, all the productive powers of social labour appear as the productive powers of capital, just as the general social form of labour appears in money as the property of a thing. Thus the productive power of social labour and its special forms now appear as productive powers and forms of capital, of materialised labour, of the material conditions of labour – which, having assumed this independent form, are personified by the capitalist in relation to living labour. Here we have once more the perversion of the relationship, which we have already, in dealing with money, called fetishism. (1863a: 389)

Marx could scarcely be plainer. The ‘traditional’ conception of productive forces defended by Cohen, I submit, splendidly exemplifies this ‘perversion of the relationship’. It illustrates both features of fetishism discussed above: materialisation of the social, and consequent universalisation of the historical. This conception certainly takes productive forces to be not ‘the productive powers of social labour’, but ‘productive powers ... of the material conditions of labour’. Cohen defines productive forces in terms of the latter, as labouring capacity and means of production, and moreover takes pains to insist that these are material as distinct, quite specifically, from social forces. He then generalises this concept of productive forces to all modes of production. A misleading and specifically capitalist appearance – for Marx – thereby gets transmuted into a general theoretical postulate, in ways exactly parallel to those Marx himself criticised in his economic predecessors.

The consequences are serious. Not only does this mystify the productive forces of capitalism itself, at least as Marx conceived them, in ways evidently conducive to capitalist apologetics. Once capital is perceived as a thing, whose contribution to social productivity is independent of that of labour, profit can be understood as its equally independent and thoroughly deserved reward (see Sayer, 1983: ch. 3). The fetishising of Marx’s concept also denies us access to what actually were the productive forces of previous epochs, if by that we understand those things (and ideas, and relations) which actually did enhance the productivity of labour, and thereby give humanity a history which is, despite everything, meaningfully a progress. Marx numbers here such phenomena as ‘the power of Asiatic and Egyptian kings, Etruscan theocrats, &c.’ in relation to ‘the colossal effects of simple co-operation’ evident in such ‘gigantic structures’ as the pyramids (1867a: 333-4; see also the long quote there from Richard Jones).

But, as a famous passage in Capital remarks – à propos commodity fetishism, but the point applies equally to other fetishes as well:

Man’s reflections on the forms of social life, and consequently, also, his scientific analysis of those forms, take a course directly opposite to that of their actual historical development. He begins, post festum, with the results of the process of development ready to hand before him. The characters that stamp products as commodities ... have already acquired the stability of natural, self-understood forms of social life, before man seeks to decipher, not their historical character, for in his eyes they are immutable, but their meaning ... The categories of bourgeois economy consist of such like forms. They are forms of thought expressing with social validity the conditions and relations of a definite, historically determined mode of production. The whole mystery of commodities, all the magic and necromancy that surrounds the products of labour as long as they take the form of commodities, vanishes therefore, as soon as we come to other forms of production. (1867a: 75-6)

It seems that ‘traditional historical materialism’ has yet to learn this lesson, and is no more immune than bourgeois economics to the fetishistic seduction of ‘natural, self-understood forms of social life’ (1867a: 75), in which, for Marx, our ‘social forms of consciousness’ are embedded. One might say of the traditional conception of productive forces, as Marx himself said of political economists’ notions of value, that ‘these formulae... bear it stamped upon them in unmistakable letters that they belong to a state of society, in which the process of production has the mastery over man, instead of being controlled by him’ (1867a: 81). One might also take seriously, sixth and lastly perhaps, his cautionary invocation of Shakespeare’s Dogberry, for whom ‘To be a well-favoured man is the gift of fortune; but reading and writing comes by nature’ (Much Ado AboutNothing, III. iii, as quoted in Marx, ibid.).

Further Reading:
The Relations of Production, (next chapter) | Marx's 1859 Preface
The Concepts of Capital, Geoff Pilling | Marx's Grundrisse and Hegel's Logic, by Hiroshi Uchida