For Marx, value and value production are eminently social qualities, referring to relations between men, and not ‘physical’ attributes adhering to things once and for all. Thus, when Marx writes that the value of a commodity is the embodiment of human labour expended in its production, and when he goes on to say that its value is equal to the socially necessary labour contained within it, he is not making two different statements, but simply repeating the same thesis. For the value of a given commodity is determined only by that portion of labour spent in its production which corresponds to the social average (both the average productivity of labour and the average socially recognized need), that is to say, which is recognized by society as socially necessary labour. Labour expended in the production of a given commodity, but not recognized by society, is not productive of value for the owner of that commodity.

However, precisely because value and the production of value refer ultimately to the distribution and redistribution of the total available labour-power of society engaged in production, that macro-economic aggregate is a basic economic reality, a basic ‘fact of life’. If five million workers work 2,000 hours a year in material production, the total value product is ten billion hours, independently of whether the socially recognized value of each individual commodity is equal to, or larger or smaller than, the actual number of labour hours expended in its production. It follows that if the value of a given commodity is less than the labour actually spent on its production, then there must be at least one other commodity whose value is greater than the quantity of labour actually embodied in it.40 Social recognition of labour expenditure and actual labour expenditure can differ only for individual commodities, not for the total mass.41 In that sense, Morishima is right when he stresses that, in the last analysis, and for the capitalist mode of production (as distinct from petty commodity production), Marx’s law of value is fundamentally an aggregate, macro-economic concept.42

The nexus between the reproduction schemas (and the problem of the circulation of capital in general) and the theory of value leads us back to one of the most hotly disputed issues of Marxist economic theory: the exact delimitation between productive and unproductive labour. As the schemas are value schemas, they express only value production, and automatically exclude economic activities which are not productive of value. What precisely are these activities?

It has to be admitted that the solution of this problem was made more difficult by Marx himself. There are undeniable differences – if only of nuance – between, on the one hand, the long section of Theories of Surplus-Value dealing with the problem of productive and unproductive labour and, on the other, those key passages of Capital (especially Volume 2) which treat the same subject. One striking illustration of this is the analysis of commercial agents and travellers. They are classified as productive workers in the Theories, and as unproductive workers in Capital Volumes 2 and 3.43 In recent years, a long and often confused debate among Marxists has further complicated the matter.44 It is also intertwined with differences in judging the so-called service industries – which, to take one example, are not included in Soviet and East European accounting as contributing to national income, on the basis of a particular interpretation of Marx’s theory of productive labour.45 How then shall we unravel the problem?

A preliminary distinction which we need to draw goes to the heart of the matter. When Marx classifies certain forms of labour as productive and others as unproductive, he is not passing moral judgement or employing criteria of social (or human) usefulness. Nor does he even present this classification as an objective or a-historical one. The object of his analysis is the capitalist mode of production, and he simply determines what is productive and what is unproductive for the functioning, the rationale of that system, and that system alone. In terms of social usefulness or need, a doctor provides labour which is indispensable for the survival of any human society. His labour is thus eminently useful. Nevertheless, it is unproductive labour from the point of view of the production or expansion of capital. By contrast, the production of dum-dum bullets, hard drugs or pornographic magazines is useless and harmful to the overall interests of human society. But as such commodities find ready customers, the surplus-value embodied in them is realized, and capital is reproduced and expanded. The labour expended on them is thus productive labour.

In the framework of this socially determined and historically relativized concept, productive labour may then be defined as all labour which is exchanged against capital and not against revenue, i.e. all labour which enriches one or several capitalists, enabling them to appropriate a portion of the total mass of surplus-value produced by the total mass of value-producing wage-labour.46 We could call it ‘labour productive from the point of view of the individual capitalist(s)’. All wage-labour engaged by capitalist enterprise – as opposed to labour functioning for private households, for consumption needs – falls into that category. This is the level at which Theories of Surplus-Value stops.

But when he returns to the same problem in Capital Volume 2, from the point of view of the capitalist mode of production in its totality, and especially from that of the growth or accumulation of capital, Marx now distinguishes labour productive for capital as a whole from labour productive for the individual capitalist. For capital as a whole, only that labour is productive which increases the total mass of surplus-value. All wage-labour which enables an individual capitalist to appropriate a fraction of the total mass of surplus-value, without adding to that mass, may be ‘productive’ for the commercial, financial or service-sector capitalist whom it allows to participate in the general sharing of the cake. But from the point of view of capital as a whole it is unproductive, because it does not augment the total size of the cake.

Only commodity production makes possible the creation of value and surplus-value. Only within the realm of commodity production, then, is productive labour performed. No new surplus-value can be added in the sphere of circulation and exchange, not to speak of the stock exchange or the bank counter; all that happens there is the redistribution or reapportionment of previously created surplus-value. This point is made clear in Capital Volumes 2 and 3.47 Most of the relevant passages from Volume 2 were drawn by Engels from Manuscripts II and IV. In other words, they were written in 1870 or between 1867 and 1870, some time after the Theories of Surplus-Value of 1861–3 (and even after the rough manuscript of Volume 3), and may therefore be considered to express Marx’s definitive views on the question.