Productive and unproductive labour: an attempt at clarification and classification.

THE RENEWED INTEREST in the foundations of the Marxist critique of political economy since the early seventies has resulted in a closer examination of Marx's distinction between productive and unproductive labour (henceforth PUPL). This closer look has, however, given rise to serious debates on questions of definition, on the criteria to be used and even on the very desirability of such a distinction. The theoretical atmosphere of the entire decade of the seventies, in particular, was after all characterised by the controversy raging between the proponents and the opponents of the so-called Sraffa-based critique of Marx's theory of value. The various viewpoints in the more specific debate on PUPL did not, to be sure, correspond in a strict way to the great divide in the larger debate. Nevertheless, unsparing scrutiny of the fundamentals was the sign of the times. We believe that the particular debate concerning PUPL was abandoned at a point where the issues under discussion were far from resolved. There was, in our opinion, much confusion and misunderstanding between the various sides with respect to the series of distinctions on which the whole idea of a counterposition of productive to unproductive labour is based. Much remained to be clarified. Moreover, at least for those of us who think the distinction PUPL is essential to the analysis of capitalist accumulation, comprehensive classification of all the major types of labour under capitalism is a burning question, especially in the context of the recent increased effort to develop an empirical analysis of the various facets of the capitalist economy on the basis of Marxist categories.(FN1) Hence we face the twin tasks of clarification and classification.

This combined nature of our objective leads us to forgo a polemical approach in putting forth our understanding of the issue. It could, indeed, hardly be considered a fruitful way of tackling the issues if we were to engage in a point by point critical evaluation of all the different viewpoints defended by the various sides to the debate. Nor is a survey of the issues and/or debates the most productive way to take up the twin tasks we pose ourselves. So in what follows we will attempt to provide what we believe to be a systematic and precise, logically sound and theoretically rigorous reconstruction of the distinction PUPL, on the basis of which we will then proceed to a classification of the major types of labour under capitalism. We would like to point out, however, that this reconstruction is so designed as to answer effectively and on logical grounds the major strands of criticism addressed to the overall conception of the distinction PUPL to be found in Marx's work. Moreover, after having completed our presentation we will explicitly take up what we perceive to be the major criticisms and try to show that our conceptual scheme, which is but a careful restatement of Marx's, is immune to the arguments offered by way of criticism.

We believe that Marx's thinking on this issue is entirely coherent. One can, to be sure, point to a careless mistake here or there or to an occasional awkward formula, but the existence of these can reasonably be attributed to the fact that many of his manuscripts were not prepared for final publication by himself. We will not be dealing with the charges, abundant in the literature on the subject, with respect to inconsistencies in his presentation of the issue, not only because that would again take us away from our central objective, but also because, irrespective of whether he himself was consistent or not, what is more important is to see whether a logically and theoretically consistent distinction between PUPL can be made starting out from the foundations of the Marxist theory of value.(FN2).

In what follows, we will first try to point out why the distinction PUPL is important, indeed crucial, both for the analysis of the trajectory of capitalism in general and for an understanding of the impact of some peculiar features of late twentieth century capitalism on the accumulation process. We will then proceed to clarify the distinction and the relationship between the concepts 'productive labour in general and 'productive labour for capital , a step that is methodologically essential to a correct understanding of PUPL. Subsequent sections will deal with all the major types of labour in a capitalist socio-economic formation, ranging from self sufficient peasant households, housework and petty commodity production to hired domestic labour, production and circulation labour, in order to concretize the content of the category productive labour for capital. Separate sections will then be devoted to two thorny questions, i.e. the status of labour in the services sector and state provision of social services. A final section will take up some common criticism to be found in the literature concerning Marx's distinction between PUPL. As we have already pointed out the purpose of that final section is not to provide a critical survey of the existing literature, nor to engage in polemics, but rather to provide some further clarification on questions that have been raised in the discussion on PUPL since the seventies.

Let us then start out with a discussion of the importance of the distinction PUPL, since one aspect of the earlier debate has centred precisely on whether there is any meaningful purpose to be achieved by pursuing this distinction. We believe that far from being esoteric and pointless, the distinction is crucial to an adequate analysis of the trajectory of accumulation in capitalist economies.

THE IMPORTANCE OF THE DISTINCTION BETWEEN PRODUCTIVE AND UNPRODUCTIVE LABOURThe capitalist economy is centred around an accumulation process grounded in the self expansion of value through the continuous production of surplus-value and its subsequent reconversion into capital. The goal of the accumulation process is not only the maintenance of the previously produced value but also the creation and productive reinvestment of surplus-value. However, in order to carry on this process of self expansion, capital has to be continuously exchanged with a certain kind of labour which can produce surplus-value. It is in this sense that 'the distinction between PUPL is vital for accumulation since only the exchange for productive labour can satisfy one of the conditions for the reconversion of surplus-value into capital. (Marx, 1976a: 1048).

While productive labour is essential for the production and reinvestment of surplus-value, unproductive labour does not create surplus-value and hence is not a source of accumulation. Not only that but, as we shall see in the subsequent discussion, the wages of unproductive workers have to be paid out of the surplus-value created by productive workers so that the mass of unproductive labour employed in a capitalist economy is in fact a positive restraint on capital accumulation.(FN3) Hence, PUPL have diametrically opposite consequences for what is the central process of a capitalist economy, i.e. capital accumulation. At this level of generality, then, the distinction PUPL has to be conceived as a fundamental theoretical element of the Marxist analysis of capitalism.

At a more concrete level, the division of total social labour between productive and unproductive uses plays a major role in the determination of the respective magnitude of various crucial variables of the capitalist economy. Foremost among these are variable capital, total surplus-value and consequently the rate of surplus-value. Since variable capital is that element of capital which produces more value than it itself contains, its magnitude at the social level is determined not by the total wage bill paid across the economy but exclusively by the wages of productive workers. Surplus-value, on the other hand, consists not only of its various component parts pocketed by the different fractions of the propertied classes (such as profits, interest, ground rent, etc.) but also includes the wages of unproductive workers. Consequently, the rate of surplus-value, being the ratio of surplus-value to variable capital, cannot be calculated directly by recourse to national income categories such as profits and wages.(FN4) A calculation of this latter consistent with Marxist categories must carefully take into consideration the distinction PUPL.(FN5) As Mandel points out '(a) precise definition of productive labour under capitalism is not only of theoretical importance. It also has major implications for social bookkeeping (calculation in value terms of the national income) . (Mandel, 1981:46).

If the division of total social labour into productive and unproductive uses influences the rate of surplus-value, it immediately follows that it also has an impact on the rate of profit, since the former rate is one of the fundamental determinants of the latter. This result can be expressed in a different manner: The fact that surplus-value is exclusively the product of productive labour implies that an increase in the proportion of unproductive to productive labour entails a reduction in the mass of surplus-value produced and, ceteris paribus, a fall in the rate of profit. This has the further consequence that, at least within the falling rate of profit framework, the distinction between PUPL is essential for an understanding of capitalist crises.

The distinction is also relevant to the analysis of state intervention and the redistribution of income that flows therefrom. This is because a correct evaluation of the net impact of state intervention in the sphere of income distribution can only be based on the precise identification of the specific sources of state revenue. In other words, the calculation of the respective shares of the state revenue derived from variable capital, on the one hand, and from surplus-value, on the other, is crucial in assessing the direction and magnitude of income distribution flowing from state intervention. But we have just pointed out that the magnitude of both variable capital and surplus-value has as one of its determinants the division of total social labour into PUPL. On the expenditure side as well the distinction is necessary.(FN6).

There is another reason why the distinction is important. For reasons that will become clear in the course of this article, the status of labour employed in the service sector with respect to the distinction PUPL is highly controversial. It is, on the other hand, common knowledge that the so-called 'tertiary sector accounts for an evergrowing part of total social labour expended in all capitalist economies, whether advanced or underdeveloped. Understanding what kind of impact this growth of the service sector will have on the overall development of capitalist economies, how, for instance, the rate of profit will be affected, hinges on a correct classification of services between the alternatives PUPL.

These, then, are the reasons why the PUPL distinction is crucial to an understanding of the development of the capitalist economy at a general level. There are, however, also reasons that make the distinction relevant to an understanding of certain recent trends peculiar to late twentieth century capitalism.

Foremost among these is, of course, the explosion of financial services giving rise to what is popularly, but somewhat misleadingly, known as the 'casino economy . In addition to the expansion of brokerage and banking activities due to the proliferation of financial instruments, the unprecedented growth of pension, mutual and hedge funds and the international integration of financial markets, the gradual demise of the so-called welfare state has resulted in a tremendous growth of insurance activities. These recent trends make it imperative to understand the nature of the labour employed in financial sectors (as a subset of circulation activities) in order to assess the impact of this explosion on the accumulation of capital.

A second area where considerable expansion can be observed is a further growth in consumer services, under the combined impact of so-called globalization, technological progress and changes in lifestyles. The explosion in domestic and international broadcasting, the tremendous growth of tourism and the catering business, new mass consumed forms of sports and physical exercise organized along capitalist lines are but the most salient examples of this kind of expansion. The status of services with respect to the distinction PUPL is crucial to an understanding of the positive and negative impact of the expansion of consumer services on capital accumulation.

The recent expansion of consumer services is, after all, a continuation of a trend well established for at least decades. Qualitative change, on the other hand, can be said to have come about in the area of business services as a combined result of technological development, internationalization of capital and a further progress in the division of labour within total social capital. Alongside the explosion of telecommunication services and electronic means of communication, activities such as branding, patenting, marketing, advertising, information related services, public relations services, human resources, training, consultancy and legal services have all led to specific sectors of activity growing at a rapid pace. These activities of different natures have a differential impact on the process of accumulation and the specific impact of each can only be assessed on the basis of a correct understanding of both services and the distinction between production and circulation in relation to the distinction PUPL.

Another crucial recent trend obviously has to do with the structural changes that have come about in the provision of what are commonly regarded as social services, i.e. health, education, housing etc. In this area, the bourgeoisie and its governments have started a sustained drive towards the gradual privatisation of activities which once belonged predominantly to the public sector and the parallel commodification of those services that were until recently provided by the state free of charge. As will subsequently become clear, this implies the transformation of the unproductive labour of state employees into the productive labour of both the workers of private companies in question and the employees of the public agencies that are practically commercialized through the commodification of the services that they provide. At the level of abstraction of this article, this phenomenon increases, ceteris paribus, the magnitude of surplusvalue and thus acts as a counteracting tendency with respect to the general increase in unproductive activities and the consequent decline in the amount of productively investible surplus-value observable in contemporary capitalism. The phenomenon of privatization gives rise to many other important issues related to the general welfare of working people, unemployment, deunionization, the status of the employees of the agencies in question, in short to many facets of class struggle in present-day capitalist society. We submit that these issues are politically very important but fall outside the scope of this paper.

Other novel features of late twentieth century capitalism (some of them with deep historical roots) such as the tremendous growth of subcontracting, the resurgence of the putting out system even in the imperialist countries, the drastic change in the necessary inventory kept due to just-in-time systems and the accompanying change in storage magnitudes, the redrudescence of hired domestic labour etc. all require clarity on the question of the PUPL distinction.

To sum up, then, far from being useless, the categories PUPL are essential to a clear understanding of the functioning of the capitalist economy.(FN7) In what follows we shall try to demonstrate that it is possible to draw a clear and lucid conceptual distinction between the two categories which can then provide the basis for a comprehensive classification of the different types of labour to be found in a capitalist economy.

PRODUCTIVE LABOUR IN GENERALThe starting point for the discussion of the distinction between PUPL has to be a definition of productive labour applicable to all modes of production. It should be made clear from the outset that this is not because productive labour in this sense is identical with productive labour under the capitalist mode of production. On the contrary, as we shall see in the next section, the distinction between these two cocnepts, i.e. productive labour in general and productive labour for capital, is vital for an understanding of the functioning of the capitalist economy. Ironically, however, it is a neglect of the concept productive labour in general that has led astray so many contemporary commentators on the question. It is because insufficient attention has been paid to a general definition of productive labour under all socio-economic formations that all economic activities under capitalism itself have come to be instinctively regarded as productive labour. Only through a clear understanding of the fact that productive labour for capital is a subset of productive labour in general can one avoid this type of theoretical impasse which leads to a total obliteration of the distinction between PUPL under capitalism. This is a question we will study in detail when we take up the criticisms directed at Marx's approach in the last section of the article.

That the distinction between productive labour in general and productive labour for capital is crucial for Marx is obvious from the following passage:.

Only bourgeois narrow-mindedness, which regards the capitalist forms of production as absolute forms--hence as eternal, natural forms of production--can confuse the question of what is productive labour from the standpoint of capital with the question of what labour is productive in general, or what is productive labour in general; and consequently fancy itself very wise in giving the answer that all labour which produces anything at all, which has any kind of result, is by that very fact productive labour. (Marx, 1963: 393)(FN8).

But in order not to confuse the two concepts one has to have a clear definition from the outset of productive labour in general and this is what we shall first attempt to provide.

There are certain activities which, whatever the historical form of social organization, have to be carried out in order to assure biological and social reproduction of the members of society and of the socio-economic formation itself. The definition of productive labour in general presupposes a careful distinction among the different types of such activities. The basic set of activities, which all (or in some cases most) societies have to carry out are the following: production, circulation, distribution of the product (so-called 'income distribution ), personal and social consumption, and the reproduction of the social order.(FN9).

For our present purposes, these activities can first be grouped according to whether they involve the expenditure of labour properly speaking. It is immediately clear that two of these activities, i.e. consumption and distribution of the product, are not activities that involve the expenditure of labour.(FN10) Consumption, whether due to biological or historically developed needs, is an activity which is common to all living species and does not involve labour which is the differentia specifica of the human race. Distribution of the product being closely connected to the relations of production valid under an historically given socio-economic formation is predicated upon the expenditure of labour on the part of the direct producers but is not itself an act which involves labour. The logical proof of this is that the propertied classes, who to the extent that they derive their means of existence exclusively from property income are non-producers,(FN11) also partake of the distribution of the product. The latter concerns a social relationship mediated by the product of labour activity.(FN12).

What is more complicated is the case of the other two activities, that of circulation and that of the reproduction of the social order. Circulation, in the strict sense of the term covering the sphere of the various metamorphoses between the commodity-form and the money-form, is an activity which has been an aspect of economic life throughout all class societies. However, private exchange and the circulation of commodities and of money come into full force only under conditions of capitalist production: Exchange, which is only of partial importance in pre-capitalist class societies because an overwhelming portion of production is the production of use-values and not of commodities, becomes the necessary form of the socialization of private labour under capitalism. As for activities directed to the reproduction of the social order, these again acquire a specific significance in class societies as opposed to primitive communal society. The state as the public sphere of the reproduction of class domination with its administrative, military and financial apparatus, along with religious institutions, is the locus where this type of activity is organized and carried out.

Now it is clear that at least in the modern era a certain part of the agents who carry out these activities of circulation and of the reproduction of the social order are involved with labour of a specific kind. The public employee working for the tax administration or for local government, the bank clerk, the employee of an insurance company, to bring up the most obvious examples, no less toil away their working day than an industrial or agricultural worker.(FN13) However, the labour expended by the former is of a different nature when compared with that expended by the latter. The industrial and agricultural worker is involved with the creation of use-values (i.e. objects which satisfy a certain need, either in the sphere of consumption as articles for personal consumption or in the sphere of production as inputs for productive consumption) and (s)he does this through the transformation of nature. This is, in fact, what distinguishes the last type of activity under discussion, that of production, from the rest of all other social activities: Production is the 'appropriation of nature on the part of an individual with and through a definite form of society (Marx, 1973: 87). Those engaged in production, in other words, mediate the relationship of society to nature. Those, on the other hand, who carry out, within the context of a given social division of labour, the activities of circulation and the reproduction of the social order, simply execute tasks which flow from a historically determined set of socio-economic relations among human beings within a definite society. Production is radically different from all other types of activities in that it is the unique activity which, through an intercourse with nature, provides human society with the indispensable material elements of its reproduction. No society can live on the edicts of kings or on contracts of life insurance without drawing from nature the means of its livelihood. And since it is only through production that these means are acquired, only labour which is engaged in production in this specifically materialist sense can be regarded as productive labour in general. Once human activities are classified accordingly (see Figure 1), it immediately becomes obvious that certain types of activities can under no type of social organization be regarded as productive labour. Social agents who carry out tasks relating exclusively to the reproduction of the social order or to the circulation of commodities and of money are unproductive by definition. The list of such agents is a long one but we should note in passing that priests and all other religious officials, kings, presidents and politicians, public employees of the administrative and financial organs of the state, judges, lawyers and all juridical professionals, generals and soldiers, policemen and prison wardens, to the extent that they are engaged in labour at all, are unproductive labourers in all types of social organizations.

PRODUCTIVE LABOUR FOR CAPITALThe definition of productive labour in general provides only the starting point for an examination of the distinction between PUPL under the relations of the capitalist mode of production. For as a historically determined mode of production, capitalism is first and foremost characterized by the production of commodities and the production of surplus-value. Within this particular formation, production is only carried on with a view to the self-expansion of capital, that is the expansion of capital through the production and appropriation of surplus-value produced by the direct producer, i.e. the wage-labourer.

Hence, a definition of productive labour based on the concrete character of the labour spent in the production process is manifestly insufficient within the context of capitalism. The result of capitalist production is not simply use-value, but specifically exchange-value and surplus-value. The capitalist production process is not a simple labour process, but much more decisively a process of valorization, of self-expansion, the latter subsuming the former under its requirements. In Marx's words:.

Since the immediate purposes and the authentic product of capitalist production is surplus-value, labour is only productive, and an exponent of labour-power is only a productive worker, if it or he creates surplus-value directly, i.e. the only productive labour is that which is directly consumed in the course of production for the valorization of capital. (Marx 1976a: 1938).

Hence, the specific nature of the capital relation provides the clue to the distinction between PUPL from the standpoint of capital. Productive labour for capital is that labour which produces surplus-value.(FN14) In other words, productive labour within the context of a determined mode of production cannot be defined solely on the basis of the interaction of humanity with nature. A meaningful definition has to incorporate, along with this general determination, characteristics which are specific to the social relations that are dominant within that mode of production.(FN15).

This simple definition, however, is only a starting point since many problems and confusions arise once the distinction is applied to more concrete phenomena. It is, nonetheless, the indispensable foundation on which alone can be based an adequate discussion of the distinction in question. It also immediately shows why productive labour for capital is, as we have already noted, a subset of productive labour in general. Once we remind ourselves that surplus-value can only be produced in the immediate process of production and has no other source, it becomes obvious that only labour which is productive in general, that is only labour which produces use-values through the purposeful transformation and appropriation of nature, can produce surplus-value. In other words, the quality of being productive in general is a necessary (but not sufficient) condition for labour to be productive for capital (the sufficient condition is being productive of surplus-value).(FN16).

The criterion of the production of surplus-value provides us with certain guidelines in our quest for a more detailed and concrete distinction between PUPL. Several types of labour can immediately be shown to be irrelevant to this distinction. First of all, it is obvious that labour expended with the sole purpose of producing use-values, that is labour not engaged in the production of commodities, cannot be regarded as productive labour from the standpoint of capital for the very simple reason that the production of surplus-value is predicated upon the production and exchange of commodities. Hence, the labour of self-sufficient peasant households or housework under capitalism cannot serve as productive labour.(FN17).

Secondly, the criterion of the production of surplus-value also rules out the labour of petty commodity producers. Petty commodity production is, by definition, based on the ownership of the means of production by the direct producers themselves and hence on the exchange of the products of their labour rather than the sale of their labour-power. Surplus-value can only be produced if labour-power is sold as a commodity and the buyer of this specific commodity consumes its unique use-value, i.e. the capacity to produce more value than it embodies. Hence, only labour which is predicated upon the sale of labour-power as a commodity can serve as productive labour under capitalism. This removes the labour of the artisan and of the small holding peasant outside the picture in a discussion of PUPL under capitalism.

Likewise, of the different methods of outsourcing, more and more a feature of present-day capitalism, that based on the organization of the producers' (predominantly women's and children's) labour at home (similar to the 'putting out system of capitalism at its dawn) is not subsumed under a capitalist form per se, even in those cases where the instruments of labour and raw materials are provided by the capitalist. Hence labour so expended is not productive in the capitalist sense. Subcontracting, on the other hand, still another form of outsourcing, is usually based on the activities of capitalist firms, albeit small, and the labour involved here is productive or unproductive, depending on the specific type of activity carried out (see below).

Even the sale of labour-power, however, is not a sufficient condition for the existence of productive labour. And here we come to the distinction PUPL properly speaking. For this, one should also make a distinction between labour exchanged against capital, on the one hand, and that exchanged against revenue, on the other. This distinction was, in fact, the point around which the original debate on the definition of productive labour turned in the late 18th and throughout the 19th centuries. Without going into the intricacies of that debate, the distinction can be expressed in a straightforward manner through the use of the general formula for the circuit of capital and of revenue.

The formula for the circuit of capital, M - C ... P ... C' - M' is the most general representation of the process of self-expansion (the process of valorization) of capital. Within the framework of this general representation it is easy to see that labour-power as a commodity (C) is exchanged with a part of M (money functioning as capital): that is, it is exchanged against capital. Subsequently, in the process of production (represented by P) it not only reproduces the equivalent of its original value but also produces a new magnitude of value appropriated by capital as surplus-value.

This is the type of relationship between labour-power and capital which transforms labour into productive labour for capital. In Marx's words:.

Productive labour is exchanged directly for money as capital, i.e. for money which is intrinsically capital, which is destined to function as capital. Thus productive labour is labour which for the worker only reproduces the value of his labour-power as determined beforehand, while as a value-creating activity it valorizes capital and confronts the worker with the values so created and transformed into capital. The specific relationship between objectified and living labour that converts the former into capital also turns the latter into productive labour.

(Marx, 1976: 1043).

However, this is not the only possible type of exchange between labour-power and money (in other words, it is not the only possible type of the sale of labour-power). Consider the next period of production: the surplus-value appropriated by capital (i.e. M' minus M) is now reconverted into capital (that is the sense in which Marx in the above quotation says that it 'confronts the worker ). But not the entire surplus-value produced is converted into capital; even abstracting from the appropriation of certain portions of it by other capitalists and landlords in the form of interest, commercial profit, ground rent, etc., a certain part of it is, of necessity, converted into revenue for the industrial capitalists with the purpose of consumption. Now this revenue can be consumed in different ways: it can either be spent on articles of consumption of all sorts, and/or it can be used to employ workers (domestic servants, cooks, drivers, gardeners, etc.) to make life more comfortable for the capitalists and their families. In the former case, capitalists' revenue is exchanged against commodities, while in the latter it is exchanged against labour-power. Hence we have here a second type of exchange between money and labour-power as a commodity. (Marx, 1976a: 1041) This type of exchange, however, in contradistinction to the exchange of labour-power against capital, does not create a surplus-value for the capitalist. The use-value of the labour-power in question is consumed in the form of personal service by the capitalist and his family. The product of the labour does not take the form of a commodity which can then be converted into money incorporating surplus-value. Hence, labour-power exchanged against revenue, as opposed to that exchanged against capital, does not produce surplus-value and is, therefore, unproductive from the standpoint of capital. Thirdly, therefore, only labour-power exchanged against capital can serve as the source of productive labour under capitalism.

One can, of course, see right away why this specific distinction was of such great importance to classical political economy (Adam Smith, in particular, but also Ricardo and others). The use of domestic servants was quite widespread in that early period of the development of capitalism.(FN18) Moreover, it was a much more common practice among landlords and the nobility than among the more frugal manufacturer of early capitalism. Classical political economy found the distinction particularly useful as an instrument in its attack on landed interests on behalf of the industrial bourgeoisie. Things have become a bit more complex since then: While the use of domestic servants is still a structural social feature of underdeveloped countries, it had been diminishing to a considerable extent in the imperialist heartlands until recently, when it started to grow again with the rise of unemployment and social inequality. Whatever the present scale of hired domestic service, though, this distinction between labour exchanged against capital and labour exchanged against revenue has now lost its overpowering importance in favour of another problem, to which we now turn.

The distinction exchange with capital/exchange with revenue does not exhaust the difference between PUPL. At the stage we have reached, the criteria for productive labour can be summarized as (1) that of commodity production, (2) that of the sale of labour-power, and (3) that of the exchange of labour-power against capital as opposed to its exchange against revenue. Yet these three conditions are still not sufficient to guarantee the production of surplus-value. Put in another way, there may be an important range of cases in which the sale of labour-power does not result in the production of surplus-value, and hence labour is not transformed into productive labour, even though these three conditions may have come together. The secret of the paradox lies, of course, in the distinction between production and circulation.

Let us go back to the circuit of capital, M - C ... P ... C' - M'. In order to complete this circuit and reproduce itself, capital has to go through a certain number of metamorphoses or changes of form; at first, it appears in the form of money capital ready to buy labour power and the elements of constant capital; then with the purchase of these elements of capital, it is converted into productive capital which serves as the basis at once of the labour process and the process of valorization of capital through which commodities of a higher value are produced by the extraction of surplus labour from the workers; these commodities are then thrown onto the market where capital appears as commodity capital. Once these commodities are sold on the market, the whole process returns to its starting point, capital is once again converted into money capital, expanded of course by the amount of surplus-value extracted, ready to start the process anew. Since in order to expand, capital has to go through this whole process, Marx calls these forms of appearance of the one and the same capital its 'functional forms. .

Now, the point which has to be noted carefully is that it is only during a single phase of this process, that of productive capital, that surplus-value is produced. In the other moments of this constant metamorphosis of capital, there is only a change of form between commodities and money, and hence, no creation of value or of surplus-value. Hence, the tasks carried out in these specific moments of the overall process do not pertain to production per se and are, therefore, unproductive circulation tasks by definition. In practice, the same worker may carry out both productive and unproductive tasks and this does cause difficulties of measurement at an empirical level. But at the conceptual level there exists a clear and unambiguous distinction between the two types of activity. Moreover, the empirical measurement aspect is itself simplified by a certain type of division of labour among capitals with the advance of capitalism.

In the course of historical development, different units of capital specialize on an increasing scale in the different functions carried out originally by the same capital in its diverse functional forms. It is thus that a certain part of the functions of money capital is taken over by interest bearing capital (commercial and investment banks, brokerage firms, mortgage companies, insurance and reassurance companies, etc.) and the function of commodity capital by commercial capital (wholesale merchants, department stores, other large outlets, retailers etc.). In order to carry out these tasks, themselves necessary and integral parts of the overall process of the reproduction of capital, these units of capital have to employ wage workers. The question immediately arises: are these workers productive workers and is their labour productive? The whole analysis presented in this paper points to a negative answer to this question.

The activities carried out by these workers are purely circulation activities and therefore by their very nature nonproductive in the general sense. Moreover, the analysis of the circuit of capital has already shown us that these workers, because they are engaged in circulation tasks, or more precisely to the extent that they are engaged in circulation tasks, do not produce surplus-value. The conclusion, then, is inevitable: Workers employed by capital working in the sphere of circulation are unproductive as is their labour. In Marx's words:.

If by a division of labour a function, unproductive in itself although a necessary element of reproduction is transformed from an incidental occupation of many into the exclusive occupation of a few, into their special business, the nature of the function itself has not changed. (Marx, 1956: 134).

Those who regard circulation tasks as productive because these tasks are necessary for the overall process of reproduction seem to forget that the necessity of certain tasks under certain types of social organization is no sound proof of their productive nature. Policemen are no less 'necessary for the capitalist state in order to preserve the much cherished 'law and order of bourgeois society, but no one (including those authors who regard circulation activities as productive) considers them to be productive. In the same vein, circulation activities in the strict sense of the term are not an inseparable ingredient of production in general but are only necessary under the given conditions of capitalism and its indissociable companion, generalized commodity production. Now, however necessary circulation may be for capitalism, however important the fetishistic forms of exchange, money, credit, etc. for this type of economy, it should be remembered that all this is grounded in the production of surplus-value. And capital cannot appropriate as surplus-value but a certain portion of the material wealth of society produced in a certain period. No society, not even capitalist society despite its fetishized relationships, can increase its wealth, and thereby the amount of products to be distributed, by changing the form of this wealth into money and then back into commodities. The labour so expended is, in fact, an additional social cost of production due to the in-built wastefulness of capitalism. It is no less a drain on the social surplus produced in the immediate production process. Marx expresses this idea in the following manner:.

(A buying and selling) agent expends his labour-power and labour-time in the operations of C - M and M - C. And he makes his living in that way, just as another does by spinning or by making pills. He performs a necessary function, because the process of reproduction itself includes unproductive functions. He works as well as the next man but intrinsically his labour creates neither value nor product. He belongs himself to the faux frais of production. His usefulness does not consist in transforming an unproductive function into a productive one, nor unproductive into productive labour. It would be a miracle if such a transformation could be accomplished by the mere transfer of a function. (Marx, 1956: 134-5).

It is true that certain aspects of circulation (by allowing, for instance, a more rapid turnover of capital) can contribute to an overall increase in production. This increase in production finds its specifically capitalist expression in the additional mass of surplus-value produced by the workers employed in production thanks to this more rapid turnover. This additional amount of surplus-value is thus the product of the labour of these production workers. Were it to be equally attributed, even indirectly, to the productivity of the workers of the sphere of circulation, we would be face to face with a blatant case of double counting because the one and the same increase in surplus-value would then have been attributed to two distinct sets of workers. What has in fact happened is the following: circulation workers, through their specific activities, have simply made an additional amount of capital available for valorization by releasing this amount from the unproductive tasks of circulation (which is itself testimony that circulation tasks are unproductive). Their specific contribution is, therefore, that of having added to the overall amount of money functioning as capital. This being the case, to regard the activities of circulation workers as productive would simply amount to attributing the magical power of productivity to objectified labour and to the fetishized category of capital. In other words, a more rapid turnover of capital implies the existence of more capital in a given period of time available for exploiting workers in the production process: The surplus-value additionally created is exclusively the product of the labour of these production workers in the same manner as the surplus-value created in the absence of this more rapid turnover of capital.

Two brief remarks are in order before we draw the conclusions of this distinction between the spheres of production and circulation. First, transportation and storage activities should in no way be conflated with the sphere of circulation.(FN19) These are essentially a necessary element of the production process itself under any form of social production and, a fortiori, in those socio-economic formations based on an advanced social and geographic division of labour.(FN20) No consumer can consume, nor a producer use as an input, a commodity produced in another part of the world but not transported to where it is to be put to use. Hence, to the extent that transportation is the indispensable final stage of production, the labour employed in this sector of activity is productive labour (in so far, of course, as it is employed by capital).(FN21) However, those transportation and storage activities which are due purely to motives peculiar to circulation (e.g. speculation or re-exportation due to differential government regulation) are immaterial to the production process and the labour employed therein counts as unproductive.

Secondly, it is obvious that the workers employed by capital in the sphere of circulation make it possible for this fraction of capital to pocket profits. It is in fact this aspect of the matter which complicates the question by making it seem that these workers produce surplus-value for their respective capitalists. The fact of the matter is that the profit obtained by these capitalists (commercial profit and interest) is simply a portion of the total surplus-value produced in the sphere of production. What is more, even the wages of the workers employed by circulation capital are themselves paid out of surplus-value within the immediate process of production.

The distinction between production capital and circulation capital finally gives us the sufficient condition for the definition of productive labour. All labour that is exchanged against capital employed in the sphere of production is productive for capital. We have thus, through a series of concrete distinctions concerning use-value production vs. commodity production, exchange of commodities vs. exchange of labour-power, exchange against revenue vs. exchange against capital, and finally, exchange against circulation capital vs. exchange against production capital reached our original definition of production labour under capitalism as labour that produces surplus-value for capital.

It is our opinion that this type of procedure in defining productive as opposed to unproductive labour has the merit of avoiding a confusion between criteria of different orders. The criterion of the production of surplus-value is the absolute guiding principle in a discussion of productive labour under capitalism. Other criteria such as exchange against revenue vs. exchange against capital or circulation vs. production are of a different order: They are, each of them, only partial in the sense that none of them is, on its own, a sufficient condition for a comprehensive definition of productive labour and unproductive labour.

The foregoing discussion enables us to classify labour carried out under different types of social and technical relations within a concrete, historically given capitalist economy. Figure 2, summing up the argument thus far, shows us that a large portion of the total labour of society, although necessary for the reproduction of the system in its totality, does not perform productive labour for capital and hence does not contribute directly to the accumulation process.

Figure 2 should be of some assistance in gaining conceptual clarity on the question of the distinction between PUPL. However, there are still other categories of workers within capitalist society which we have not discussed. Two of these categories stand out by the sheer importance of their size and by the confusion in the literature concerning their status vis-à-vis the distinction under discussion: These are, first, workers in the service sector and, secondly, wage labourers employed by the state. We will now discuss these two categories in turn.

THE PROBLEM OF SERVICESOne of the most contentious areas in the discussion of PUPL has always been the treatment of services. Even authors who otherwise are in perfect agreement on other aspects of the question have been found to be at loggerheads on this specific aspect.(FN22).

It is our contention that a major part of the confusion and misunderstanding concerning the status of services is due to the ambiguity of the concept 'service itself. This concept, as we shall shortly see, designates two very different types of things and failure to distinguish between the two leads to errors and confusion. In fact, the problem goes all the way back to Adam Smith and that is where we propose to begin in order to clarify the problem.

It is widely known that there are two easily distinguishable definitions of productive labour in Smith's work. The first one, quoted below, puts a stress on capitalist social relations in a characteristically Smithian fashion:.

...(T)he labour of a manufacturer adds, generally, to the value of the materials which he works upon, that of his own maintenance, and of his master's profit. The labour of a menial servant, on the contrary, adds to the value of nothing. Though the manufacturer has his wages advanced to him by his master, he, in reality, costs him no expense, the value of those wages being generally restored, together with a profit, in the improved value of the subject upon which his labour is bestowed. But the maintenance of a menial servant never is restored. A man grows rich by employing a multitude of manufacturers: he grows poor, by maintaining a multitude of menial servants (Smith, 1976: 351).(FN23).

Smith's second definition, however, emphasized the intrinsic quality of the labour in question:.

...(T)he labour of the manufacturer fixes or realizes itself in some particular subject or vendible commodity, which lasts for some time at least after the labour is past. It is, as it were, a certain quantity of labour stocked and stored up to be employed, if necessary, upon some other occasion. That subject, or what is the same thing, the price of that subject, can afterwards, if necessary, put into motion a quantity of labour equal to that which had originally produced it. The labour of the menial servant, on the contrary, does not fix or realize itself in any particular subject or vendible commodity. His services generally perish in the very instance of their performance, and seldom leave any trace or value behind them, for which an equal quantity of service could afterwards be procured (Smith, 1976: 351-2).

Marx was of course quick in pointing out the error in the second definition: The material aspect of the product of labour or the concrete content of the labour itself has nothing to do with the definition of productive labour (so long, of course, as the labour in question produces a use value through the appropriation of nature).

...(T)he designation of labour as productive labour has absolutely nothing to do with the determinate content of the labour, its special utility, or the particular use in which it manifests. The same kind of labour may be productive or unproductive. (Marx, 1963: 401)(FN24).

So long as the labour in question transforms a particular aspect of nature with the purpose of satisfying a need, so long, that is, this activity is an aspect of production in general, labour engaged in such a process can, if it is employed by capital, serve as productive labour. This implies that activities that are commonly regarded as services, such as education, health provision, catering, art performance, hairdressing, etc. can also be the basis of the extraction of surplus-value and, therefore, of the existence of productive labour. In Marx's words:.

The only worker who is productive is one who produces surplus-value for the capitalist, or in other words contributes towards the self-valorization of capital. If we may take an example from outside the sphere of material production, a schoolmaster is a productive worker when, in addition to belabouring the heads of pupils, he works himself into the ground to enrich the owner of the school. That the latter has laid out his cpaital in a teaching factory, instead of a sausage factory, makes no difference to the relation. (Marx, 1976b: 644).

This criticism of Marx against the second definition of Smith is widely acknowledged to be valid. What is generally disregarded, however, is the fact that in this second definition Smith treats as identical two different concepts of 'service . In order to see this, let us go back and read the crucial sentence which embodies this conflation: 'The labour of the menial servant, on the contrary, does not fix or realize itself in any particular subject or vendible commodity . Smith is here talking of two different things at once. On the one hand, he is talking about a definite social relationship: The concept 'menial servant implies that the labourer in question is connected to his/her master within a definite relationship, i.e. the exchange of his/her labour-power against the revenue of the master. On the other hand, he is talking about the nature of the product of the labourer in question, i.e. about the fact that a service is not materialized in an object. His formulation implies that those two very different things are necessarily related as if only 'menial servants could provide services and as if they could provide nothing but services. It is obvious that there is no such necessary relationship. 'Menial servants can produce material products; for instance, a wealthy person may employ a tailor on a permanent basis. The tailor in this case will not be producing surplus-value not because the product of his labour is intangible but because this product is not sold as a commodity. But much more important than this somewhat marginal possibility is the fact that services can be provided not only by 'menial servants but also by wage workers employed by capitalists. We in this age of the increasing 'tertiarization of capitalist economies should certainly know better.

Unfortunately, errors are obstinate. This conflation of the two meanings of the term 'services was later to be repeated by many Marxists, leading them to a rejection of the possibility of conceiving labour employed in services as productive (and this despite Marx's numerous warnings to the contrary). A striking example is to be found in Poulantzas' discussion of PUPL.

Also to be considered as unproductive labour is that taking the form of services, whose products and activities are directly consumed as use-values and which are not exchanged against capital but rather against revenue or income.... (Poulantzas, 1975: 213)(FN25).

It is obvious that Poulantzas commits the same error as Smith. That is, he imagines a necessary connection between the immaterial nature of the product of labour and the specific social relationship deriving from the fact that wages are being paid out of revenue. The social relationship in question immediately implies that the labour expended is unproductive; the false necessary connection then leads to the corollary that all labour which provides services, even when it is expended within the framework of a capitalist firm, is unproductive.(FN26).

Having identified the source of the confusion, let us now try to define the concept 'services in a rigorous manner. The first thing to be noted is that in the case of a domestic servant etc., what is exchanged for money (for a wage) is labour-power itself, while in the case of capitalistically organized services (a private school, a hotel etc.) the 'service is the product of labour and not labour-power. In other words, the service is a commodity as good as any material object.(FN27).

This is totally consistent with a second point which goes to reinforce the first. It is not only labour but numerous other inputs which go into the production of a service: For instance, a person who has a hair cut buys not only the labour time of the barber but also 'scissors time , 'chair time , 'mirror time , so to speak--so many material inputs without which the service could not have been produced. Thirdly, not only on the input side but equally on the output side a service is not made up of only labour: what are commonly called services are more often than not materialized in a good of corporeal qualities; e.g. a beef Stroganov, a clean hotel bed, a cleaned coat instead of a dirty one, etc.

Hence, what defines a service is not the fact of buying either the labour-power or the labour of the worker in question. What defines it is the simultaneity of production and consumption, independently of the material form or otherwise of the commodity, that is of the product of labour.(FN28).

Since what is sold is therefore a commodity, the consumption point of view is irrelevant to the discussion of the distinction PUPL. If the correct point of view, that of production, is taken, it can immediately be seen that the labour-power of the service worker is bought (and hence his/her wages paid) not by the customer-consumer but by the capitalist who organizes the production of services and that these workers produce surplus-value. Labour so expended is obviously productive labour.

This whole discussion implies that the recent growth in consumer services (with the exception naturally of financial services, e.g. credit cards) should in no way be regarded as a growth in the share of unproductive labour in total social labour. The case of business services is more complicated, the status of each of these depending on whether the service in question is one relating to a productive function or to a circulation activity. Clear examples of the former are human resource and training services (but not temp agencies or head hunters) and some but not all information services. As to the latter, the most clear cut cases are of course marketing, advertising and financial consultancy services.

STATE EMPLOYEESThe secular expansion of the state sector both in the imperialist and in the dependent, underdeveloped regions of the capitalist world makes imperative an assessment of the nature of the labour of state employees. Fortunately, the criteria we have thus far proposed for the distinction PUPL provide sufficient basis for a solid evaluation of this question.

Any discussion concerning the nature of the labour of state employees has to start out by noting the diversity of state activities under capitalism. For our present purposes, the various activities of the state can be regrouped under three headings. There are, first, those which relate exclusively to the reproduction of the social order, such as the activities of the administrative bureaucracy of central and local government, the military, the courts, the police force, the prison system, etc. The second type of activity is the organization of production activities in the framework of corporations and companies owned wholly or partially by the state, central or local. Finally, there is a growing section of state activities directed to the provision of social services, services which are related to the so-called 'Welfare State , i.e. education, health, housing, etc.

Given our earlier criteria for the definition of productive labour, it is quite easy to identify the nature of the labour of state employees in these three areas (although, of course, at a more concrete level there are bound to come up complex border line cases which do not fit easily into any one of these groups). Those employees who carry out tasks which are directly related to the production of the social order are unproductive labourers by definition. Their labour is not productive in the general sense, as we have seen. It does not act upon nature to transform certain aspects into use-values with a view to satisfying human needs, directly or indirectly. The tasks these people carry out are meant to serve the survival and reproduction of a conflict ridden society based on class exploitation and gender and racial oppression. The wages and salaries of people so employed should therefore be simply regarded as the faux frais of class society, in this case of capitalism.

In order to be able to assess the nature of the labour of the second group of state employees, it is necessary to understand that, whatever other differences may exist between production organized by private capital and that organized within the state sector, there is no difference whatsoever from the point of view of the production of surplus-value. State enterprises are capitalist enterprises which employ workers in order to organize a labour process with a view to extract surplus-value. Political tampering with the prices of their output, overmanning for political purposes, chronic losses and similar phenomena which distinguish the behaviour of state enterprises from private capitalist firms do not, in the least, change the fact that these are capitalist firms which produce commodities on the basis of the exploitation of wage labour, whether the surplus-value thus produced is pocketed by the state enterprises themselves or by other capitalists. If this is the case, it follows that workers employed by state enterprises in the production sphere should be considered, exactly like those employed by private firms, as productive labourers.

The most difficult case to deal with is the third category. On the one hand, social services are (usually) not sold as commodities on the market and therefore the national education system or the national health service of a capitalist country cannot be regarded as capitalist enterprises. Consequently, the workers they employ cannot be classified as productive labourers. On the other hand, teachers, doctors, nurses and other health workers do produce use-values (services) with a view to the satisfaction of human needs and are, in that sense, in a different position than those wage-labourers (such as prison wardens or tax-collectors) whose exclusive task is the reproduction of the existing social order. Thus, as opposed to the latter whose labour is unproductive by definition (because it is not an element of productive labour in general), the labour of the former (that is of healthworkers, etc.) is unproductive in a contingent sense, as a result of the nature of the social relationship within which their labour is organized. That is to say, the labour of a tax-collector cannot be considered to be productive under any type of social organization, while the labour of the health worker can, depending on the circumstances of its expenditure, be productive or unproductive. The former is the case when medical services are so organized as to be sold on the market and are thereby transformed into commodities that bring the owner of the hospital a profit. Under such circumstances, the hospital, clinic etc. become capitalist firms and the labour of the health workers becomes productive. The problem becomes identical with that of defining the nature of the labour of workers in the capitalistically organized service sector, the sole difference due to the different nature of the employer (state vs. private) being, as we have just seen, irrelevant to the question we are examining.

The foregoing has certain implications concerning recent trends in capitalist societies. The widespread assault on the so-called welfare state has resulted in a shift from the unproductive labour of state employees producing public services free of charge for the recipient to the productive labour of the employees of private hospitals, schools etc. in the case of the privatisation of social services. As for the practice of social services being offered not free of charge but on the basis of fees, tuition etc., the more the fees in question come closer to market or 'shadow prices of the service in question, the closer the hospital, school, university etc., in question approximates to a capitalist enterprise and the more the employees of such establishments become productive labourers. The point after which this becomes the case is a question of empirical methodology for which different criteria can be offered but the choice between these criteria need not detain us in the context of this article.

Hence, the status of the labour of wage workers engaged in the provision of social services has to be decided on the basis of concrete social conditions which may differ from one country or from one period to another.

SOME MISCONCEPTIONS REGARDING PRODUCTIVE AND UNPRODUCTIVE LABOURAs we stated at the outset, a general survey of the literature on PUPL lies outside the scope of this article. More relevant for our purposes are the major types of criticism addressed to Marx's conception of this distinction, since this conception is what lies at the basis of our own attempt at classification. To attempt to tackle all the different types of criticism directed at Marx's conception would require a detailed point by point response which can only be the subject of another article. What we intend to do in this section instead is to take up the two most important areas where the controversy has concentrated, i.e. the characterization of circulation activities and the position of those workers whose activities have an indirect impact on the mass and rate of surplus-value. Most, if not all, of the issues raised by the critics are directly or indirectly related to these two areas so that an examination of these will provide us with the opportunity to touch upon the most sensitive aspects of an involved and complex controversy. We will not refer to specific authors as many of the critics share common positions on these questions although their argumentation may vary in detail.(FN29).

To take up the question of circulation activities first, the position of many critics on this issue is that what are commonly regarded as circulation activities are no different in nature from production activities and that therefore to classify them as unproductive is unwarranted. There are several arguments put forward in order to substantiate this view. Let us take them up one by one.

The first argument is that circulation workers contribute in their own way to the production of use-values. An example cited is that of salespeople who carry out tasks such as classification and display of available commodities, provision of knowledge and guidance to the consumer etc. This argument is based on a misunderstanding with respect to what activities are considered circulation activities by the proponents of a rigorous distinction between production and circulation. It is true to say (and Marx himself said it on so many occasions) that any activity which is a necessary component of the process whereby the consumer gains access to the object of consumption should be considered to be a productive activity. But precisely because this activity is a necessary link in the chain between the point of production and the consumer, it is part of the production process on an equal footing for instance with transportation. The sphere of circulation is restricted to those activities which are instrumental exclusively in the change of form between commodities and money. At a conceptual level, those aspects of sales activity which are necessary for the completion of the chain from producer to consumer can quite clearly be distinguished from pure circulation activities. The best example would be the function of cashiers. Their whole activity relates to sale and purchase and is by no means a necessary link in the chain. (Any recording necessary for purposes of social bookkeeping could be carried out in alternative ways, e.g. through computers the consumers themselves use.) The cashier is of course one example among many since an overwhelming part of the functions of the personnel in any capitalist commercial enterprise relates to pure circulation activities. It is true that at the empirical level some difficulty may arise when calculating the ratio of production activities to circulation activities in commercial enterprises, but at a conceptual level the distinction is crystal clear. And that is all that is needed for our present purposes. (The careful reader will already have noted that when we discussed above the status of workers employed by circulation capital, we explicitly stated that 'these workers ... to the extent that they are engaged in circulation tasks, do not produce surplus-value. ).

A second argument in favour of considering circulation activities as productive is to contend that these activities are necessary in all societies, thereby implying that they should be considered productive. Leaving aside for the moment the tenuous connection between the premise and the conclusion (to which we will return in another context below), it can be said that this argument of universality is in fact a more general and theoretical expression of the misunderstanding concerning the distinction between production and circulation which we pointed out with respect to the first argument. The idea that circulation is necessary in all types of societies (strictly speaking, one should restrict this statement to those societies with an advanced division of labour) obviously refers to the indisputable fact that in every such society there arises the inevitable necessity of flows of use-values between the different branches of production and between producers and consumers. But unless one ascribes immutability to capitalist forms of production and distribution, there is no reason to suppose that this flow of use-values will necessarily assume in all societies the form of the purchase and the sale of commodities. Whatever one's views on the efficiency and desirability of central planning, it is surely not debatable that this is a possible alternative form of the interrelationships among the different branches of economic activity. The error in question is hence transparent: this type of criticism confuses and conflates in an unjustifiable manner the circulation of use-values and the circulation of commodities, money and capital. The former is certainly a necessary aspect of production in all advanced socio-economic structures; the latter is historically transitory and, what is more important, involves much more than the provision of supplies to the various branches and to consumers--in fact, taken in its pure form, it is a totally different type of activity which concentrates exclusively on change of form.

While this last argument contends that circulation should be treated on a par with production because it is necessary in all kinds of advanced social formations, there is another line of reasoning which accepts the idea that pure circulation in Marx's sense (i.e. circulation of commodities, money and capital) is peculiar to the capitalist economy, but then proceeds to criticize its exclusion from the range of productive activities on the ground that to do so would be normative, evaluative, moralistic (one would be saying 'there can be a more rational order from the standpoint of which these activities are unproductive ), and therefore regards PUPL to be unacceptable as a scientific distinction. We leave aside the complex debate concerning the relationship between social science and so-called value judgments. It is true that Marx regards the expenses of circulation (which obviously include the wages of workers employed in this sphere) as the faux frais of capitalism, as an expression of its irrational nature, as one of its 'evils etc. But the crux of the matter is the following: It is not because capitalism with its plethora of commercial, banking, brokerage etc. activities is irrational relative to communist society that these circulation activities under capitalism are deemed unproductive. On the contrary, it is because these activities are unproductive that capitalism is regarded as irrational. What then is the ground of saying that they are unproductive? We have already examined this in detail but by way of recapitulation it must be emphasized that a society can only increase its wealth through the purposeful transformation of nature and only that amount that has thus been produced can be distributed among the individual members or social classes of a society. No amount of exchanging parts of the social product already produced can increase this product itself. Circulation activities in the strict sense of the term do nothing but that. Of course, the nature of capitalism renders this type of labour necessary but this does not change in the least that these activities do not increase overall production.(FN30) The characterization of labour expended in the sphere of circulation as unproductive labour is nothing but a logical extension of the Marxist distinction between the spheres of production and circulation. Whoever accepts the latter distinction has to accept equally the former. It is indeed astonishing to find someone saying simultaneously that there can be no production of surplus-value in circulation and that labour employed in circulation is productive, this being defined as labour that produces surplus-value. This position defies all logic. To conclude this point then: The distinction PUPL has nothing normative, evaluative or moralistic about it any more than the prior distinction between production and circulation.

One final line of argument that is commonly used in order to prove that circulation activities are no different from production activities from the standpoint of the distinction PUPL is to point out an alleged contradiction in Marx's thinking. Critics recall that Marx constantly emphasized that the distinction PUPL is indifferent to the type of use-value produced. They then proceed to claim that circulation activities result in specific usevalues which are admittedly of a different nature but, on Marx's own admission, this difference in nature should not be a basis for excluding this type of labour from the overall set of productive labour. It is here that the full force of our earlier insistence on the importance of the concept 'productive labour in general reveals itself. To recapitulate, the quality of being productive in general is a necessary (though not sufficient) condition for labour to be productive for capital. This means that any activity which is not directly necessary for humanity's intercourse with nature in order to transform aspects of it in accordance with human needs cannot be regarded as productive labour in general, nor, therefore, as productive labour under capitalism. In other words, this double determination of the concept productive labour implies that productive labour under capitalism is a subset of productive labour in general. On the basis of this understanding it is a simple task to prove that the critics are wrong in accusing Marx of contradicting his own definition. Since productive labour for capital is a subset of productive labour in general, it follows that no activity that falls outside of the latter can be deemed productive. As we stressed above such an activity is by definition unproductive in all types of socio-economic organization. Therefore, each and every one of Marx's propositions concerning productive labour for capital can only cover those activities which lie within the set of types of labour which are productive in general. In other words, only labour which is productive in general may or may not be productive for capital depending on contingent conditions. Hence, when Marx emphasizes that the distinction PUPL is independent of and indifferent to the type of use-value produced, his statement is, by the logic of his whole position, restricted to that set of use-values which corresponds to the set of productive labour in general. It does not cover use-values 'produced" (the term is eminently misplaced in this context) by circulation activities. In short, the type of use-value that results from a certain activity is immaterial to the distinction PUPL only when the activity in question is part of production.

The second major area of contention concerns those categories of labour which in different ways contribute to an increase in the mass and rate of surplus-value. Since productive labour for capital is defined as that labour which produces surplus-value, it has been held that any type of worker (health workers, teachers, scientists and research workers to name but a few examples) who contributes to an increase in surplus-value for capital should be considered either as productive, on an equal footing with, for instance, industrial workers, or as indirectly productive. There is a certain deceptive coherence here. However, closer examination shows that it would be a serious mistake to go along with this line of reasoning.

Surplus-value is nothing but the historically specific social form of the surplus labour the worker expends over and above necessary labour, i.e. over and above that part of his labour that goes to reproduce the equivalent to his wages. Those workers who are supposed to be indirectly productive do not expend surplus labour for capital (unless they themselves are wage workers for a certain capitalist in which case they are already productive labourers). To say, therefore, that they contribute to the production of extra surplus-value for capital becomes something of a riddle: where does the extra surplus-value come from if capital does not appropriate a part of their labour as surplus labour? The paradox disappears as soon as we realize that the workers in question contribute to the rise in surplus-value through their contribution to an increase in the social productivity of labour. This increase in productivity finds its immediate expression in a diminution of the value of the labour power of production workers and, thereby, in an increase in the surplus labour and hence of surplus-value appropriated by capital. So the extra surplus-value is the value form of the extra surplus labour extorted out of the production workers. It has already been counted as their contribution to surplus-value so that to count it a second time as a specific contribution of scientists, educators etc. would be a blatant case of doublecounting. And since the distinction PUPL has no normative or moralistic significance but is important in analysing the trajectory of accumulation, this type of double-counting would have been unacceptable. This whole debate was foreshadowed by Marx in a passage we have already quoted above where he specified the specific importance of the direct participation of the worker in the productive process:.

...labour is only productive, and an exponent of labour-power is only a productive worker, if it or he creates surplus-value directly, i.e. the only productive labour is that which is directly consumed in the course of production for the valorization of capital. (Marx, 1976a: 1038a; emphasis added).

CONCLUSIONWe believe that the above discussion shows that a systematic and rigorous theoretical reconstruction of the distinction PUPL renders it immune to the major criticisms to which it has been subjected in the course of the debate of the last two decades. It is, in our opinion, the main contribution of this article to have provided this type of systematic attempt at reconstruction, with particular emphasis on the key importance, neglected heretofore by most participants in the debate, of the category 'productive labour in general as the foundation upon which the more specific and narrow concept 'productive labour for capital rises.

The above discussion also provides us with the criteria necessary to classify as productive or unproductive the major groups of workers within a capitalist economy. This classification may be summed up as in Figure 3.

The first point to be noted is that Figure 3 is an attempt to classify only those workers (or producers) who are employed on the basis of a wage contract and excludes both labour expanded outside the circuits of money-commodity circulation (housewives, self-sufficient peasant households) and that of petty-commodity producers. This is because the division productive/unproductive is meaningful only when a worker is a wage labourer, since the aim of the distinction is to see what portion of the total social labour so employed creates (is productive of) surplus-value and what portion is simply engaged in activities that are paid out of surplus-value. The production of surplus-value being predicated upon wage-labour, the two categories of use-value production and petty-commodity production are irrelevant to the distinction under examination. That is why they were excluded from the discussion at an earlier stage (see Figure 2).

Figure 3 brings out the fact that major sections of the working class in capitalist society are unproductive workers. Naturally, this does not imply in any sense that they are less important either for the well-being of society or for the class struggle and for revolutionary strategy. The distinction PUPL is significant exclusively for the analysis of the various significant variables of the capitalist economy such as the value of labour-power, the rate and mass of surplus-value and hence the rate of capital accumulation. This is sufficient ground for taking the distinction seriously for it is only through an examination of these variables that Marxists can adequately analyze the historical trajectory and the cyclical fluctuations of capitalist accumulation.

Added material.

Figure 1.

Figure 2.

Figure 3.

FOOTNOTES1. For empirical estimations of Marxian categories see, for example, among many recent works Shaikh and Tonak (1994), Moseley (1986).

2. Since we base our analysis on the labour theory of value, the question of whether the distinction PUPL can be meaningful in alternative theoretical frameworks lies outside the scope of this article.

3. The wages of unproductive labourers 'constitute a reduction of the surplus-value that is available for reinvestment as capital (Yaffe, 1973a: 2).

4. For diametrically opposite approaches to this question see Glyn and Sutcliffe (1972); Yaffe (1973b); Shaikh (1978a); Shaikh (1978b).

5. In this context, an alternative macroeconomic accounting framework based on the PUPL distinction expectedly produces a completely different set of measures of surplus product, consumption, investment and productivity. Hence, one should expect to arrive at a fundamentally different empirical picture of a given economy. To illustrate the dramatic extent of this difference in empirical measurements, it suffices to compare the rate of surplus-value (by taking into consideration the PUPL distinction) with its conventional counterpart, the profit-wage ratio. Between 1948-89, the U.S. rate of surplus-value increased by approximately 44% when the conventional measure profit wage ratio declined by 22% (Shaikh and Tonak, 1994: 151).

6. For a discussion of questions relating to redistribution due to state action see Tonak (1984).

7. On the other hand, the distinction PUPL is not necessary for establishing (or even reinforcing) the distinction between spheres of production and circulation. This latter distinction is conceptually independent of and prior to the distinction PUPL which itself is, in fact, predicated upon a solid understanding of the difference between production and circulation. We emphasize this point because there seems to be some confusion regarding this aspect, a point we will be returning to in the penultimate section.

8. Need it be pointed out that the concept of 'productive labour in general employed here by Marx is an element of the conceptual sphere of 'production-in-general which defines characteristics common to all modes of production (the labour process, for instance) which Marx takes as his starting point in his analysis of the capitalist mode of production (Marx, 1973: 85-8).

9. The idea of grounding the distinction between PUPL on such a basic set of activities was first put forward by Shaikh (1980). The first four concepts employed here are due to Marx's discussion in the Grundrisse, (Marx, 1973: 88-100). Consumption, however, is here taken up only in its 'personal and social form ('consumption proper in Marx's words), since other types of consumption, i.e. consumption of means of production and of raw materials (what Marx calls 'productive consumption ) refer immediately back to production.

10. It must be clearly stated that housework, predominantly carried out by women, is not a consumption activity but a necessary element of the production of the material to be consumed. Without housework consumption goods cannot acquire the final form in which they can be consumed (including the socially created forms of consumption). Therefore housework is a production activity (see further footnote 17).

11. On the other hand, to the extent that the capitalist also works as a manager of his own firm, he acts, in that capacity, as worker and that part of his income due to his labour should be regarded as belonging to the same category as the salaries of the rest of the managerial staff. Marx is of the opinion that in such cases the capitalist should even be considered to be a productive worker: 'As the director of the labour process the capitalist performs productive labour in the same sense that his labour is involved in the total process that is realized in the product (Marx, 1976a, 1048). This whole discussion clearly shows that in the distinction PUPL the decisive criterion concerns the type of activity carried on and not the position of the individual who carries out this activity within the overall social hierarchy or division of labour.

12. See Marx: 'The structure of distribution is completely determined by the structure of production. Distribution is itself a product of production, not only in its object, in that only the results of production can be distributed, but also in its form... (Marx, 1973: 95).

13. Others, such as kings or priests, to give the most prominent examples, can hardly be said to 'work in any meaningful sense of the word.

14. And since capital is but a converted form of surplus-value, 'productive labour is only that which produces capital...Labour becomes productive only by producing its own opposite (Marx, 1973: 305n).

15. See Marx: '...these definitions are therefore...derived...from the definite social form, the social relations of production, within which the labour is realized (Marx, 1963: 157).

16. This double determination of the concept productive labour for capital is expressed by Marx in the following manner: 'The concept of a productive worker therefore implies not merely a relation between the activity of work and its useful effect, between the worker and the product of his work, but also a specifically social relation of production, a relation with a historical origin which stamps the worker as capital's direct means of valorization (Marx, 1973: 644).

17. The fact that housework, which swallows up a large part of the life of women, is not productive for capital does not, of course, mean that it is not productive in the general sense or that it is not useful for society at large. It certainly is an essential element within the overall expenditure of total social labour under all societies and should be recognized as such. However, it does not directly and immediately create surplus-value and this is what counts when deciding what is productive labour under capitalism.

18. It is estimated that, before the first world war, hired domestic service accounted for as much as one-sixth of Britain's workforce (The Economist, 1998, 20).

19. See Marx: '...circulation of commodities can take place without physical motion by them, and there can be transportation of products without circulation of commodities, and even without a direct exchange of products. A house sold by A to B does not wander from one place to another, although it circulates as a commodity (Marx, 1956: 152).

20. Since labour spent in storage activities in this sense is part of the production process, just in-time techniques which result in a reduction of the expenditure of such labour simply raise the productivity of labour in the production process and thereby bring down the relative value of the commodities produced through such processes.

21. This is what Marx has to say on the question: 'Quantities of products are not increased by transportation. Nor, with a few exceptions, is the possible alteration of their natural qualities, brought about by their transportation, an intentional useful effect; it is rather an unavoidable evil. But the use-value of things is materialized only in their consumption and may necessitate a change of location of these things, hence may require an additional process of production, in the transport industry. The productive capital invested in this industry imparts value to the transported products, partly by transferring value from the means of transportation, partly by adding value through labour performed in transport. This last-named increment of value consists, as it does in all capitalist production, of a replacement of wages and of surplus-value (Marx, 1956: 153).

22. This may partly be due to the fact that the national income accounts of the Soviet Union and, following its example, of other countries where capitalism has been abolished were so organized as to relegate the production of all services to the domain of unproductive activities. Presumably, some authors assume on the strength of this example that the provision of services under capitalism also falls in that domain. (Whether criteria for the distinction PUPL applicable to capitalism are equally valid in a post-capitalist society is a question we cannot discuss here.).

23. Marx says of this definition 'Adam Smith was essentially correct with his productive and unproductive labour, correct from the standpoint of bourgeois economy (Marx, 1973: 273).

24. The foregoing quotation should, however, be handled carefully. Lest it might be read as saying that all labour can be productive irrespective of its position within the social division of labour and, thereby, accepted as licence to claim that workers employed by circulation capital or even those engaged in the reproduction of the social order are productive, we feel it necessary to draw attention to Marx's careful wording (of what is after all an unpublished text): He talks of 'special utility and 'particular use-value implying thereby that the product in question does have a use-value and therefore the labour which produces it is engaged in production. In other words, it is not the presence or otherwise of a use-value, but its particular nature that Marx regards as irrelevant to the discussion of PUPL.

25. The translation is here corrected. The word 'which italicized in the above quotation was omitted by the translator and this changes the meaning of Poulantzas' discussion. Its absence makes it seem that it is 'use-values which are exchanged against revenue while the French text shows that it is labour, or rather labour-power, that is so exchanged.

26. Having equated the working class with the category of productive workers, Poulantzas goes even further to declare that the wage workers employed by capital in the service sector do not belong to the working class (Poulantzas, 1975: 212-14).

27. See Marx: 'The materialization, etc., of labour is however not to be taken in such a Scottish sense as Adam Smith conceives it. When we speak of the commodity as a materialization of labour...this itself is only an imaginary, that is to say, a purely social mode of existence of the commodity which has nothing to do with its corporeal reality... (Marx, 1963: 171).

28. This criterion can also be expressed by saying that in the case of services 'the product is not separable from the act of producing. (Marx, 1973: 1048).

29. Some prominent examples of the critical approach to PUPL are the following: Gough (1972), Harrison (1973), Hunt (1979), de Vroey (1982) and Laibman (1992). It should be noted that the line of critique which we attempt to respond to in the text is a composite picture and there is not the least implication that the authors mentioned share each and every idea presented below.

30. Lest it be retorted that circulation does contribute to increased production by accelerating the turnover of capital, it should be recalled that we have already taken this aspect into consideration.

REFERENCESde Vroey, Michel (1982) 'On the Obsolescence of the Marxian Theory of Value. in Capital & Class 17, Summer.

Glyn, A. and B. Sutcliffe (1972) British Capitalism, Workers, and the Profit Squeeze. Penguin Press, Middlesex.

Gough, Ian (1972) 'Marx's Theory of Productive and Unproductive Labour. in New Left Review 76, December.

Harrison, John (1973) 'Productive and Unproductive Labour in Marx's Political Economy. in Bulletin of the Conference of Socialist Economists, Autumn.

Hunt, E.K. (1979) 'The Categories of Productive and Unproductive Labour in Marxist Economic Theory. in Science and Society, Fall.

Laibman, David (1992) Value, Technical Change, and Crisis: Explorations in Marxist Economic Theory. M. E. Sharpe, New York.

Mandel, Ernest (1981) 'Introduction Capital, Vol.II, Vintage Books, New York.

Marx, Karl (1956) Capital, Vol.I. Progress Publishers, Moscow.

Marx, Karl (1963) Theories of Surplus-Value, Part I. Progress Publishers, Moscow.

Marx, Karl (1973) Grundrisse. Penguin, Harmondsworth.

Marx, Karl (1976a) 'Results of the Immediate Process of Production. published as an Appendix in Capital, Vol.I.Penguin, Harmondsworth.

Marx, Karl (1976b) Capital, Vol.I. Penguin, Harmondsworth.

Moseley, Fred (1986) 'Estimates of the Rate of Surplus-Value in the Postwar United States Economy, in Review of Radical Political Economics, Vol.18, No:1/2.

Poulantzas, Nicos (1975) Classes in Contemporary Capitalism. New Left Books, London.

Shaikh, Anwar (1978a) 'An Introduction to the History of Crisis Theories. in U.S. Capitalism in Crisis, Union for Radical Political Economics, Monthly Review Press, New York.

Shaikh, Anwar (1978b) National Income Accounts and Marxian Categories, New School for Social Research: Unpublished Paper.

Shaikh, Anwar (1980) Towards a Critique of the Keynesian Theory of the Role of the State, Unpublished Paper.

Shaikh, Anwar and E. Ahmet Tonak (1994) Measuring the Wealth of Nations: Political Economy of National Accounts. Cambridge University Press, New York.

Smith, Adam (1976) The Wealth of Nations. University of Chicago Press, Chicago.

The Economist (26/09/98).

Tonak, E. Ahmet (1984) A Conceptualization of State Revenues and Expenditures: U.S. 1952-1980, Unpublished Ph. D. Dissertation, New School for Social Research.

Yaffe, David (1973a) 'The Marxian Theory of Crisis, Capital, and the State, in Economy and Society.