Reading Capital. Louis Althusser, Étienne Balibar 1968

Part III: The Basic Concepts of Historical Materialism
(Étienne Balibar)

First published: by Librairie François Maspero, Paris, 1968;
Translated: by Ben Brewster;
This translation first published New Left Books 1970.

The preceding papers have already formulated the idea that Marx’s work contains a general scientific theory of history. In particular, they have shown that, in the formulation of this theory, Marx’s construction of the central concept of the ‘mode of production’ has the function of an epistemological break with respect to the whole tradition of the philosophy of history. For in its generality it is absolutely incompatible with the principles of idealism, whether dogmatic or empiricist, and it progressively revolutionizes the whole problematic of society and history.

If this is the case, we know that it is because Marx’s ‘historical materialism’ gives us not only elements of scientific historical knowledge (e.g., elements restricted to the history of ‘bourgeois’ society, in its economic and political aspects), but, in principle, a true theoretical science, and therefore an abstract science. The concept of the ‘mode of production’ and the concepts immediately related to it thus appear as the first abstract concepts whose validity is not as such limited to a given period or type of society, but on which, on the contrary, the concrete knowledge of this period and type depends. Hence the importance of defining them at the level of generality that they demand, i.e., in fact, the importance of posing a number of problems which the science of history has been waiting for since Marx.

Althusser however, in his paper, has shown us that the explicit formulation (and therefore recognition) of an abstract theory of history is surrounded by difficulties and ambiguities. He has shown the historical and philosophical reasons for this. Marx’s theory was able to realize the paradox of having as its constant object the very history whose scientific knowledge it inaugurated, and yet of offering nowhere the adequate concept of this history, reflected for itself. I should like first to add a few specifications of this point, which will serve as a direct introduction to my particular problem.

It is not quite accurate to say that this theoretical formulation is missing: several texts give a remarkable outline of it, e.g., the first section of The German Ideology (which already contained a whole new definition of ‘production’), the various preparatory drafts for Capital collected into the Grundrisse der Kritik der politishen Ökonomie,[1] and above all the Preface to A Contribution to the Critique of Political Economy, the terms of which have been constantly discussed in the Marxist tradition. These are very general, prospective or summary texts; texts in which the sharpness of the distinctions and the peremptoriness of the claims are only equalled by the brevity of the justifications, the elliptical nature of the definitions. By an unfortunate accident, which is really a true historical necessity, the only expositions of the principles of the theory of history and the main expositions of its method (the 1857 Introduction) are of this type, and most of them were also intentionally left as incomplete and unpublished manuscripts. So despite the malicious critical intentions that inspire those readers of Marx who have asked ‘Where precisely did Marx set out his conception of history?’, they have not been completely unfair.

The reader will be familiar with the young Lenin’s answer in What the Friends of the People Really Are:[2] this theory is everywhere, but in two forms; the Preface to A Contribution presents ‘the hypothesis of historical materialism’; Capital sets this hypothesis to work and verifies it against the example of the capitalist social formation. These concepts enable Lenin to formulate what is for us a decisive commentary: in the expression ‘historical materialism’, ‘materialism’ means no more than science, and the expression is strictly synonymous with that of ‘science of history’. But at the same time, these concepts belong organically to the empiricist, even pragmatist theory of science, and this text of Lenin’s is throughout an application of such a theory (hypothesis/verification). However, let us reconsider its movement in other terms.

In reality, this Preface to A Contribution, if it is read attentively, does not present us with the form of a hypothesis, but explicitly that of an answer, an answer to a question we must try to reconstitute.

As an example, let us take a familiar text, one of those programme-texts whose interest I have just discussed, in which Marx states what was new in what he had proved: his letter to Weydemeyer on 5 March 1852:

No credit is due to me for discovering the existence of classes in modern society, nor yet the struggle between them. Long before me bourgeois historians had described the historical development of this struggle of the classes, and bourgeois economists the economic anatomy of the classes. What I did that was new was to prove: (1) that the existence of classes is only bound up with particular historical phases in the development of production... .

Here we find a procedure characteristic of Marx when he wants to think his own ‘novelty’, i.e., his rupture, his scientificity: the delimitation of a classicism. Just as there is an economic classicism (in England), there is a historical classicism, represented by the French and German historians of the early nineteenth century (Thierry, Guizot and Niebuhr). This, therefore, is Marx’s point of departure: their point of arrival. Historical knowledge in its most advanced form shows the succession of ‘civilizations’, ‘political regimes’, ‘events’, ‘cultures’, organized and rationalized by a series of class struggles, a general form whose patterns can be listed: slaves and free citizens, patricians and plebeans, serfs and feudal lords, masters and journeymen, land-owners and bourgeois, bourgeois and proletarians, etc. This heritage, this fact, proposed by history, but itself already the result of a labour of knowledge, is reflected in the famous opening of the Manifesto: ‘The history of all hitherto existing society is the history of class struggles.’ This sentence is not the first statement of Marx’s theory, it predates it, it summarizes the raw material of its work of transformation.

This is a very important point, for it enables us to formulate Marx’s question more precisely, the question contained in the Preface to A Contribution: on what conditions can the claim that history is the history of class struggles be a scientifc utterance? In other words what classes are these? what are classes? what is their struggle?

If we turn to the text of the Preface itself, we do indeed find an exposition of a relationship between the ‘social formation’ (Gesellschaftsformation) and its ‘economic base’ or ‘economic structure’ (Struktur), the anatomy of which is constituted by the study of the mode of production. The social formation is the site of a first ‘contradiction’ between the classes which Marx describes in terms of struggle, war, and opposition, a ‘contradiction’ which can be ‘now hidden, now open’, and whose terms are ‘in a word, oppressor and oppressed’ (The Communist Manifesto). Here it is related just as to its essence to a second form of ‘contradiction’ which Marx is always very careful not to confuse with the first, even terminologically: he calls it an ‘antagonism’, ‘not in the sense of individual antagonism’ (nicht im individuellen Sinn), i.e., not a struggle between men but an antagonistic structure; it is inside the economic base, typical of a determinate mode of production, and its terms are called ‘the level of the productive forces’ and ‘the relations of production’. The antagonism between the productive forces and the relations of production has the effect of a revolutionary rupture, and it is this effect which determines the transition from one mode of production to another (‘progressive epochs in the economic formation of society’), and thereby the transformation of the whole social formation. Marx himself chose to restrict his study to the level of the relatively autonomous sphere or stage of this ‘antagonism’ inside the economic structure.

But it remains strictly impossible for us to locate this sphere, since the terms that define it do not yet have any meaning. Indeed, it would be absolutely wrong to take the descriptive style of some of these terms or the direct simplicity with which Marx presents them as a pretext for believing them to be given in immediate experience and of obvious significance. On the contrary, they have been produced by Marx (who is careful to remind us – notably in his use of the term ‘civil society’ – that a considerable part of the raw material of this production had been constituted by economic and philosophical tradition), and they are so little obvious that it is extremely difficult to make use of them in actual sociological analyses without first mastering the definitions that Marx gave of them elsewhere. That is why they are often described from the standpoint of bourgeois empiricist sociology as paradoxical, heteroclite or inconsistent, or else assimilated without further ado to other terms: technology, economics, institutions, human relations, etc.

Taking this textual reading further, we can draw from it the two principles on which is based the transformation of history into a science: the principle of periodization and the principle of the articulation of the different practices in the social structure. One diachronic principle, it seems, and one synchronic principle. The principle of the articulation of the practices refers to the construction (Bau) or mechanism of ‘correspondence’ in which the social formation is presented as constituted out of different levels (we shall also speak of them as instances and practices). Marx lists three: the economic base, the legal and political superstructures, and the forms of social consciousness. As for periodization, it distributes history according to the epochs of its economic structure. These two principles introduce a double reduction of temporal continuity. Leaving aside the problem of primitive societies (i.e., the way Marx conceived the origin of society: there is no allusion to this here, any more than there is in the Manifesto), there is, first, a reduction to an absolute invariance in the elements which are found in every social structure (an economic base, legal and political forms, and ideological forms); second, there is a division into periods which replaces historical continuity with a discontinuity, a succession of temporarily invariant states of the structure which change by rapid mutation (‘revolution’): the antagonism that induces the mutation can only be defined by this invariance itself, i.e., by the permanence of the terms which it opposes.

These states of the structure are the modes of production, and the history of society can be reduced to a discontinuous succession of modes of production.

Now it is essential to pose the question of the theoretical status of these concepts. Are they all positive concepts? Does the text as a whole have a homogeneous content of theoretical knowledge, at the level of scientific abstraction which I have just discussed, as Gramsci thinks, for example, regarding it as he does as the most exact exposition of the ‘philosophy of praxis’?

I think, on the contrary, that within theoretical practice itself, this text has the status of what is called a set of practical concepts.[3] In other words, this text offers us concepts which still depend in their formulation precisely on the problematic which has to be displaced; at the same time, without being able to think it in its concept they indicate where we must go in order to pose otherwise (and at the same stroke solve) a new problem which has arisen within the old problematic.

To demonstrate this characteristic, I shall take as my main example the concept of periodization. This concept belongs completely to the traditional conception of history which Marx is questioning here. It is the concept of discontinuity in continuity, the concept which fragments the line of time, thereby finding the possibility of understanding historical phenomena in the framework of an autonomous totality (in this general form, the problem does not change whether we look for ‘civilizations’ or for ‘structures’ as opposed to ‘conjunctures’). Thus the concept of periodization gives theoretical form to a problem which historians have never been able to evade in their practice, but without itself providing them with a theoretical solution, a precise theoretical methodology, for fundamental reasons which the rest of this paper will reveal. A problem which manifestly haunts these texts of Marx’s, too: the problem of the ‘right break’. If the right break or breaks are found, history, without ceasing to unfold in the linear flux of time, becomes intelligible as the relationship between an essential permanence and a subordinate movement. The questions necessarily contained in this problematic do not differ in their essence whether it is economic structures or ages (the ‘age of Louis XIV’) that have to be distinguished. The latter formulation even has the advantage that it constantly reminds us that these problems are constrained to respect the conditions imposed on them by the linearity of time: or in other words to transpose all discontinuities onto the plane of temporal discontinuities. It is in this way that it has been possible for the main instrument of historical conceptualization which emerged in modern economic history to have been a distinction between the long term and the short term, i.e., a distinction entirely ‘rotated’ into the linearity of time. The historian seeks to distinguish the long-term phenomena from the short-term phenomena, and to show how the latter are inserted into the movement of the former and into their determinism. At the same time, he perpetuates two kinds of difficulties: those relating to the notion of the historical event, which is assessed according to the single criteria of brevity (suddenness) and is therefore almost of necessity confined to the sphere of political events; and those relating to the impossibility of making clean breaks.

Marx therefore seems to treat matters in exactly the same way; simply proposing a new criterion of periodization, a means of making the right break, the one which gives the best periods, the periods which must not be described as artificial but not arbitrary, but which correspond to the very nature of historical social reality.[4] In fact, if we are to take the idea of an epistemological rupture seriously, we should have to say that the very nature of the criterion chosen (epochs in the economic structure) implies a complete transformation of the way the problem has to be posed. Marx would say: in order to periodize the history of mankind, we must approach it from the side of economic science rather than from that of art, politics, science or law. But it is then clear that what is theoretically essential in this concept, what is new in its contribution, what defines it differentially, cannot lie in the general form that it has in common with all the other periodizations, but in its particular answer to the question.

We must now think in all its epistemological singularity the form in which Marx proposes his own theory to us here: the theoretical specificity of Marx’s own concept of periodization lies solely in the fact that it is a particular answer to a question which, for its part, belongs to an old problematic, a question which is not decisive in the constitution of the science. Such a situation necessarily implies and envelops Marx’s own inability to justify his particular answer at this level – in fact it is impossible to justify it at this level – and that is perhaps why the text we are discussing is so dogmatically brief; and also Marx’s inability to formulate the true theoretical concept of this periodization, since it would be the concept of the only way to periodize which abolishes the earlier problematic of periodization based on the linear conception of time and at grips with it.

What is true of the concept of periodization is also necessarily true of the concepts in the Preface which designate the different instances of the social structure other than the economic base (which, as we have seen, is designated by new concepts which are specific if not yet defined: productive forces, relations of production, mode of production). These concepts and all the terms which designate the peculiar articulation of their objects (‘corresponds ‘, ‘on which rises’, etc.) are remarkably vague and yet they have sustained all Marxist reflection on the problem of ideologies and superstructures. They have no other function than to indicate where, provisionally, Marx is not going to go on this occasion; they do not therefore constitute a knowledge of these levels and their mutual relations, but merely a practical registration (practical in the sense of theoretical practice, of course) which disengages the level of the economic structure which Marx is now undertaking to study, in its relative autonomy. Nevertheless, if this registration is to be possible, certain theoretical conditions must be met which constitute its real meaning: on condition that its concept is redefined, the economic structure must really possess the relative autonomy which allows us to delimit it as an independent field of research. A plurality of instances must be an essential property of every social structure (but we shall regard their number, names and the terms which designate their articulation as subject to revision); the problem of the science of society must be precisely the problem of the forms of variation of their articulation.[5]

Finally, these same comments are valid for the concept ‘men ‘: the ‘men’ who support the whole process. Let me say without prevarication that all the rest of this paper is governed by a principle of critical reading, which I hope will be granted me: I shall refrain from pre-judging the meaning of such a term (‘men’) until I have elucidated its conceptual function in the theoretical structure which contains it – since its theoretical meaning depends entirely on this function. The ‘obviousness’, the ‘transparency’ of the word ‘men’ (here charged with every carnal opacity) and its anodyne appearance are the most dangerous of the traps I am trying to avoid. I shall not be satisfied until I have either situated it and founded it in the necessity of the theoretical system to which it belongs, or eliminated it as a foreign body, and in this latter case, replaced it by something else. The formulations in this Preface (‘In the social production of their life, men enter into definite relations ... their material productive forces ... It is not the consciousness of men that determines their being ... ideological forms in which men become conscious of this conflict ...’) must be compared with many others in The German Ideology, in The Poverty of Philosophy, in the correspondence (notably in Engels’s letter to Bloch: ‘We (=men) make our history ourselves, but, in the first place, under very definite assumptions and conditions ...’). All these formulations are the matrices of the idea that it is men who make history on the basis of previous conditions. But who are these ‘men’? A first, ‘naïve’ reading of our Preface suggests that they are firstly the agents of the process of the historical transformation of the social structure via the mediation of the activity of economic production. We are to understand that men produce their material means of survival, and at the same time, the social relations in which they produce, which are either maintained or transformed. In consequence, they are secondly the real (concrete) supports of the different practices articulated into the social structure: this articulation is precisely given only by the men who at one and the same time take part in the production process, are legal subjects and are consciousnesses. The importance of this concept can thus be measured by the function of structural cohesion it fulfils in theory. But its ambiguity is revealed in the fact that it belongs simultaneously to several incompatible systems of concepts: theoretical and non-theoretical, scientific and ideological. The concept of ‘men’ thus constitutes a real point where the utterance slips away towards the regions of philosophical or commonplace ideology. The task of epistemology here is to stop the utterance slipping away by fixing the meaning of the concept.

If this really is the ambiguous status of these concepts, if they really are practical concepts, signal concepts within a still unbalanced problematic (periodization, correspondence – articulation of the practices, men), then this task becomes necessary. I propose to begin this work here, an explicit labour which transforms these ‘practical’ concepts into theoretical concepts of the Marxist theory of history, a labour which strips them of their present theoretical form in order to make them theoretically adequate to their practical content. At the same time, those concepts, which are no more than expressions of the exigencies of the old ideological problematic, will disappear completely. And at the same time, too, weak and open points will appear which will demand the production of new theoretical concepts even in the region explored by Marx, and make this production possible. For, at the most abstract level, the fruitful incompleteness of Marx’s work is the necessary effect of its scientific character.

Since the theoretical concepts of the Preface to A Contribution have this compound status as the anticipations and summaries (or ‘results’) of an analysis, the text of Capital cannot therefore constitute a mere ‘verification’ or application of them. The text of Capital, in its necessary order of exposition, is the process of the production, construction and definition of these theoretical concepts, or at least of some of them. If we take the ‘mode of production’ as the main object of our analysis, it is because in that very exposition Marx himself designates the theoretical object of Capital as the concept of the capitalist mode of production.

Chapter 1. From Periodization to the Modes of Production

In my reconstitution of the concept of a mode of production, I shall start with what seem the most external and formal determinations, and attempt to enrich them progressively. I shall therefore return to the first question of the theory of history, the question of the breaks, of the right break. Scattered throughout Marx’s writings is a series of comments with a common form: they all begin as follows: ‘What defines a historical epoch of production is ...’ or again, ‘what defines a historical mode of production is the specific way in which it ...’; then follow several phrases whose comparison is only too likely to be quite instructive, for they are all equivalent in principle, without this equivalence being at all tautological. In other words, we can try to extract from these equivalent answers to a single question which depends in principle on a method of comparison, the determination of the criteria for the identification of a ‘mode of production’ (for the moment this term is still no more than a name, as far as we are concerned, the name of the unit of periodization peculiar to Marx), the determination of the pertinent differences which make it possible to define the concept of each mode of production. If we do reveal such pertinent differences, we shall face a second task, that of characterizing the ensembles within which these differences act.[6]


Even more than its French or English equivalent, the German term Produktionsweise retains some echo of the simple and original meaning of the word Weise, mode, i.e., manner, way to do something (there is a standard German expression for this, the doublet Art und Weise). This warns us immediately what kind of analysis we are dealing with: a descriptive analysis which isolates forms or qualities. Thus the mode ‘of production’ first exists on the same plane as the many other modes we find in the course of an analysis of Capital. For example:

Modes of exchange: ‘It is not the economy, i.e., the process of production itself that is emphasized as the distinguishing mark of the two categories, money-economy and credit-economy, but rather the mode of exchange ... between the various agents of production or producers’ (Verkehrsweise) (Capital, Vol. II, p. 116). Modes of circulation: ‘What determines that a portion of the capital-value invested in means of production is endowed with the character of fixed capital is exclusively the peculiar manner in which this value circulates. This specific manner of circulation (diese eigene Weise der Zirkulation) arises from the specific manner in which the instrument of labour transmits is value to the product, or in which it behaves (sich ...verhält) as a creator of value during the process of production. This manner again arises from the special way in which the instruments of labour function in the labour-process (aus der besonderen Art der Funktion der Arbeitsmittel)’ (Capital, Vol. II, p. 160). Modes of consumption: ‘Even the number of so-called natural needs, as also the modes of satisfying them (die Art ihrer Befriedigung), are themselves a historical product’ (Capital, T.I, p. 174; Vol. I, p. 171).

I could give other examples, too, taken from the ‘economic’ sphere and elsewhere.

This descriptive and comparative character indicates that the expression ‘mode of production’ does not initially contain any reference to the breadth of its application other than in the form of a tendency towards generality: we find the capitalist mode of production, in the narrow sense of the industrial mode of production, the utilization of machinery, steadily extended to the various branches of industry:

But when surplus-value has to be produced by the conversion of necessary labour into surplus-labour, it no longer suffices for capita!, while leaving intact the traditional labour process, simply to prolong the duration of that process. The technical and social conditions of the process, and consequently the very mode of production must be transformed. Only then can the productivity of labour be increased, thus decreasing the value of labour-power, and thereby shortening the time necessary for the reproduction of that value (Capital, T.II, p. 9; Vol. I, p. 315).

This text is preceded by the following definition:

a revolution in the conditions of production, i.e., an alteration in his tools or his mode of working, or in both.

Here we have descriptions of processes, manners, methods, forms – all expressions which have meaning only by what they exclude. Firstly, quantitative measurements. Thus the productivity of labour, which determines the relative quantities necessary for the satisfaction of the producer’s needs and for surplus-value, only intervenes here insofar as it depends in each historical epoch, on a certain form of the labour process, i.e., on the relationship between certain instruments (means of labour) and certain forms of labour organization (which include non-organizations, such as when the individual producer alone sets to work the tools which enable him to obtain an actual useful product). Then they exclude any consideration of the material nature of the objects which produce or undergo a transformation, insofar as such a consideration refers to the special features of branches of the social division of production which produce special use-values with peculiar technological characteristics. In this sense, Marx had already written in the 1857 Introduction that ‘political economy is not technology’ in the sense that the latter term had acquired at the beginning of the nineteenth century, and whose historical origins he reveals in the chapter in Volume One on Modern Industry. These two negative determinations are to be found in the text of the chapter on the labour process:

Relics of by-gone means of labour possess the same importance for the investigation of extinct economic forms of society, as does the structure of fossil bones for a knowledge of the organization of extinct species of animals. It is less what is produced than how it is produced (Nicht was ... sondern wie), and by what means of labour, that enables us to distinguish different economic epochs. Means of labour supply a standard of the degree of development of the labourer and they are indicators of the social relations in which he labours (Nicht nur Gradmesser der Entrwicklung der menschlichen Arbeitskraft, sondern auch Anzeiger der gesellschaftlichen Verhältnissen, worin gearbeitet wird) (Capital, T.I, p. 182; Vol. I, pp. 179-80).

If means of labour are to be ‘indicators’ of social relations, they must obviously be justifiable by a type of analysis different from the measurement of their effectivity or the technological description of their elements. Otherwise we should fall back into Proudhon’s error and take machines for social relations (cf. The Poverty of Philosophy, op. cit., p. 133).

We can define this analysis as a differential determination of forms, and define a ‘mode’ as a system of forms which represents one state of the variation of the set of elements which necessarily enter into the process considered. This definition, which I am about to put to the test, is true for all modes, and on each occasion it requires two things: a listing of the places (or functions) which feature in the process concerned, and a determination of the pertinent criteria which enable us to distinguish between the forms occupying these places. Thus, if we return to the above-mentioned example of the mode of circulation (Capital, Vol. II, p. 160), we find that this criterion consists of the fact that it transmits its value to the product either in toto or only in parts spread over several periods of production. At the same time, we can derive from it the concepts by which Marx designates existence as an element of the process: function, factor. But in order to list these places we must refer to another ‘mode’, the ‘mode of production’ itself; we are not dealing with a relatively autonomous process with its own consistency. It is different with the mode of production itself, and there we find that consistency.


In the case, therefore, of the mode of production (in the strict sense), we still have to identify these elements. Here we shall find it necessary to compare several of Marx’s texts which complement one another, and even to suggest interpretations of them whose well-foundedness will, I hope, emerge later in the paper.

We find a first extremely clear text in Capital Volume Two:

Whatever the social form of production, labourers and means of production always remain factors (Faktoren) of it. But in a state of separation from each other either of these factors can be such only potentially (der Möglichkeit nach). For production to go on at all they must combine (Verbindung). The specific manner in which this combination is accomplished distinguishes the different epochs of the structure of society one from another (Capital, Vol. II, p. 34 – modified).

Two of the elements we are seeking are indicated here:

(1) The labourer (labour power);

(2) The means of production.

The text goes on:

In the present case, the separation of the free worker from his means of production is the starting-point given, and we have seen how and under what conditions these two elements are united in the hands of the capitalist, namely, as the productive mode of existence of his capital.

Here we find straightaway a third element which, like the other two, also deserves to be called a ‘factor’:

(3) The non-worker, appropriating surplus-labour. Elsewhere, Marx describes him as the representative of the ‘class of large proprietors’ (Grossbesitzerklasse – Capital, T.II, p. 185; Vol. I, p. 511). This is the capitalist. Besides this, we find here an element of a different kind which we could call a connection (relation) between the preceding elements: it can take two exclusive values: separation (Trennung)/property.

If we compare the results of our analysis of this text with a series of other texts, particularly those contained in Marx’s unpublished draft Pre-Capitalist Economic Formations (op. cit.), and in the Chapter in Volume Three of Capital on the ‘Genesis of Capitalist Ground Rent’, we find the same elements and long descriptions of their combinations. The labourer is specified as the

direct producer ; the property relation is itself specified according to several complex forms, notable the duality of ‘possession’ (use, enjoyment) and ‘property’ (property strictly speaking).

But the essential interest of these texts is that they oblige us to introduce into the structure a second connection distinct from the first, a second relation between the ‘factors’ of the combination. This is a very important point, for it governs our whole understanding of the structure. We must therefore try to define the nature of this connection very dearly, starting from Marx’s texts themselves. This connection corresponds to what Marx designates by various terms such as the real material appropriation of the means of production by the producer in the labour process (Aneignung, Appropriation, wirkliche Aneignung), or simply as the appropriation of nature by man. Two points must be clearly established:

(1) this connection is distinct from the preceding one;

(2) this, too, really is a connection, a relation between the previously listed elements.

The relative looseness of Marx’s vocabulary on this point in the texts I have mentioned (particularly Pre-Capitalist Economic Formations) makes it difficult to prove the first point. Marx uses a whole series of practically equivalent terms (Aneignung, Appropriation ; Besitz, Benutzung, etc.) to describe all the connections between the producer and his means of production. This looseness depends in reality on the difficulty Marx felt in clearly thinking the distinction between the two connections, a difficulty I shall explain. Nevertheless, let us take the text of Volume One of Capital on absolute surplus-value and relative surplus-value (T.II, pp. 183ff.; Vol. I, pp. 508ff.): there we find two uses of the word Aneignung (appropriation) less than two pages apart but with obviously different meanings corresponding to the two connections I have been discussing:

in der individuellen Aneignung von Naturgegenständen zu seinem Lebenszwecken kontrolliert er sich selbst. Später wird er kontrolliert (In the individual appropriation of natural objects the labourer controls himself. Afterwards his labour is controlled by others);

‘die Aneignung dieser Mehrarbeit durch das Kapital’ (the appropriation of that surplus-labour by capital).

The second ‘Aneignung’ describes a property relation, the one we first met. It describes one of the presuppositions of capitalist production: capital is the owner of all the means of production and of labour, and therefore it is the owner of the entire product.

But the first does not designate a property relation: it belongs to the analysis of what Marx called the ‘labour process’, or rather it situates the analysis of that labour process as part of the analysis of the mode of production. Nowhere in it does the capitalist intervene as an owner, but only the labourer, the means of labour and the object of labour.

In the light of this distinction, we can now re-read for example the chapter on the labour process (T.I, pp. 186-7; Vol. I, pp. 184-5). Marx writes:

The labour process, turned into the process by which the capitalist consumes labour-power, exhibits two characteristic phenomena. First, the labourer works under the control of the capitalist ... Secondly, the product is the property of the capitalist and not that of the labourer, its immediate producer ... (T.I, p. 187; Vol. I, p. 185).

In these ‘two phenomena’ characteristic of the capitalist mode of production, we find precisely the two connections in the specific form they take in the capitalist mode of production.

From the point of view of property, the labour process is an operation between things which the capitalist has purchased. ‘The product of this process belongs, therefore, to him, just as much as does the wine which is the product of a process of fermentation completed in his cellar.’

In the capitalist mode of production, the labour process is such that individual labour does not set to work the society’s means of production, which are the only means of production able to function as such. Without the capitalist’s ‘control’, which is a technically indispensable moment of the labour process, labour does not possess the fitness (Zwetkmässigkeit) it requires if it is to be social labour, i.e., labour used by society and recognized by it. The fitness peculiar to the capitalist mode of production implies the cooperation and division of the functions of control and execution. It is a form of the second connection I have discussed, which can now be defined as the direct producer’s ability to set to work the means of social production. In the pages of Capital, Marx defines several forms of this connection: the autonomy (Selbständigkeit) of the direct producer, and the forms of mutual dependence of the producers (co-operation, etc.).

We can already see that recognition of this second connection in its conceptual independence, in its difference from the ‘property’ connection (A), is the key to several very important theses of Capital. Notably the double function of the capitalist as the exploiter of labour-power (‘property’) and as the organizer of production (‘real appropriation’); a double function expounded by Marx in the chapters on co-operation, manufacture and modern industry (Volume One). This double function is an index of what I shall call the double nature of the division of labour in production (the ‘technical’ division of labour and the ‘social’ division of labour); at the same time, it is an index of the interdependence or intersection of these two divisions, which itself reflects the fact that the two connections which I have distinguished both belong to a singleVerbindung ‘, to a single combination, i.e., to the structure of a single mode of production.

That is why the distinction between these two connections finally enables us to understand what constitutes the complexity of the combination, the complexity which characterizes the Marxist totality as opposed to the Hegelian totality. When the concept of structural complexity was introduced,[7] it was a question of the complexity of the social structure as a whole, insofar as several relatively autonomous levels were articulated in it. Now we find that production itself is a complex totality, i.e., that nowhere is there a simple totality, and we can give a precise meaning to this complexity: it consists of the fact that the elements of the totality are not linked together once, but twice, by two distinct connections. What Marx called a combination is not therefore a simple relationship between thefactors’ of any production, but the relationship between these two connections and their interdependence.

Finally, therefore, we can draw up a table of the elements of any mode of production, a table of the invariants in the analysis of forms:

(1) labourer;

(2) means of production;

(i) object of labour;

(ii) means of labour;

(3) non-labourer;

(A) property connection;

(B) real or material appropriation connection.

Marx’s difficulty in clearly distinguishing between the two connections in certain historically retrospective texts can be explained by the particular form these connections take in the capitalist mode of production. In the capitalist mode of production, both connections can indeed be characterized by a ‘separation ‘: the labourer is ‘separated’ from all the means of production, he is stripped of all property (save that of his labour-power); but at the same time, as a human individual, the labourer is ‘separated’ from any ability to set in motion the instruments of social labour by himself; he has lost his craft skill, which no longer corresponds to the means of labour; as Marx says, the labour is no longer ‘his property’. In the capitalist mode of production, strictly speaking, these two ‘separations’, these two distinctions overlap and coincide in the image of the opposition between the ‘free’ labourer and the means of production instituted as capital, to the extent that the labourer himself becomes an element of capital: that is why Marx constantly confounds them in a single concept, the concept of the separation of the labourer from his condition of labour. Now in all the historical inquiries which trace the history of the constitution of the elements of the capitalist mode of production back to earlier modes of production, Marx takes this concept as his guiding thread. This explains his difficulty, a difficulty which is patent in the semantic hesitations of Pre-Capitalist Economic Formations, in isolating the two connections; for the homology between the two connections, the overlap between their forms, which characterizes the capitalist structure, does not so characterize those earlier modes of production. Marx only finds it again in the hypothetical ‘natural community’ which inaugurates history: then the form of each of the two connections was, on the contrary, the union, the belonging together of the labourer and the means of production: on the one hand the almost biological collective property of the land, on the other the biological naturalness of the labour (the earth as ‘man’s laboratory’, indistinctly object and means of labour).

But the entire difficulty, and any looseness in Marx’s terminology, disappear once our analysis deals with the effects of this double articulation of the mode of production, i.e., with the double nature of the ‘immediate production process’ as a labour process and (in its capitalist form) as a process of self-expansion (Verwertung) of value (the distinction between these two constitutes the object of Volume One, Chapter VII).

By varying the combination of these elements according to the two connections which are part of the structure of every mode of production, we can therefore reconstitute the various modes of production, i.e., we can set out the ‘presuppositions’ for the theoretical knowledge of them, which are quite simply the concepts of the conditions of their historical existence. In this way, we can even to a certain extent generate modes of production which have never existed in an independent form, and which do not therefore strictly speaking form part of our ‘periodization’ – modes of production such as Marx called the ‘mode of commodity production’ (the reunion of individual small producers owning their own means of production and setting them to work without co-operation); or modes of production for which it is only possible to foresee the general conditions, such as the socialist mode of production. The final result would be a comparative table of the forms of different modes of production which all combine the same ‘factors’.

However, this is by no means a combinatory in the strict sense, i.e., a form of combination in which only the places of the factors and their relations change, but not their nature. Before we go on to prove this in a second section, we can nevertheless draw from what has already been established a number of conclusions as to the nature of the ‘determination in the last instance’ of the social structure by the form of the production process; which amounts to a justification of what I announced when I referred to the Preface to A Contribution: that the new principle of periodization proposed by Marx contained a complete transformation of the historian’s problematic.


By a double necessity, the capitalist mode of production is both the mode of production in which the economy is most easily recognized as the ‘motor’ of history, and the mode of production in which the essence of this ‘economy’ is unrecognized in principle (in what Marx calls ‘fetishism’). That is why the first explanations of the problem of the ‘determination in the last instance by the economy’ that we find in Marx are directly linked to the problem of fetishism. They occur in the texts in Capital on the ‘fetishism of commodities’ (T.I, pp. 88-90; Vol. I, pp. 76-8), on the ‘genesis of capitalist ground rent’ (Vol. III, pp. 763-93) and on the ‘trinity formula’ (Vol. III, pp. 794-811), where Marx replaces the false conception of this ‘economy’ as a relation between things by its true definition as a system of social relations. At the same time, he presents the idea that the capitalist mode of production is the only one in which exploitation (the extortion of surplus-value), i.e., the specific form of the social relation that binds classes together in production, is ‘mystified’, ‘fetishized’ into the form of a relation between the things themselves. This thesis follows directly from his proof where the commodity is concerned: the social relation which constitutes its reality, knowledge of which enables us to assess its fetishism, is precisely the commodity relation as a relation of production, i.e., the commodity relation as generalized by the capitalist mode of production. A social (‘human’) relation cannot therefore be found behind ‘things’ in general, but only behind the thing of this capitalist relation.[8]

At this point there is a refutation of an objection raised against the general thesis of the Preface to A Contribution, which introduces the general idea of determination in the last instance. We shall only find this refutation intelligible if we constantly think the ‘economy’ as the structure of relations that I have defined:

According to these objections: ‘my view ... that the mode of production of material life dominates the development of social, political and intellectual life generally ... is very true for our own times, in which material interests preponderate, but not for the middle ages, in which Catholicism, nor for Athens and Rome, where politics, reigned supreme. In the first place it strikes one as an odd thing for anyone to suppose that those well-worn phrases about the middle ages and the ancient world are unknown to anyone else. This much, however, is clear, that the middle ages could not live on Catholicism, nor the ancient world on politics. On the contrary, it is the economic conditions of the time that explain why here politics and there Catholicism played the chief part. It requires but a slight acquaintance with the history of the Roman Republic, for example, to be aware that its secret history is the history of its landed property. On the other hand, Don Quixote long ago paid the penalty for wrongly imagining that knight errantry was compatible with all economic forms of society’ (Capital, T.I, p. 93n; Vol. I, p. 81n).

We can therefore first make a specification that can be added to those that the preceding papers have proposed with respect to fetishism: Marx’s thesis does not mean that in modes of production other than capitalism the structure of the social relations is transparent to the agents. ‘Fetishism’ is not absent from them, but displaced (onto Catholicism, politics, etc.). In reality certain of Marx’s formulations leave no doubt on this point. For example, at the beginning of the text on Pre-Capitalist Economic Formations, Marx writes about the so-called ‘primitive’ community:

The earth is the great laboratory, the arsenal which provides both the means and the materials of labour, and also the seat, the basis of the community. Men relate to it naïvely as the property of the community, and of the community producing and reproducing itself in living labour. Only insofar as the individual is a limb or member of such a community, does he regard himself as an owner or possessor. Real appropriation by means of the process of labour takes place under these pre-conditions, which are not the product of labour but appear as its natural or divine pre-conditions (Grundrisse, p. 376; PCEF, p. 64).

In other words, the transparency which characterizes the relation between the direct producer and his product in non-commodity modes of production has as its counterpart this specific form of ‘naïvety’ in which the existence of a community, i.e., certain kinship relations and forms of political organization, can appear as ‘natural or divine’ and not as implied by the structure of a particular mode of production.

But this point, which Marx touches on only too briefly (for lack of historical material), is in principle quite clearly linked to the problem of determination in the last instance. Indeed, it emerges that the ‘mystification’ applies not to the economy (the mode of material production) as such, but precisely to that instance of the social structure which, according to the nature of the mode of production, is determined as occupying the place of determination, the place of the last instance.

We can now understand why analogous causes produce analogous effects here: in the event, it is possible to give this formulation a precise sense; that is to say, whenever the place of determination is occupied by a single instance, the relationship of the agents will reveal phenomena analogous to ‘fetishism’. Perhaps it is not an exaggeration to say that this is the sense of the following passage from Pre-Capitalist Economic Formations on the ‘Asiatic’ mode of production:

In most Asiatic fundamental forms ... the all-embracing unity (Einheit) which stands above all these small communities may appear as the higher or sole proprietor, the real communities only as heredity possessors. Since the unity is the real owner, and the real pre-condition of common ownership, it is perfectly possible for it to appear as a particular being

page 219

above the numerous real, particular communities. The individual is then in fact propertyless, or property ... appears to be mediated by means of a grant from the total unity – which is realized in the despot as the father of the many communities – to the individual through the mediation of the particular community. It therefore follows that the surplus-product (which, incidentally, is legally determined as a consequence of the real appropriation through labour) belongs of itself (von sich selbst) to this higher unity ... (Grundrisse, pp. 376-7; PCEF, pp. 69-70).

This ‘of itself’ must be taken in the strongest sense, noting that in other modes of production, e.g., the feudal mode of production, the surplus-product does not ‘of itself’ belong to the representatives of the ruling class. As we shall see, something further is explicitly required for the feudal mode of production: a political relationship, either in the ‘pure’ form of violence, or in the adapted and improved forms of law. In the ‘Asiatic’ mode of production and the capitalist mode of production, on the contrary, to modes of production as far apart chronologically, geographically, etc., as possible, and despite the fact that the agents who enter into the relationship are different in other respects (here capitalist and wage-labourer, there State and communities), the same direct determination by the functions of the process of production produces the same effects of fetishism: the product belongs ‘of itself’ to this higher ‘unity’ because it appears to be the work of that unity. This is what Marx writes a little further on in the same text:

The communal conditions of real appropriation through labour, such as irrigation systems (very important among the Asian peoples), means of communication, etc., will then appear as the work of the higher unity – the despotic government which is poised above the lesser communities.

This reasoning recurs in the chapter in Capital on co-operation, where Marx systematically compares the Asiatic forms of despotism with capitalist forms of ‘despotism’, i.e., the joining of the function of control or direction, indispensable to the performance of the labour process (the real appropriation of the object of labour), with the function of ownership of the means of production.

Because social labour power costs capital nothing, and because, on the other hand, the wage-labourer himself does not develop it before his labour belongs to capital, it appears as a power with which capital is endowed by nature – a productive force that is immanent in capital. The colossal effects of simple co-operation are to be seen in the gigantic structures of the ancient Asiatics, Egyptians, Etruscans, etc... . This power of Asiatic and Egyptian kings, Etruscan theocrats, etc., has in modern society been transferred to the capitalist, whether he be an isolated or a collective capitalist (Capital, T.II, p. 26; Vol. I, pp. 333-4).

It would therefore be possible and legitimate to look in Asiatic despotism for an analogy to the forms of appearance which mean that in the capitalist mode of production, ‘all faculties of labour are projected as faculties of capital, just as all forms of value commodity are projected as forms of money’ (Capital). We should then in fact be basing ourselves on the analogy of the relations between the two connections with the ‘combination’ in these two modes of production, i.e., on the analogy of the articulation of the double division of labour (see above).

But above all, these texts imply that all the levels of the social structure have the structure of a ‘mode’ in the sense in which I have analysed the mode of production strictly speaking. In other words, they are themselves presented in the form of specific complex combinations (Verbindungen). They therefore imply specific social relations, which are no more patterns of the inter-subjectivity of the agents than are the social relations of production, but depend on functions of the process concerned: in this sense, I shall be rigorous in speaking of political social relations or ideological social relations. In the analysis of each of these modes of combination, I shall appeal to criteria of pertinence specific to each occasion.

The problem which I wish to approach is therefore the following: how is the determinant instance in the social structure in a given epoch itself determined, i.e., how does a specific mode of combination of the elements constituting the structure of the mode of production determine the place of determination in the last instance in the social structure, i.e., how does a specific mode of production determine the relations between the various instances of the structure, i.e., ultimately, the articulation of that structure? (What Althusser has called the matrix role of the mode of production.)

In order to answer this question, at least in principle, I shall consider, not an ideal, but a reduced case: that of a social structure reduced to the articulation of two different instances, an ‘economic’ instance and a ‘political’ instance, which will enable me to follow closely certain passages where Marx compares, vis-à-vis ground rent, the feudal mode of production with the capitalist mode of production.

On the simplest form of feudal ground rent, labour rent (corvée), Marx writes:

It is ... evident that in all forms in which the direct labourer remains the ‘possessor’ of the means of production and labour conditions necessary for the production of his own means of subsistence, the property relationship must simultaneously appear as a direct relation of lordship and servitude (als unmittelbares Herrschafts- und Knechtschaftsverhältnis), so that the direct producer is not free; a lack of freedom (Unfreiheit) which may be reduced from serfdom with enforced labour to a mere tributary relationship. The direct producer, according to our assumption, is to be found here in possession of his own means of production, the necessary material labour conditions required for the realization of his labour and the production of his own means of subsistence. He conducts his agricultural activity and the rural home industries connected with it independently ...

Under such conditions the surplus-labour for the nominal owner of the land can only be extorted from them by other than economic pressure, whatever the form assumed may be ... Thus, personal conditions of personal dependence are requisite, a lack of personal freedom, no matter to what extent, and being tied to the soil as its accessory (Zubehör), bondage in the true sense of the word... .

The specific economic form, in which unpaid surplus-labour is pumped out of direct producers, determines the relationship of rulers and ruled, as it grows directly out of production itself and in turn, reacts upon it as a determining element. Upon this, however, is founded the entire formation of the economic community which grows up out of the production relations themselves, thereby simultaneously its specific political form. It is always the direct relationship of the owners of the conditions of production to the direct producers ... which reveals the innermost secret, the hidden basis of the entire social structure, and with it the political form of the relation of sovereignty and dependence (Souveränitäts- und Abhängigkeitsverhältnis), in short, the corresponding specific form of the State ...

So much is evident with respect to labour rent, the simplest and most primitive form of rent: Rent is here the primeval form of surplus-labour and coincides with it. But this identity of surplus-value with unpaid labour of others need not be analysed here, because it still exists in its visible, palpable form, since the labour of the direct producer for himself is still separated in space and time from his labour for the landlord, and the latter appears directly in the brutal form of enforced labour for a third person (Capital, Vol. III, pp. 771-2).

This text contains four major points (I shall take them in a different order):

– a new formulation of the principle of periodization: ‘what distinguishes one historical epoch from another’. Here it is the mode of dependence of the social structure with respect to the mode of production, i.e., the mode of articulation of the social structure, which Marx gives us as equivalent to the previous determinations, from the point of view of its concept;

– the specific difference in the relation between labour and surplus-labour implied by the difference between the social relations in the feudal mode of production and in the capitalist mode of production (property/possession of the means of production): in the latter case there is a coincidence ‘in space and time’, simultaneously of labour and surplus-labour, but not in the former;

– the non-coincidence of the two processes, the labour process and the surplus-labour process, requires ‘other than economic pressure’ if surplus-labour is actually to be carried out;

– these other than economic pressures take the form of the feudal master/slave relationship.

It seems to me that several conclusions follow.

Firstly, Marx tells us that surplus-value exists in its visible, palpable form (in sichtbarer, handgreiflicher Form existiert) in this mode of production, although surplus-value can only be recognized in its essence in the capitalist mode of production where it is hidden and therefore needs to be ‘analysed’. Surplus-value is par excellence a category of the capitalist mode of production which takes its meaning from the analysis of the ‘process of producing value’ (Verwertungsprozess), i.e., of a production process whose aim is an increase in exchange value (the latter, by the same token, being generalized as a form of value).

The justification for this statement is the fact that surplus-value is not aform’ in the same way that profit, rent and interest are; surplus-value is no more nor less than surplus-labour. The specific mode of exploitation of this surplus-labour in capitalist production, i.e., ultimately the mode of constitution of revenues (the mode of distribution), and therefore of the classes, is the constitution of profit, interest and capitalist rent, i.e., of what Marx calls the ‘transformed forms’ of surplus-value. In the capitalist mode of production, the forms of class struggle are first inscribed in the forms of the production process in general, they appear as a confrontation of forces within certain limits which are directly determined in the production process and analysable in it (limits of the working day, of wages, of profit and its sub-divisions).

In other words, if we inquire about the structure of the class relations in a given society of which we have already said that it was distinguished by a certain mode of extraction of surplus-value, we are inquiring first of all about the ‘transformed forms’ peculiar to that society.[9]

But it is no accident that the point which this passage singles out as the characteristic difference between the feudal mode of production and the capitalist mode of production – the coincidence and non-coincidence of necessary labour and surplus-labour – is also the essential point of the whole of Marx’s analysis in Capital of the capitalist mode of production alone: this coincidence is another way of expressing the term by term coincidence of the labour process and the process of producing value. The distinction between constant capital and variable capital which defines the process of producing value will always be found to correspond to the distinction between labour power and means of production peculiar to the labour process. Many examples from Capital could be adduced to show how the analysis demands reference to this correspondence (notably in the whole analysis of turnover). The worker’s labour materially transforms raw materials into a product by setting to work the means of production; the same labour transfers to the product the value of the means of production and materials consumed, and produces a new value, part – but only part – of which is equal to the value of the labour-power. In the last analysis, therefore, the dual character of the production process, which expresses this coincidence, refers to the dual character of ‘living’ labour.

It is easy to see that in the case Marx is describing here, the case of a form of feudal production, the coincidence exists in neither of the two forms: not only are labour and surplus-labour distinct ‘in time and space’, but even given a retrospective projection of the category of value, neither of the terms can strictly speaking be called a process of producing value.

In other words:

– in the capitalist mode of production, the two processes coincide ‘in time and space’, which is an intrinsic feature of the mode of production (of the economic instance); this coincidence is itself the effect of the form of combination of the factors of the production process peculiar to the capitalist mode of production, i.e., of the form of the two relations of property and real appropriation. The corresponding ‘transformed forms’ in this social structure, i.e., the forms of the relations between classes, are then directly economic forms (profit, rent, wages, interest), which implies notably that the State does not intervene in them at this level.

– in the feudal mode of production there is a disjunction between the two processes ‘in time and space’, which is always an intrinsic feature of the mode of production (of the economic instance) and an effect of the form of combination peculiar to it (the property relation appears in it in the dual form of ‘possession’ and ‘property’). Surplus-labour cannot then be extorted without ‘other than economic pressure’, i.e., without ‘Herrschafts- und Knechtschafstverhältnis ‘. Even before we have analysed the ‘transformed forms’ for themselves, we can conclude that in the feudal mode of production they will not be the transformed forms of the economic base alone, but of the ‘Herrschafts- und Knechtschaftsverhältnis ‘. Not directly economic, but directly and indissolubly political and economic ;[10] which means, finally, that different modes of production do not combine homogeneous elements, and do not allow differential divisions and definitions like the ‘economic’, the ‘legal’ and the ‘political’. Historians and ethnologists today often attest the discovery of this effect, though usually in a theoretically blind fashion.

We may also be able to understand why this politics was not conscious as such, why it did not think its relative autonomy, even in the moment when it occupied the determinant place, either in the form of ‘pure’ violence, or in the forms of a law, because it emerged as one of the presuppositions of the mode of production itself. Indeed, as we know, this relative autonomy of politics was not recognized in thought until much later: it is peculiarly a ‘bourgeois’ thought.

I think that it is possible to draw from this, one of Marx’s most detailed texts, the principle explicitly present in Marx of a definition of the determination in the last instance of the economy. In different structures, the economy is determinant in that it determines which of the instances of the social structure occupies the determinant place. Not a simple relation, but rather a relation between relations; not a transitive causality, but rather a structural causality. In the capitalist mode of production it happens that this place is occupied by the economy itself; but in each mode of production, the ‘transformation’ must be analysed. Here I merely suggest that we could try to re-read the first pages of The Origins of the Family in this perspective, the pages in which Engels expresses the following notion which he presents as a mere ‘correction’ of Marx’s general formulations:

According to the materialist conception, the determining factor in history is, in the last resort, the production and reproduction of material life. But this itself is of a two-fold character. On the one hand, the production of the means of subsistence, of food, clothing and shelter and the tools requisite therefore; on the other, the production of human beings themselves, the propagation of the species. The social institutions under which men of a definite historical epoch and of a definite country live are conditioned by both kinds of production: by the stage of development of labour, on the one hand, and of the family, on the other. The less the development of labour ... the more preponderatingly does the social order appear to be dominated by ties of sex (Marx-Engels: Selected Works, pp. 455-6).

A surprising text, which not only plays impudently on the term production, but demands the application of the technological model of the advance of the productive forces to the forms of kinship, presented as social relations of procreation! Perhaps it would be more worthwhile, as a number of Marxist anthropologists have been attempting, to show how, in certain ‘primitive’ or ‘self-subsistent’ societies, the mode of production determines a certain articulation of the social structure in which the kinship relations determine even the forms of transformation of the economic base.[11]

Chapter 2. The Elements of the Structure and their History

The definition of every mode of production as a combination of (always the same) elements which are only notional elements unless they are put into relation with each other according to a determinate mode, and the possibility this affords of periodizing the modes of production according to a principle of the variation of these combinations, are two propositions which of themselves alone deserve our attention. In fact, they convey the radically anti-evolutionist character of the Marxist theory of the history of production (and therefore of society). Nothing conforms less to the dominant ideology of the nineteenth century, the century of history and evolution to which Marx belonged, if we are to believe chronology. As we shall see better later, this is because Marx’s concepts are not intended to reflect, reproduce and mimic history, but to produce the knowledge of it: they are the concepts of the structures on which the historical effects depend.

In consequence, here there is neither a progressive movement of differentiation of the forms, nor even a line of progress with a logic akin to a destiny. Marx does tell us that all the modes of production are historical moments, but he does not tell us that these moments descend one from the other: on the contrary, the way his basic concepts are defined excludes such a facile solution. As Marx says in the 1857 Introduction that we have already quoted, ‘certain determinations are common to the most modern and to the most ancient epochs’ (e.g., co-operation and certain forms of direction, of accountability, which are common to ‘Asiatic’ modes of production and to the capitalist mode of production more than to all the others). This breaks the identity between chronology and a law of the internal development of forms which is at the root of evolutionism as of all historicisms of ‘supersession’. Marx’s aim was to show that the distinction between different modes is necessarily and sufficiently based on a variation of the connections between a small number of elements which are always the same. The announcement of these connections and of their terms constitutes the exposition of the primary theoretical concepts of historical materialism, of the few general concepts which form the rightful beginning of his exposition and which characterize the scientific method of Capital, conferring on its theory its axiomatic form; i.e., the announcement of a determinate form of this variation, one which directly depends on the concepts of labour-power, means of production, property, etc., is a constantly necessary presupposition of the economic’ proofs in Capital.

But is this some kind of ‘structuralism’? The suggestion is a tempting one, despite the risk of a confusion with thoroughly unscientific contemporary ideologies, in that it would redress the balance, for readings have traditionally leaned towards evolutionism and historicism. The ‘combination’ that Marx analyses is, to be sure, a system of ‘synchronic’ connections obtained by variation. However, this science of combinations is not a combinatory, in which only the places of the factors and their relationships change, but not their nature, which is not only subordinate to the system in general, but also indifferent: it is therefore possible to abstract from it and proceed directly to the formalization of the systems. This suggests the possibility of an a priori science of the modes of production, a science of possible modes of production, whose realization or non-realization in real-concrete history would depend on the result of a throw of the dice or on the action of an optimum principle. Historical materialism does authorize the prediction or even the reconstruction of ‘notional’ modes of production (as one might describe the ‘mode of simple commodity production’) which, never having been dominant in history, have never existed in an undeformed state. However, it does so in a different way, as will be explained later, on the basis of modifications in an existing mode of production. Otherwise, this would presuppose that the ‘factors’ of the combination were the very concepts I have listed, that these concepts directly designated the elements of a construction, the atoms of a history. In reality, as I have already said in a very general way, these concepts designate the elements of the construction only mediately: what I have called the ‘differential analysis of forms’ is an essential intermediate step in the determination of the historical forms taken by labour-power, property, ‘real appropriation’, etc. These concepts designate only what might be called the pertinences of historical analysis. It is this feature of the ‘combinatory’, which is therefore a pseudo-combinatory, that explains why there are general concepts of the science of history although there can never be a history in general.

In order to show how this pertinence works, I shall now return in a little more detail to a few of the problems of definition involving the two ‘connections’ which I have distinguished, taking the two articulations of the ‘combination’ separately in order to bring out their peculiar effects on the definition of the elements (‘factors’). These specifications are indispensable if we are to see that Marx was right to speak of a structure of the process of production, and if the combination of the factors is to be no mere descriptive juxtaposition, but an effective explanation of a functional unity.


The first connection that we inscribed in the ‘combination’ of a mode of production was designated as the ‘property’ connection, or connection of surplus-value appropriation; in fact, Marx constantly defines the ‘relations of production’ characteristic of a historical mode of production (and notably of capitalism) by its kind of ownership of the means of production, and therefore by the mode of appropriation of the social product which depends on it. The principle of this definition is well known. But a number of specifications are necessary, in order to bring out its exact structural function.

In the previous chapter, I concentrated above all on showing the difference between two concepts of appropriation, each of which refers to one aspect of the dual production process contained in every mode of production, and therefore defines one of the two connections which constitute the combination of the ‘factors’ of production. But it is no less important to take up Marx’s many hints and distinguish between the relations of production themselves, which are all that concern us here, and their ‘legal expression’, which does not belong to the structure of production considered in its relative autonomy. In this case, it is a question of distinguishing sharply between the connection that we have called ‘property’ and the law of property. This analysis is of fundamental importance in characterizing the degree of relative autonomy of the economic structure with respect to the equally ‘regional’ structure of the ‘legal and political forms’, i.e., in initiating an analysis of the articulation of regional structures or instances within the social formation.

This is also a decisive point for the history of theoretical concepts: Althusser has already recalled that the Marxist concept of ‘social relations’ marks a break with the whole of classical philosophy and with Hegel in particular, insofar as these relations do not represent forms of inter-subjectivity but relations which assign a necessary function to things as well as to men. Let us add that the Hegelian concept of ‘civil society’, adopted from the classical economists and designated by Marx as the main site of his discoveries, i.e., of his theoretical transformations, includes both the economic system of the division of labour and exchange, and the sphere of private law. There is therefore an immediate identity of appropriation in the ‘economic’ sense and legal property, and, in consequence, if the second can be designated as an ‘expression’ of the first, it is a necessarily adequate expression, or a duplication.

It is particularly interesting to note that certain of the clearest texts Marx devoted to the distinction between the social relations of production and their legal expression, concern precisely the possibility of a dislocation between base and superstructure, which, without this distinction, would obviously be incomprehensible. For example, in his analysis of the ‘Genesis of Capitalist Ground Rent’, he writes:

Since the direct producer [in the feudal mode of production] is not the owner, but only a possessor, and since all his surplus-labour de jure actually belong to the landlord, some historians have expressed astonishment that it should be at all possible for those subject to forced labour, or serfs, to acquire any independent property, or relatively speaking, wealth, under such circumstances. However, it is evident that tradition must play a dominant role in the primitive and undeveloped circumstance on which these social production relations and the corresponding mode of production are based. It is furthermore clear that here as always it is in the interest of the ruling section of society to sanction the existing order as law and legally to establish its limits given through usage and tradition. Apart from all else, this, by the way, comes about of itself as soon as the constant reproduction of the basis of the existing order and its fundamental relations assumes a regulated and orderly form in the course of time. And such regulation and order are themselves indispensable elements of any mode of production, if it is to assume social stability and indifference from mere chance and arbitrariness. These are precisely the form of its social stability and therefore its relative freedom from mere arbitrariness and mere chance... . It achieves this form by mere repetition of its own reproduction (Capital, Vol. III, pp. 773-4, modified).

Such a gap or discordance between the law and a ‘tradition’ which might seem a sub-law or a debased law, is therefore in reality the expression of a gap or discordance between the law and an economic relation (the individual producer’s necessary disposition of his plot of land), characteristic of periods of the formation of a mode of production, i.e., of the transition from one mode of production to another. A remarkable instance of the same effect is also featured in the analysis of the factory legislation that dates from the first period of the history of industrial capitalism and codifies the conditions of the ‘normal’ exploitation of wage labour-power (see Capital, T.II, pp. 159ff.; Vol. I, pp. 480ff.).

Since such gaps are possible, or more precisely, since contradictions are induced within the law itself by its non-correspondence with the relations of production, law must be distinct and second in order of analysis to the relations of production. And this is confirmed if we compare the passages where Marx reveals the specificity of ‘bourgeois’ property, e.g.:

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

with those that recall the chronological precedence, the precession of the (‘Roman’) legal forms of the right of property with respect to the capitalist mode of production, which alone generalizes the private ownership of the means of production. On this point I could refer to the text of Pre-Capitalist Economic Formations that has already been quoted (and is a very legal text, both in its object and in its terminology), or else to a letter from Engels to Kautsky:

Roman law was the consummate law of simple, i.e., pre-capitalist commodity production, which however included most of the legal relations of the capitalist period. Hence precisely what our city burghers needed at the time of their rise and did not find in the local law of custom (26 June 1884).

This comparison retrospectively illuminates the text on ‘The Genesis of Capitalist Ground Rent’ that I quoted above. It shows that the problem of the gap between a ‘tradition’ and a ‘law’ must not be interpreted as a theory of the genesis of the law out of the economic relations: for although the transition from a custom to a law does occur in history, this transition is not a continuity, but on the contrary, a rupture, a change in the law, or better: a change in the nature of law which is achieved by re-activating an older law (‘Roman’ law) which has already been superseded once. Nor is the repetition that seems to play an essential part in the articulation of the law with the economic relations here an element of this genesis, which, would explain the formation of a codified superstructure by virtue of its duration: its function is necessarily quite different, and refers us to the theoretical analysis of the functions of reproduction found in every mode of production, which we will discuss later. What we can see from the reproduction of economic relations is the necessary function of the law with respect to the system of economic relations itself, and the structural conditions to which it is therefore subordinate; but not the generation of the instance of the law itself in the social formation.

It is difficult, firstly, to distinguish clearly between the relations of production and their ‘legal expression’; this very concept of expression is difficult, too, once it no longer means duplication but rather the articulation of two heterogeneous instances; finally, so is the possible dislocation between the economic relations and the legal forms. All these preliminary difficulties are not accidental, they explain the method of investigation which must necessarily be followed here (and to which Marx himself shows the way, notably in his texts on pre-capitalist modes of production, which are closer to investigations than to systematic expositions). This method consists of looking for the relations of production behind the legal forms, or better: behind the secondary unity of production and law, which has to be disentangled. Only by this method will it eventually be possible to trace the theoretical boundary while still taking into account the ambivalent function that Marx assigns to legal forms: they are necessary and yet ‘irrational’, expressing and codifying the ‘economic’ reality which each mode of production defines in its own way, and yet simultaneously masking it. This represents a commitment to a regressive course – another attempt to determine gaps or differences which will be expressed negatively on the basis of the forms of the law, but this time within a completely self-contemporaneous system (a highly determinate mode of production: here the capitalist mode of production). Hence a difficult terminological problem as well, since the concepts in which the relations of production are expressed are precisely concepts in which the economic and the legal are indistinct, starting with the concept of property. What is ‘property’ insofar as it forms a system within the relatively autonomous structure of production, and logically precedes the law of property peculiar to the society considered? Such is the problem which must be initiated for capitalism too.

This commitment to an analysis of the relations between the economic structure of the capitalist mode of production and the law that corresponds to it demands a complete study of its own: that is why I must be satisfied here by giving a few hints which will serve as reference points. The steps in a proof can be outlined as follows:

(1) the whole of the economic structure of the capitalist mode of production from the immediate process of production to circulation and the distribution of the social product, presupposes the existence of a legal system ; the basic elements of which are the law of property and the law of contract. Each of the elements of the economic structure receives a legal qualification in the context of this system, notably the various elements of the immediate production process: the owner of the means of production, the means of production (‘capital’), the ‘free’ labourer, and the process itself, characterized legally as a contract.

(2) the peculiarity of the legal system we are discussing here (but not, of course, of every historical legal system) is its abstract universalistic character: by which I mean that this system simply distributes the concrete beings which can support its functions into two categories within each of which there is no pertinent distinction from the legal point of view: the category of human persons and the category of things. The property relation is established exclusively between human persons and things (or between what are reputed to be persons and what are reputed to be things); the contract relation is established exclusively between persons. Just as, in law, there is no diversity between persons, who are all or can all be owners and contractors, so there is no diversity between things, which are all or can all be property, whether they are means of labour or means of consumption, and whatever the use to which this property is put.

(3) this universality of the legal system reflects, in the strict sense, another universality which is part of the economic structure: the universality of commodity exchange, which as we know is only realized on the basis of the capitalist mode of production (although the existence of commodity exchange and the forms that it implies are much older); only on the basis of the capitalist mode of production is the set of elements of the economic structure distributed entirely as commodities (including labour-power) and exchangers (including the direct producer). These two categories thus correspond adequately to those which define the legal system (persons and things).

Thus the general problem of the relationship between the capitalist mode of production and the legal system which its functioning presupposes depends historically and theoretically on another problem: that of the relationship between the economic structure of the immediate process of production and the economic structure of the circulation of commodities. This necessary presence of ‘commodity categories’ in the analysis of the process of production explains the necessary presence of the corresponding legal categories.

(4) the social relations of production which are part of the structure of the capitalist mode of production can be characterized on the basis of their legal expression, by comparison, uncovering a series of dislocations between them.

Firstly, whereas the ‘law of property’ is characterized as universalistic, introducing no differences between the things possessed and their uses, the only property which is significant from the point of view of the structure of the production process is the ownership of the means of production, to the extent that, as Marx constantly reiterates, the latter function as means of production, i.e., are consumed productively, combined with ‘living’ labour and not hoarded or consumed unproductively. Whereas legal property is a right of consumption of any kind (in general: the right ‘to use and abuse’, i.e., to consume individually, to consume productively, to alienate – exchange – or to ‘squander’ – Capital, Vol. III, p. 804), the economic ownership of the means of production is not so much a legal ‘right’ to them as the power to consume them productively, depending on their material nature, on their adaptation to the conditions of the labour process, as a means of appropriating surplus-labour. This power does not come down to a law, but, as Althusser has already suggested, to a distribution of the means of production (notably a suitable concentration in quantity and quality). The economic relation is not based on the indifference of ‘things’ (and, correlatively, of commodities), but on an appreciation of their differences, which can be analysed according to two lines of opposition:

elements of individual consumption

elements of productive consumption


labour-power/means of production

(the reader will realize that this system of differences recurs in the analysis of the departments of aggregate social reproduction). Thus the gap between the social relations of production and the law of property can be characterized as a movement of extension or protraction, as an abolition of the divisions required by the structure of production: from ‘ownership of the means of production’ to property ‘in general’.

Secondly, the relationship established between the owner of the means of production (the capitalist) and the wage-labourer is, legally, a special form of contract: a labour contract. This is established on condition that labour is legally reputed to be an exchange, i.e., that labour-power is legally reputed to be a ‘commodity’, or a thing. Note that in its concept this transformation of labour-power into a commodity and the establishment of the labour contract are completely independent of the nature of the labour in which the labour is consumed. That is why the legal form of the wage-earner is, just as before, a universal form which applies both to productive labour, the work of transformation that produces surplus-value, and to all the other forms of labour that can generally be designated by the term ‘services’. But only ‘productive’ labour determines a relation of production, and productive labour cannot generally be defined by the relationship between the employer and the wage-earner, a relationship between ‘persons’: it presupposes that the economic sphere in which it takes place is taken into account (the sphere of immediate production, the source of surplus-value), i.e., the material nature of the labour and its objects, i.e., the nature of the means of labour with which it is combined. A few moments ago the ownership of the means of production, in the form of a legal relation between a person and a thing, appeared to us as a power over ‘living’ labour through the disposition of the means of production (which alone confer this power); in the same way, wage labour, insofar as it is a relationship inside the structure of production, in the legal form of a wage-service contract, appears to us now as a power over the means of production through the disposition of productive labour (which alone confers this power, i.e., determines an adequate consumption, not just any consumption). Thus the gap between wage labour as a social relation of production and the law of labour can be characterized as a movement of extension or protraction formally similar to the preceding one.

Hence two conclusions of the first importance:

– whereas from the legal point of view (from the point of view of the law implied by the capitalist mode of production, of course) the property relation, a relation between a ‘person’ and a ‘thing’, and the contract relation, a relation between a ‘person’ and a ‘person’, are two distinct forms (even if they are based on a single system of categories), the same is no longer the case from the point of view of the economic structure: the ownership of the means of production and productive wage labour define a single connection, a single relation of production. This follows directly from the two analyses outlined above.

– because this social relation is not legal in nature, although, for reasons that lie in the very nature of the capitalist mode of production, we are obliged (and Marx first of all) to describe it in the peculiar terminology of legal categories, it cannot be supported by the same concrete beings. The legal relations are universalistic and abstract: they are established between ‘persons’ and ‘things’ in general; it is the systematic structure of law which defines its supports as individuals (persons) confronted by things. Similarly, it is through their functions in the production process that the means of production are the supports of a connection in the economic structure, and this connection (as opposed to property and contract) cannot be defined for individuals, but only for social classes or representatives of social classes. The definition of the capitalist class or of the proletarian class therefore does not precede that of the social relations of production, but vice versa, the definition of the social relations of production implies a ‘support’ function defined as a class.

But a class cannot be the subject of property in the sense in which – legally – the individual is the subject of his property, nor a partner, nor ‘third party’, of a contract. We are not dealing here with the inherence of the object in its subject, or with the mutual recognition of subjects, but with the mechanism of the constant distribution of the means of production, hence with the entire capital and in consequence the entire social product (as Marx shows in the penultimate chapter of Volume Three of Capital: ‘relations of production are relations of distribution’). Classes are not the subjects of this mechanism but its supports, and the concrete characteristics of these classes (their types of revenue, their internal stratification, their relations to the different levels of the social structure) are the effects of this mechanism. The economic relation of production appears therefore as a relation between three functionally defined terms: owner class/means of production/class of exploited producers. Confirmation of this may be found especially in Part 7 of Volume One (‘The Accumulation of Capital’), where Marx shows how the mechanism of capitalist production, by productively consuming the means of production and the workers’ labour power, produces the labourers’ existence as an appendage of capital and makes the capitalist the instrument of accumulation, capital’s functionary. There is nothing individual about this connection, it is in consequence not a contract, but ‘invisible threads’ which bind the worker to the capitalist class, the capitalist to the working class (Capital, T.III, pp. 16, 20; Vol. I, pp. 573-4, 577-8). We therefore find that the social relation which determines the distribution of the means of production is instituted as a necessary relation between each individual of one class and the whole of the opposing class.


Among the general concepts to whose systematic articulation by Marx I referred in my analysis of the Preface to A Contribution, none, perhaps, presents such difficulties, despite all its apparent simplicity, as that of the productive forces, or, more exactly, of the level of the productive forces (or their degree of development). Indeed, the announcement of the concept alone immediately suggests two consequences which have been the source of fundamental misconstructions of Marx’s theory, but of which it must be said that they are not easy to avoid: first, to speak of ‘productive forces’, ‘forces’ of production, immediately suggests the possibility of a list – ‘the productive forces are the population, the machines, science, etc.’; at the same time, it suggests that the ‘advance’ of the productive forces may take the form of a cumulative progress, an addition of new productive forces or a replacement of certain of them by other, more ‘powerful’ ones (the craftsman’s tool by the machine). This leads to an interpretation of the ‘level’ or ‘degree of development’ which is all the more tempting in that it seems to be implied by the words themselves: a linear and cumulative development, a quasi-biological continuity. But if that were so, how could we explain the historical discontinuities expressly contained in the general theory, except by a theory of ‘qualitative change’, of the transformation ‘of quantity into quality’, i.e., a descriptive theory of the pattern of a movement which does not suppress its general structure? How could we avoid a mechanistic theory of historical movement in which the ‘dialectic’ is merely another name for a periodic, and periodically compensated and adjusted, dislocation or lateness of the other instances with respect to this development against which they are measured?

However, such a distribution quickly runs into remarkable difficulties: and all of them are related to the heteronomy of the ‘elements’ that must be added together to make Marx’s concept coincide directly with a description of the ‘facts’. Marx’s bourgeois critics have not failed to note that the ‘productive forces’ ultimately include not only technical instruments, but also the application of scientific knowledge to the perfection and replacement of those instruments, and ultimately science itself; not only a population of working strengths, but also the technical and cultural customs of this population, which history (for earlier modes of production) and industrial social psychology show to be more and more historically and sociologically ‘dense’ and complex; not only techniques, but also a certain organization of labour, or even a social and political organization (‘planning’ is an obvious example), etc. These are not arbitrary difficulties: they reflect the fact that Marx’s concept cannot be made to coincide with the categories of a sociology which, for its part, does proceed by the distribution and adding together of levels – the technological, the economic, the legal, the social, the psychological, the political, etc. – and which bases its peculiar historical classifications on these distributions (traditional societies and industrial societies, liberal societies and centralized-totalitarian societies, etc.). Moreover, these difficulties provide us with an index to an essential formal difference between Marx’s concept and categories of this kind: the fact that the concept of the productive forces has nothing to do with a distribution of this type. We must therefore start looking for its real features.

First let us stop and examine Marx’s formulation itself: ‘level’ and ‘degree’, are certainly expressions which suggest the possibility of at least a notional measurement, and the measurement of a growth. These expressions are thought to characterize the essence of the productive forces, and in consequence to define them in the specificity of a historical mode of production. But it is a common-place to note that the productivity of any labour, i.e., the ‘measure’ of this development, increased more in a few decades of industrial capitalism than in centuries of previous modes of production, whereas the ‘relations’ of production and the legal and political forms maintained a comparable rate of change; the same is true of the transformation of the means of labour (the equipment) which Marx calls the ‘Gradmesser der Entwicklung der menschlichen Arbeitskraft ‘. Besides, Marx says much more correctly, and whenever this level plays a direct part in economic analysis: the productive power of labour, the productivity of the power of labour (Produktivkraft).

In other words, as we shall see, the ‘productive forces’ are not really things. If they were things, the problem of their transport, their importation, would, paradoxically enough, be easier to resolve for bourgeois sociology (with the exception of a few ‘psychological’ problems of cultural adaptation) than it is for Marx – since his theory claims that there is a necessary connection or correlation between certain productive forces and a certain type of society (defined by its social relations). Bypassing the verbal illusion created by the term, we can already say that the most interesting aspect of the ‘productive forces’ is no longer their distribution or composition, but the rhythm and pattern of their development, for this rhythm is directly linked to the nature of the relations of production, and the structure of the mode of production. What Marx proved, notably in Capital, and what is alluded to in some well-known sentences in the Manifesto, is not the fact that capitalism has liberated the development of the productive forces once and for all, but the fact that capitalism has imposed on the productive forces a determinate type of development whose rhythm and pattern are peculiar to it, dictated by the form of the process of capitalist accumulation. It is this pattern which best characterizes, descriptively, a mode of production, rather than the level attained at any moment. (‘The law of increased productivity of labour is not, therefore, absolutely valid for capital. So far as capital is concerned, productivity does not increase through a saving in living labour, but only through a saving in the paid portion of living labour, as compared to labour expended in the past’ – Capital, Vol. III, p. 257).

But from the theoretical point of view, the ‘productive forces’, too, are a connection of a certain type within the mode of production, in other words, they, too, are a relation of production: precisely the one I have tried to suggest by introducing into the constitutive connections inside the mode of production, as well as a ‘property’ connection, a connection, B, of ‘real appropriation’, between the same elements: means of production, direct producers, even ‘non-labourers’, i.e., in the context of the capitalist mode of production, the non-wage-earners. I should now like to show that this really is a connection, or more rigorously a relation of production, by tracing the analysis to be found in the chapters of Capital devoted to the methods of formation of relative surplus-value; at the same time, we shall see better what the differential analysis of forms is.

Marx’s analysis takes up three chapters of Capital (Volume One, Chapters XIII, XIV and XV in the English translation) which are devoted to the forms of co-operation in manufacture and modern industry, and the transition from the one to the other which constitutes the ‘industrial revolution’. But this development is incomprehensible unless we refer it on the one hand to the definition of the labour process (Volume One, Chapter VII) and on the other to Chapter XVI of Volume One (‘Absolute and Relative Surplus-Value’) which is its conclusion.

The transition from manufacture to modern industry inaugurates what Marx calls the ‘specific mode of production’ of capitalism, or again the ‘real subsumption’ of labour beneath capital. In other words, modern industry constitutes the form of our connection which belongs organically to the capitalist mode of production.

At first, capital subordinates labour on the basis of the technical conditions given by historical development. It does not change immediately the mode of production. The production of surplus-value in the form considered by us – by means of a simple extension of the working day, proved, therefore, to be independent of any change in the mode of production itself (Capital, T.I, p. 303; Vol. I, p. 310).

The production of relative surplus-value revolutionizes out and out the technical processes of labour, and the forms of social grouping (die gesellschaftlichen Gruppierungen). It therefore presupposes a specific mode, the capitalist mode of production, a mode which, along with its methods, means and conditions, arises and develops itself spontaneously on the basis provided by the formal subsumption of labour under capital. In the case of this development, the formal subsumption is replaced by the real subsumption of labour under capital (Capital, Vol. I, p. 510, retranslated from Marx-Engels: Werke, Bd. XXIII, pp. 532-3).

The following considerations may be regarded merely as a commentary on these texts.

Firstly, the difference between formal subsumption and ‘real’ subsumption indicates the existence of a chronological dislocation in the formation of the different elements of the structure: capital as a ‘social relation’, i.e., the capitalist ownership of the means of production, exists before and independently of the ‘real’ subsumption, i.e., the specific form of our connection (real appropriation) which corresponds to the capitalist mode of production. The explanation for this dislocation and for the possibility of such dislocations in general is found in a theory of the forms of transition from one mode of production to another, which I shall leave aside for the moment. Let me merely underline the following: the simple, purely chronological dislocation is indifferent to the theory that we are studying; the ‘synchrony’ in which the concept of a mode of production is given simply suppresses this aspect of temporality and hence excludes from the theory of history every mechanical form of thought where time is concerned (any theory which asserts that anything featured at the same level in a chart of chronological concordances belong to the same time). Not only is there a dislocation between the emergence of the capitalist ownership of the means of production and the ‘industrial revolution’, but the industrial revolution is itself dislocated from one branch of production to another. The second dislocation is also suppressed by the theory. Finally, within a single branch, it proceeds by successive replacements of manual labour by ‘mechanized’ labour, in a rhythm subject to structural and conjunctural economic necessities; so much so that the ‘transition’ which is our object here appears as a tendency in the strict sense Marx gave that term, i.e., as a structural property of the capitalist mode of production: the essence of the ‘productive forces’ in the capitalist mode of production is to be constantly in the process of transition from manual labour to mechanized labour.

Let us recall in what this transition from manufacture to modern industry consists.

Both are forms of co-operation between the labourers (the direct producers), and this co-operation is only possible through their subjection to capital, which employs them all simultaneously. Both therefore constitute what can be called organisms of production, instituting a ‘collective labourer’: the labour process which is defined by the delivery of a finished use product (whether this use be an individual consumption or a productive consumption) requires the intervention of several labourers in a specific form of organization. Manufacture and modern industry are thus equally opposed to the individual handicraft. However, that is not the real break.

All co-operation may take simple or complex forms: in simple co-operation, there is a juxtaposition of labourers and operations. ‘Numerous labourers work together side by side, whether in one and the same process, or in different but connected processes.’ This form of co-operation is still found, particularly in agriculture. In the workshop of the guild master, the labour of the journeymen is usually performed in simple co-operation. The same is true of the primitive forms of manufacture, which consist simply in gathering the artisans into a single place of work. Complex co-operation, on the contrary, consists of an imbrication, of an intertwining of the labour. The operations performed by each worker successively or simultaneously are complementary, and only together do they give birth to a finished product. This form of co-operation (which is found in quite distant times in some sectors, e.g., metallurgy) constitutes the essence of the division of labour in manufacture: one piece of work is divided among the workers (until the eighteenth century this was called a single ‘oeuvre’ or ‘ouvrage’ in France).

Obviously, this division may have different origins. It may derive from a real ‘division’, after the complex operations of a single handicraft have been shared out among different labourers who thus become specialists in one fraction of the labour: or it may derive from the junction of several different handicrafts, subordinated to the production of a single useful product to which they all contribute, thus transforming these handicrafts post festum into fractions of a single labour. Marx analyses examples of both (the manufacture of pins, the manufacture of carriages); they depend on the physical properties of the product, but in any case, this process of formation disappears in the result which is a division of labour of the same form. The basic principle, the importance of which we shall soon discover, is the fact that the fractional operations can be performed as manual labour.[12] All the advantages of the manufacturing division of labour are derived from the rationalization of each component operation which is made possible by its isolation and by the specialization of the labourer: the improvement of movements and tools, increased speed, etc. It is therefore essential that this specialization is in fact possible, that each simplest possible operation is individualized. Instead of a break, we therefore find a continuity between handicraft and manufacture: the manufacturing division of labour arises as the extension of the analytical movement of specialization peculiar to handicrafts, a movement which simultaneously affects both the perfection of technical operations and the psycho-physical characteristics of the workers’ labour-power. These are merely two aspects, two faces of one and the same development

Indeed, manufacture is merely the extreme radicalization of the distinctive feature of handicrafts: the unity of labour-power and means of labour. On the one hand, the means of labour (the tool) must be adapted to the human organism; on the other, a tool is no longer a technical instrument in the hands of someone who does not know how to use it: its effective use demands of the worker a set of physical and intellectual qualities, a sum of cultural habits (an empirical knowledge of the materials, of the tricks of the trade up to and including the craft secret, etc.). That is why handicrafts are indissolubly linked to apprenticeship. Before the industrial revolution, a ‘technique’ was the indissociable ensemble of a means of labour or tool, and a worker, molded to its use by apprenticeship and habit. The technique is essentially individual, even if the organization of labour is collective. Manufacture retains these properties and pushes them to the limit: the inconveniences denounced from the beginning of fractional labour arise precisely from the fact that it maintains a rigorous coincidence of the technical process, which gives rise to more and more differentiated operations, adapted to more and more numerous and distinct materials, with the anthropological process, which makes individual abilities more and more specialized. The tool and the worker reflect one and the same movement.

The main consequence of this immediate unity is what Marx calls ‘manual labour as a regulating principle of social production’. This means that co-operation in manufacture brings workers into relationship, and only through their mediation, means of production. This fact emerges clearly if, for example, we consider the constraints to which the constitution of the ‘organisms of production’ must conform where the proportion of workers employed in different tasks is concerned: these are dictated by the characteristics of labour-power. The number of manual operations into which it is most advantageous to divide the labour, and the number of workers detailed to each functional task so that there is ‘work’ for all of them all the time, must be established empirically. This will fix the composition of a unity-group which is paralysed if even one of its members is missing, in exactly the same way as an artisan would be paralysed in the continuity of his labour-process if for some reason he could not perform any one of the operations required for the manufacture of his product (See Capital, T.II, p. 37; Vol. I, p. 347).

By replacing human strength in the function of tool-bearer, i.e., by suppressing its direct contact with the object of labour, mechanization produces a complete transformation of the connection between the labourer and the means of production. From then on, the information of the object of labour no longer depends on the culturally acquired characteristics of the labour-power, but is pre-determined by the form of the production instruments and by their functioning mechanism. The basic principle of the organization of labour becomes the necessity to replace the operations of manual labour as completely as possible by the operations of machines. The machine-tool makes the organization of production completely independent of the characteristics of human labour-power: at the same stroke, the means of labour and the labourer are completely separated and acquire different forms of development. The previous relationship is inverted: rather than the instruments having to be adapted to the human organism, that organism must adapt itself to the instrument.

This separation makes possible the constitution of a completely different type of unity, the unity of the means of labour and the object of labour. The machine-tool, says Marx, makes possible the constitution of a ‘material skeleton independent of the labourers themselves’ (Capital, T.II, p. 56; Vol. I, p. 367). An organism of production is now no longer the union of a certain number of workers, it is a set of fixed machines ready to receive any workers. From now on, ‘a technique’ is a set of certain materials and instruments of labour, linked together by a knowledge of the physical properties of each of them, and of their properties as a system. The process of production is regarded in isolation as a natural labour process: within the elements of the labour process, it constitutes a relatively autonomous sub-set. This unity is expressed in the emergence of technology, i.e., the application of the natural sciences to the techniques of production. But this application is only possible on the existing basis provided by the objective unity of the means of production (means and object of labour) in the labour process.

The collective labourer acquires the determination of what Marx calls ‘socialized labour’. It is impossible to explain the totality of conditions actually required by a particular labour process (leading to a determinate useful product) without considering it as a component labour process, an element of social production as a whole. And notably, the intellectual labour which produces the knowledges which are applied in any particular labour process must appear in its analysis (in the analysis of the technical division of this labour process). There are labourers in this co-operation who are not present at the work-place. The fact that this product of intellectual labour, science, is a free element so far as the capitalist is concerned (which besides is not completely the case) and seems to be a gift of society, is a different problem, one which does not arise in the analysis of the labour process. Similarly, the set of workshops or factories in which the same technique is applied, independently of the distribution of property, tends to become its field of application and experiment, constituting what Marx calls ‘practical experience on a wide scale’:

It is only the experience of the collective worker which discovers and reveals ... the simplest methods of applying the discoveries, and the ways to overcome the practical frictions arising from carrying out the theory – in its application to the production process, etc. (Capital, Vol. III, p. 103, modified).

Thus we see that as a consequence of the relationship between the elements of the combination, the natures of those elements themselves are transformed. This ‘collective worker’ in a relationship with the unity of the means of production is now a completely different individual from the one who formed the characteristic unity of artisan-manufacturing labour with different means of labour; at the same time, the determination of ‘productive labour’ has changed it support:

Once ... the individual product has been transformed into a social product, produced by a collective labourer, each member of which participates to a very different extent and from near or far or not at all in the manipulation of the material, the determinations of productive labour, and of the productive labourer become extended as a necessary consequence. In order to labour productively, it is no longer necessary for you to do manual work yourself; it is enough that you are an organ of the collective labourer, and perform one of its subordinate functions. The first determination given above of productive labour, a definition deduced from the very nature of the production of material objects, still remains correct for the collective labourer, considered as a single person. But it no longer holds good for each of its members taken individually (Capital, T.II, pp. 183-4; Vol. I, pp. 508-9).[13]

In our pseudo-combinatory, therefore, we do not really find the same ‘concrete’ elements when we move from one variant to the next. Nor is their particularity defined by a mere place, but rather as an effect of the structure, differing every time, i.e., an effect of the combination which constitutes the mode of production. I have taken this connection as an example because the analysis in Capital unravels every inch of it, but it is clear that an analysis of the same type could be conducted for the forms of property, not in the legal sense of the term, but in the sense of the relations of production pre- supposed and formalized by the legal forms. Marx outlines a hint towards such an analysis in the retrospective texts on The Genesis of Capitalist Ground Rent (Capital, Volume Three) and Pre-Capitalist Economic Formations (Grundrisse), making use notably of a formal distinction between ‘property’ and ‘possession’. His hints are enough to show that we should find forms which are as complex as those he reveals with respect to real appropriation.[14]


Before announcing the further consequences we can draw from this analysis, I must first show that it depends entirely on criteria for the differentiation of forms which are contained in the definition of the labour process.

The simple elements (die einfache Momente) into which the labour process breaks down are: (1) the personal activity of man, or labour strictly speaking (zweckmässige Tätigkeit); (2) the object on which that labour acts (Gegenstand); (3) the means with which it acts (Mittel) (Capital, T.I, p. 181; Vol. I, p. 178).

What most people remember about Marx’s analysis of the industrial revolution is what distinguishes it from other explanations of the same ‘phenomenon’: the fact that he attributed the origin of the technical and social upheavals to the introduction of the machine-tool, to the replacement of man as the tool-bearer, instead of attributing it to the introduction of new sources of energy (the steam engine), to the replacement of man as the motor. But it is less usual to dwell on the theoretical expression of this originality, which is contained in the definition of the labour process. The industrial revolution (the transition from manufacture to modern industry) can be completely defined, with the assistance of these concepts, as the transformation of the relationship which followed from the replacement of the means of labour. Returning to what I said above about this transformation, summarizing Marx, it could be represented as the succession of two ‘material forms of existence’ of the labour process:[15]

– unity of the means of labour and of the labour power,

– unity of the means of labour and of the object of labour;

in both cases, the pattern of the relationship between the three elements is completely characterized by designating the sub-set which has a unity and relative autonomy.

– object of labour







unity of mechanization,



– means of labour







handcraft (and manufacturing) unity,


– labour power

It is obvious straightaway that the three concepts of the definition of the labour process have nothing to do with the abstraction of an empirical description (subject, object, ‘mediation’), which can always be recast by distinguishing other elements. They are not derivatives of the analysis of the two successive forms of the connection. They make that analysis possible.

Thus the movement from one form to the other can be completely analysed: not as the mere dissolution of a structure (the separation of the labourer from the means of labour), but as the transformation of one structure into another. Nor as the constitution ex nihilo of a structure although it is original (the unity of object and means of labour in a single system of physical interactions) (or as the accidental formation of that structure by the convergence of those two abstractions (’science’ and ‘technique’): for it is the forms of the labour process which have changed. The new system of the productive forces, of which modern capitalist mechanized industry is the first example, is neither an absolute end nor an absolute origin, but a reorganization of the entire system, of the relation of the real appropriation of nature, of the ‘productive forces’.

But at the same time it is quite clear that this change in form could not have been analysed at all as the linear movement of a development, as a lineage. There is such a lineage between handicrafts and manufacture, since, as we have seen, manufacture can be regarded, from the point of view which concerns us, as the continuation of a movement peculiar to handicrafts, and one which conserves all its characteristics. But the machine which replaces the ensemble of tools and educated, specialized labour-power is in no way a product of the development of that ensemble. It merely occupies the same place. It replaces the previous system by a different system: the continuity is not that of elements or individuals, but of functions. This type of transformation can be designated by the general term displacement.

Here I should like to make a digression, though not an arbitrary one, and compare this kind of reasoning with the very interesting and very surprising method followed by Freud in his texts on the history of the libido (notably the Three Essays on the Theory of Sexuality). The analogy is precise enough to encourage this comparison, which will perhaps seem even more justified if we think how akin were the ideological situations in which and against which Marx and Freud had to construct their theories, and how alike sometimes even the concepts of those ideologies were. Evolutionism reigns as supreme in the science of history as it does in ‘psychology’. The terms Freud uses in the Three Essays refer to a psychological evolutionism, exactly as Marx’s terms ‘level’, ‘degree of development’ of the productive forces, refer to a historical evolutionism (in the Preface to A Contribution, Marx speaks of the replacement of the existing social relations by ‘new, higher’ relations). Therefore (to forestall any ambiguity) I am not interested here in the articulation of the objects of psycho-analysis and historical materialism, but in the possibility of revealing epistemological analogies between Marx’s theoretical work and Freud’s.

Indeed, on the one hand we find in these texts of Freud’s a whole biological or quasi-biological theory of the stages of development of the libido (sexual instinct), a problematic of the congenital constitution and established nature of the ‘germs’ whose development will constitute the successive stages. We find a theory of development and of its intermediate degrees, which at the same time justifies a theory of the pathological as the fixation on one stage of development or a regression to it (but a regression is always merely the revelation of a fixation), etc.

But on the other hand, in contrast with what would be a real evolutionist theory, although in the very same terms, we find something completely different.

For example, in the following passage:

A difficult question and one which cannot be evaded: what is the general characteristic which enables us to recognize the sexual manifestations of children? The concatenation of phenomena into which we have been given an insight by psycho-analytic investigation justifies us, in my opinion, in regarding thumb-sucking as a sexual manifestation and in choosing it for our study of the essential features of infantile sexual activity (Three Essays, pp. 180-1).[16]

This is one example of a reasoning Freud generalizes in this study, which consists of making a series of organizations of the search for pleasure into the successive forms of a single sexual instinct. ‘The final outcome of sexual development lies in what is known as the normal sexual life of the adult’ (the formulation in the Introduction to Psychoanalysis gives a more complex chain, since Freud uses in his definition both infantile sexuality and ‘abnormal’ adult sexuality: hence the final outcome of the development is either ‘normal’ sexuality or perversion and neurosis, which have the same place in the ‘abnormal’ outcome). Paradoxically, the origins of the development are the stages which are least obviously of a ‘sexual’ character. In reality, they only acquire this character because analysis discovers for them the same function. The succession of these stages is much better analysed as a series of displacements than as a continuity: a displacement of the erotogenic zones, i.e., of the parts of the body invested with a sexual ‘value’ in a given libidinal organization (Freud tells us that there is hardly any part of the body that cannot be treated in this way); a displacement of the biological functions which ‘prop up’ the sexual instinct initially; a displacement of the objects of the instinct, from what Freud calls the absence of an object, but which is a particular modality of an object, to the object of genital love. Each of these displacements corresponds to one variant of the relations between what Freud calls the ‘component instincts’, i.e., the components of the complex sexual instinct.

In the second place we have found that some of the perversions which we have examined are only intelligible if we assume the convergence of several motive forces. If such perversions admit of analysis, that is, if they can be taken to pieces, then they must be of a composite nature. This gives us a hint that perhaps the sexual instinct itself may be no simple thing, but put together from components which have come apart again in the perversions. If this is so, the clinical observation of these abnormalities will have drawn our attention to amalgamations which have been lost to view (Three Essays, p. 162).

Each of these variants is a system of organization of the complex sexual instincts implying a relation of dominance or hierarchy within the ‘component instincts’ (pre-genital or genital organizations – primacy of the genital erotogenic zone) (see Three Essays, pp. 197ff.).

Thus Freud’s reasoning in these pages sets to work a series of concepts which only superficially have anything to do with a theory of the evolution of the individual, or with a biological model of the latter. This reasoning must answer two questions at once: what form does the development take and what is its subject, what is it that develops?[17] This reasoning seems to be inseparable from a new definition of the ‘sexuality’ which is the object of the analysis (Freud is constantly dealing with objections which are related to this ‘extension’ of the notion of sexuality and which confuse it with the protraction of ‘genital’ sexual activity to periods before puberty). Finally, it emerges that sexuality is defined quite simply by the succession of forms between which suchdisplacements’ can be analysed. Anything is sexual which in an element of an organization of the component instincts, the final outcome of whose variations is genital organization.

But what makes it possible to analyse these displacements is a set of theoretical concepts which plays a part analogous to that of the concepts which define the labour process in the analysis of the forms of the connection of real appropriation (‘productive forces’): activity/object/means of labour. In Freud, these concepts are used systematically in the Three Essays and presented systematically in the article Instincts and their Vicissitudes (Standard Edition, Vol. XIV): they are the concepts of the source (Quelle), pressure (Drang), object (Objekt) and aim (Ziel) of the instinct. Of course, there is no question of any correspondence between Freud’s concepts and those of Marx: but rather one of the same type of analysis, and hence of an identity of the functions of these concepts in the method.

Perhaps in return we shall now be able to illuminate the problems posed by Marx’s text. Notably the difficulty Marx found in isolating the connection that I have discussed, or, what amounts to the same thing, in thinking the ‘level of the productive forces’ as a connection within the combination, i.e., as a relation of production with the same status as the forms of the ownership of the means of production.[18]

This difficulty is accompanied by the temptation to list the productive forces, and, for example, to divide them between nature and man. Similarly, these texts of Freud’s contain formulations which attempt to situate the sexual instinct, as described by analysis, with respect to the domains of biology and psychology; Freud ends by defining instinct as a frontier between the biological and the psychological, and he even locates this ambiguity at the level of the ‘source’ of instinct (see Instincts and their Vicissitudes, op. cit., p. 123: ‘By the source of an instinct is meant the somatic process which occurs in an organ or part of the body and whose stimulus is represented in mental life by an instinct. We do not know whether this process is invariably of a chemical nature ... The study of the sources of instincts lies outside the scope of psychology. Although instincts are wholly determined by their origin in a somatic source, in mental life we know them only by their aims’). In the analysis of forms, the biological is therefore always absent as such. The sought after ‘frontier’ is thereby strictly non-existent. But we should add that in another sense the psychological, too, is absent: in its traditional conception, it, too, was defined by its opposition and relation to the biological. If the latter disappears as such, the psychological is transformed into something other than itself: into precisely what Freud called the ‘psychical’. We are therefore always dealing with a series of reorganizations and displacements of the domains whose links Freud himself has very clearly conceived. In the Introductory Lectures on Psychoanalysis, Freud writes:

Whereas for most people ‘conscious’ and ‘psychical’ are the same, we have been obliged to extend the concept of ‘psychical’ and to recognize something ‘psychical’ that is not ‘conscious’. And in just the same way, whereas other people declare that ‘sexual’ and ‘connected with reproduction’ (or, if you prefer to put it more shortly, ‘genital’) are identical, we cannot avoid postulating something ‘sexual’ that is not ‘genital’ – has nothing to do with reproduction. The similarity here is only a formal one, but it is not without a deeper foundation (Standard Edition, Vol. XVI, p. 32I).

We should add, simply, that this ‘extension’ is in fact a completely new definition, in content as well as in the nature of the theoretical discourse by which it is justified.

The same is true of ‘nature’ in the analysis of the productive forces. For although Marx writes that ‘labour is, in the first place, an action which takes place between man and nature ... In it man has the role of a natural power with respect to nature’, it would perhaps be equally correct to say that nature has the role of a social element. In this sense, too, ‘nature’ as such is absent.

Insofar as the Marxist analysis of the ‘productive forces’ is systematically inscribed in the definition of a mode of production, i.e., insofar as it is not a simple list or description of the ‘technical’ aspects of production or its ‘resources’, but rather the definition of a form of variation of the ‘technical’ social relations of production, it therefore achieves the same effect of displacement and rupture with respect to the traditional theoretical division of labour as that which we have found in Freud. This rupture effect is characteristic of the founding of a new science which is in the process of constituting its object and defining for it a domain which a variety of disciplines were previously occupying and for that reason ignoring completely. In the domain of historical materialism, as a scientific theoretical discipline, the analysis of the productive forces does not arise as a technical or geographical preliminary, formulating the conditions or bases on which a ‘social’ structure of human institutions and practices can be constructed, as an essential, but external limitation imposed on history: on the contrary, it is inside the definition of the social structure of a mode of production (no definition of a ‘mode of production’ can be regarded as satisfactory unless it includes a definition of the productive forces which are typical of that mode of production); it therefore completely transforms the meaning of ‘social’.

But, as we have seen, the analogy goes further: it also extends to the type of object and history that Marx and Freud defined. Just as the ‘sexual’ that Freud discusses is not the subject of the development staked out by the organizations of the instincts, just as the organizations of the instincts do not strictly speaking descend one from another, so in Marx’s analysis we are never dealing with anything other than the combination itself and its forms. Thus, in Marx’s case, too, we can say that the subject of development is nothing but what is defined by the succession of the forms of organization of labour and the displacements that it achieves. Which reflects exactly the theoretical, non-empirical character of the constitution of his object.


This analysis has very important consequences for the theory of history. Indeed, we should ask what has really been achieved by this analysis of two successive forms: we should pose the question of whether this can be called ‘a history ‘. This definition would be manifestly meaningless unless we could at the same time designate the object of this history. Whatever the mode of this designation, by a concept or by a mere name, we can never conceive history in general, but only the history of something.

We should note that most historians have, until really quite recently, avoided the necessity of giving a theoretical answer to this problem of their object. Take for example Marc Bloch’s reflections on the ‘science of history’; it is clear that all his efforts are devoted solely to the constitution of a methodology. The attempt to define the object of the historian’s work is indeed revealed as aporetic, once it has been demonstrated that this object cannot be ‘the past’, nor ultimately any pure and simple definition of time: ‘the very idea that the past as such could be the object of science is absurd’ (Marc Bloch: The Historian’s Craft, trans. Peter Putnam, Manchester 1954, p. 22). Nevertheless, after this negative and perfectly cogent conclusion (although its consequences have not always been drawn – by the philosophers), attempts such as Bloch’s are content with an incomplete definition of their science which relegates the problem of the object to the indefiniteness of a totality: ‘man, let us say rather, men’, and characterizes knowledge solely as a certain set of methods. Here is not the place to analyse the empiricism that ultimately flows from this incomplete definition, but we should note that the problem evaded theoretically is necessarily solved practically at every moment. That is why we have political histories, histories of institutions, histories of ideas, histories of the sciences, economic histories, etc.

In this perspective we could undoubtedly define the object which was the concern of the above analysis as ‘labour’ and say that it was a history of labour, or a moment of such a history.

But at the same time, we see that Marx’s analysis was presented in an essentially polemical situation with respect to what is usually called ‘labour history’ or ‘technical history’. Such histories exist, and they receive but do not constitute objects which are claimed to persist in a certain identity of nature, through all their changes. These histories require a ‘subject’ to unify them, and they find one in technology, regarded as a ‘fact’ (even as a ‘fact of civilization’), or in labour, regarded as a kind of cultural ‘behaviour’. To say that they receive these objects is quite simply to say that the moment of their constitution lies outside the historian’s theoretical practice itself, but is a part of other practices, theoretical or otherwise. From the viewpoint of theoretical practice, the constitution of the object is therefore presented as a designation, as a reference to another practice; it is therefore only possible from the point of view of the personal identities of the men who are implied in all these practices at once, in a historian’s theoretical practice, and in political, economic and ideological practices. This reference is therefore only possible as an effect of the complex historical unity and of the historical articulation of these different practices, but as it is given, as it is reflected uncritically in a privileged site, the ideology of the period. But at the same time, because they are a paradox – a discourse (supposedly critical par excellence) which depends for the constitution of its object on an uncritical operation – these histories encounter, in their conceptualization and in the nature of their explanations, the insoluble problem of the mutual frontiers of these received objects, and finally, of the relation between this component history and other histories, and the history of the totality. As Vilar says of economic history, their description of the change, the movement of their peculiar object leads them to the insertion of this movement into a reality wider than their objects considered in their ‘purity’ (the ‘pure’ economy, ‘pure’ technique, etc.), which is the totality of human relations and explains this change (see Contributions à la première Conférence Internationale d’Historie Économique, Stockholm, 1960, p. 38). They discover that their objects change, that their objects have a history because what they are not changes too. It thus appears that the constitutive problem of each history is that of the relation between its object and history in general, i.e., other historical objects, and they solve it, when they are prepared to go beyond empiricism, sometimes by the announcement of a global, undifferentiated relation, which ultimately results in a theory of the ‘spirit of the age’, a ‘historical psychology’ (see for example Francastel’s work on the history of the plastic arts, and I. Meyerson’s theories), sometimes by the complete reduction of one structure to another, which thus emerges as the absolute reference, the original text of several translations (see for example the works of Lukács and his disciple Goldmann on literary history).

When I say that Marx presents his analysis in a polemical situation with respect to this historical practice, I do not mean that this analysis suppresses the problem of the relation between component histories and general history – a problem which must necessarily be solved before it is possible to speak strictly of ‘a history’. On the contrary, it shows that this problem cannot be solved unless history really constitutes its object, instead of receiving it. In this sense, the term analysis used by Marx has exactly the same significance as that given it by Freud when he speaks of the ‘analysis of an individual history’: just as Freud’s analysis produces a new definition of his object (sexuality, the libido), i.e., really constitutes it by showing the variation of its formations, which is the reality of a history, so Marx’s analysis constitutes his object (the ‘productive forces’) by constructing the history of its successive forms, i.e., forms which have a determinate place in the structure of the mode of production.

In his determination of the object of a component theory, Marx’s method thus completely abolishes the problem of ‘reference’, of the empirical designation of the object of a theoretical knowledge, or of the ideological designation of the object of a scientific knowledge. In fact, this determination now depends entirely on the theoretical concepts which make it possible to analyse in a differential way the successive forms of a connection, and the structure of the mode of production to which this connection belongs. ‘Labour’ is presented as a connection between the elements of the mode of production, and therefore its constitution, as an object of history, depends entirely on a recognition of the structure of the mode of production. We can generalize this comment and say that each of the elements of the combination (Verbindung) undoubtedly has a kind of ‘history’, but it is a history without any locatable subject: the real subject of each component history is the combination on which depend the elements and their relations, i.e., it is something which is not a subject. In this sense we can say that the first problem for a history as a science, for a theoretical history, is the determination of the combination on which depend the elements which are to be analysed, i.e., it is to determine the structure of a sphere of relative autonomy, such as what Marx calls the process of production and its modes.

In fact, this preliminary determination provides a determination of the component object, and, at the same stroke, that of its articulation with the other component objects. Which is to say once again that the knowledge of one instance of the social formation through its structure includes the theoretical possibility of knowing its articulation with other instances. This problem then emerges as the problem of the mode of intervention of the other instances in the history of the instance analysed. On this point, too, the preceding analysis provides us with an excellent example: the example of the application of science to production, i.e., the articulation of (economic) production with another practice: the theoretical practice of the natural sciences. In his study of the ways of economizing on constant capital in order to raise the rate of profit, Marx writes:

The development of the productive power of labour in any one line of production, e.g., the production of iron, coal, machinery, in architecture, etc., which may again be partly connected with progress in the field of intellectual production, notably natural science and its practical application, etc. (Capital, Vol. III, p. 81).

A text of this kind contains absolutely no implications that ‘intellectual production’ is a branch of production in the economic sense of the term But it does mean that intellectual production intervenes in the history of the mode of production (in the strict sense) through its products, which are susceptible to importation (knowledges). And the analysis of the displacement of elements within the mode of production, which I have reproduced above, alone enables us to explain why and in what form this intervention takes place. This analysis cancels out all the questions that have been posed as to the technological ‘routine’ of the ancient world and the middle ages, since the application of science to production is not determined by the ‘possibilities’ of that science, but by the transformation of the labour process which is an organic part of the combination of a determinate mode of production. It is determined by the constitution of the system which I have called the unity of the means of labour and of the object of labour. Not only is it therefore essential to seek in the analysis of the mode of production itself for the conditions which explain its relation with other practices, but the definition of this relation depends on the same theoretical concepts as those that designate the structure of the mode of production itself, in which the specific form of the other practices is as such absent. These other practices intervene through their specific products under conditions, or more accurately, as Marx says, within limits, which express the current essence of the mode of production (we shall see this in more detail with respect to the articulation of the political practice of the class struggle with the economic structure). Such also is one of the senses of the concept of ‘methods’ which Marx uses in relation of the production of relative surplus-value (see the passage quoted above, Capital, Vol. I, p. 510) as well as in relation to the (political) ‘methods’ of primitive accumulation; perhaps one could suggest that for Marx this concept always designates the intervention of one practice in conditions determined by another – the articulation of two practices.

On this model, we can formulate the indispensability of other histories than those of the modes of production, histories whose objects remain to be constituted. Not all histories are possible: historical research, via controversies in economic history, the history of ideas, mentalities, etc., is beginning to sense this, although it has not explicitly posed the problem of this constitution. The determination of the objects of these histories must await that of the relatively autonomous instances of the social formation, and the production of concepts which will define each of them by the structure of a combination, like the mode of production. We can predict that these definitions, too, will always be polemical definitions, i.e., they will only be able to constitute their objects by destroying ideological classifications or divisions which benefit from the obviousness of the ‘facts’. Attempts like that of Foucault give us a good example of this.[19] It might be suggested – to enter the realm of conjecture – that the history of ideologies, and notably the history of philosophy, are perhaps not histories of systems, but histories of concepts organized into problematics, whose synchronic combinations it is possible to reconstitute. I am referring here to Althusser’s work on the anthropological problematic to which Feuerbach and the Young Marx belonged, and on the history of philosophy in general. Similarly, the history of literature may not be that of the ‘works’, but that of another object, a specific one, i.e., a certain relation to the ideological (itself already a social relation). In this case, too, as Pierre Macherey suggests (‘Lénine, critique de Tolstoï’, La Pensée no. 121, June 1965 or Pour une théorie de la production littéraire, Maspero, Paris 1966), the object under consideration would be defined by a complex combination whose forms all have to be analysed. Obviously, these are only programmatic hints.

If the theory of history implied by Marx’s method of analysis is really like this, we can produce a new concept which belongs to that theory: I shall call it the concept of the differential forms of historical individuality. In the example which Marx analysed, we see that the two successive forms of the ‘productive forces’ connection imply two different forms of historical individuality. In the example which Marx analysed, we see that the two successive forms of the ‘productive forces’ connection imply two different forms of individuality for the ‘labourer’, who is one of the elements of the connection (and similarly, two different forms of the means of production): in the first case, the ability to set the means of production to work belonged to the individual (in the ordinary sense), it was an individual mastery of these means of production; in the second case, the same ability only belongs to a ‘collective labourer’, it is what Marx calls a ‘social’ mastery of the means of production. The productive forces developed by capitalism thus institute a norm which is not valid for any individual. On the other hand, this historical difference is strictly relative to the combination considered, i.e., it only concerns the practice of production. We can say that each relatively autonomous practice thus engenders forms of historical individuality which are peculiar to it. This observation results in a complete transformation in the meaning of the term ‘men’, which, as we have seen, the Preface to A Contribution made the support for the whole construction. We can now say that these ‘men’, in their theoretical status, are not the concrete men, the men of whom we are told, in famous quotations, no more than that they ‘make history’. For each practice and for each transformation of that practice, they are the different forms of individuality which can be defined on the basis of its combination structure. Just as, in Althusser’s words, there are different times in the social structure, none of which is the reflection of a common fundamental time, so for the same reason, i.e., what has been called the complexity of the Marxist totality, there are different forms of political, economic and ideological individuality in the social structure, too, forms; which are not supported by the same individuals, and which have their own relatively autonomous histories.

Besides, Marx formulated the very concept of the dependence of the forms of individuality with respect to the structure of the process or the ‘mode’ of production. His terminology itself is marked by the epistemological fact that in the analysis of the ‘combination’ we are not dealing with concrete men, but only with men insofar as they fulfil certain determinate functions in the structure: – bearers of labour power (with respect to the labour process, in his exposition of the theoretical concepts which define the analysis, Marx does not, as we have seen, say ‘man’ or ‘subject’, but ‘zweckmässige Tätigkeit ‘, activity which conforms to the norms of the mode of production); – representatives of capital.

To designate these individuals, he systematically used the term Träger, which is most often translated into English as support. Men do not appear in the theory except in the form of supports for the connections implied by the structure, and the forms of their individuality as determinate effects of the structure.

We might perhaps import the term pertinence to designate this characteristic of Marxist theory, and say that each relatively autonomous practice in the social structure must be analysed according to its own pertinence, on which depends the identification of the elements which it combines. There is no reason why the elements, which are thus determined in different ways, should coincide in the unity of concrete individuals, who would then appear as the local, miniature reproduction of the whole social articulation. The supposition of such a common support is, on the contrary, the product of a psychological ideology, in exactly the same way as linear time is the product of a historical ideology. It is this ideology which supports the whole problematic of mediations, i.e., the attempt to rediscover concrete individuals, the subjects of psychological ideology, as the centres or ‘intersections’ of various progressively more external systems of determination, culminating in the structure of economic relations, systems which constitute a series of hierarchized levels. This is a modern form of what Leibniz expressed perfectly when he said that each substance with a degree of singularity, and in particular each mind, expresses the whole universe in a specific way:

Minds ... in a manner ... express and concentrate the whole into themselves, so that it may be said that minds are total parts (‘De rerum originatione radicali’, in The Monadology and other Philosophical Writings, trans. Robert Latta, Oxford 1898, p. 349).

Similarly, if men were the common supports of determinate functions in the structure of each social practice, they would ‘in a manner express and concentrate’ the entire social structure into themselves, i.e., they would be the centres from which it would be possible to know the articulation of these practices in the structure of the whole. At the same time, each of these practices would be effectively centred on the men-subjects of ideology, i.e., on consciousnesses. Thus the ‘social relations’, instead of expressing the structure of these practices, of which individuals are merely the effects, would be generated from the multiplicity of these centres, i.e., they would have the structure of a practical inter-subjectivity.

As we have seen, Marx’s whole analysis excludes this possibility. It forces us to think, not the multiplicity of centres, but the radical absence of a centre. The specific practices which are articulated in the social structure are defined by the relations of their combination before they themselves determine the forms of historical individuality which are strictly relative to them.

Chapter 3. On Reproduction

In everything that I have said so far, I have only been defining a single concept: ‘mode of production’, on the basis of the use that Marx made of it in his analysis of the capitalist mode of production. I have outlined what might be called the first theoretical effects peculiar to this concept: all the terms the function of which in Marx’s exposition I have attempted to pin-point have only acquired their meaning by reference to this first definition; their intervention in a proof thus appears as an extension of the effectivity of the ‘presuppositions’ implied by the definition of a mode of production; the transformations in the way history is thought contained in these terms, transformations which at the same time have the meaning of a transition from ideology to science, are merely the effects of a single theoretical event: the introduction of the concept of a mode of production into the traditional problematic of periodization.

But to stop there would leave us facing a difficulty which I have already referred to in my discussion of ‘component histories’ in the normal practice of historians: I have pointed out the stumbling-block of these histories, which do not constitute their object on the basis of a historical definition but receive it ready constituted, the problem of the location of that object in a totality of historical objects. This location is always something already established for theoretical discourse (for the discourse that aims to be theoretical), established by a non-theoretical operation which refers to the more or less immediate obviousness in which this object proposes its existence and consistency; thus, in the last analysis, it presents itself as a recourse to gesture, to the gesture which shows the objects of a world, whose conceptual representatives one only then proposes to deal with in a theoretical discourse. But we also know that this gesture is only apparently an innocent one, that in reality it is inhabited by an ideology which governs the division of the world into objects, and, in the same movement, the ‘perception’ of these objects, what has elsewhere been designated as the allusive nature of ideology. We know this from the moment a science breaks, constituting other objects in polemical rupture with the previous ones.

The difficulty we are now about to meet is of an analogous kind, and we shall not lack examples to persuade us that this difficulty is no fabrication. We now have the theoretical concept of a mode of production, or more precisely, we have it in the form of the knowledge of one particular mode of production, since, as we have seen, the concept only exists if it is specified. However, it seems that we still need to know something else, i.e., when and where the concept isapplicable ‘, what societies, at what moments in their histories, have a capitalist mode of production. Indeed, the whole problem of periodization seems to be concentrated in this point: it is not enough to have at one’s disposal a theoretical analysis of the effects which depend on the structure of each mode of production, once one has formulated its ‘presuppositions’ – it is also necessary to build an actual history with them, quite simply, real history, our history, which presents these different modes of production here or there, one after the other. A true knowledge tells us, i.e., we know theoretically, what the capitalist mode of production is, but we also want to know if this knowledge is really the knowledge of England in 1840 or of France in 1965, etc. This is a problem of identification or judgment: it seems that we need rules to determine which objects in experience fall within the concept of the capitalist mode of production. It is this apparent necessity which gives rise to the empiricist interpretation of theoretical practice as a practice which constitutes ‘models’: in this view, the entire theory of Capital is a study of the properties of a model, properties which are valid for every production that is an ‘example’ or ‘case’ of the structure. The identification of the cases, the actual subsumption, is, in this ideology of models, in every respect a pragmatic process, a gesture, however complicated the forms in which it is achieved (by which I mean, even if this identification is not made at one stroke, but through a series of partial identifications in which the elements of the structure and its particular effects are discovered). As such, it is a non-theoretical process which depends, not on concepts, but on properties of the identifier, properties which might well be called psychological even where a scientific consciousness is concerned. Kant already said that good judgment is a gift which cannot be learnt, and that the basis of judgment is a profound mystery (for theory).

Nevertheless, this route whose mere exercise subordinates theoretical practice to a non-theoretical faculty seems to be implied, at least negatively, like the space within a mould, in certain terms which Marx applies to his own object in Capital. I shall only recall a few of these texts here, for I have commented on them several times already. Marx tells us that he only studied the mode of production ‘in its ideal average’ (Capital, Vol. III, p. 810). Which does not only mean that one abstracts from the ‘particular’ effects, from the ‘accidental’ circumstances or ‘superficial’ traits, in order to study the general structure itself, but also that one studies a structure which is not peculiar to any particular time or place. This is also the meaning of the famous reference to England:

In this work I have to examine the capitalist mode of production, and the relations of production and exchange corresponding to that mode. Their classical ground is England. That is the reason why I have taken the chief facts and examples which illustrate the development of my theories from England. If, however, the German reader shrugs his shoulders ... I must; plainly tell him, ‘De te fabula narratur!’ (Capital, preface to the first German edition, T.I, p. 18; Vol. I, p. 8).

We must take this text strictly literally, and say that the object of the theory is itself a theoretical object at a determinate level of abstraction. The mode of production, the relations of production and exchange, these things are what is known in Capital, not England or Germany (besides, a whole book could be written on the history of the theoretical destiny of the English example in Marxism, from its function here as a paradigm to the function as an exception which Lenin gave it, basing himself on certain of Marx’s own political texts – see Lenin: ‘"Left-Wing” Childishness and Petty-bourgeois Mentality’, Selected Works in Three Volumes, Vol. II, pp. 753-5). Certain of Marx’s texts allow us to go further and say that the analysis is not only in principle independent of the national historical examples it deals with, but also of the extension of the connections that it analyses; it is a study of the properties of every possible economic system which constitutes a market subject to a structure of capitalist production:

We here take no account of export trade, by means of which a nation can change articles of luxury either into means of production or means of subsistence, and vice versa. In order to free the general analysis of all irrelevant subsidiary circumstances, we must treat the commercial world as one nation, and assume that capitalist production is everywhere established and has possessed itself of every branch of industry (Capital, T.III, p. 22n; Vol. I, p. 581n).

The same is true of every mode of production.

In the chapter on the ‘Genesis of Capitalist Ground Rent’ (Vol. III), where he analyses the successive forms of land ownership in different modes of production, Marx could therefore generalize these epistemological suggestions, and write:

This does not prevent the same economic basis – the same form from the standpoint of its main conditions – due to innumerable different empirical circumstances, natural environment, racial relations, external historical influences, etc., from showing infinite variations and gradations in appearance, which can be ascertained only by analysis of the empirically given circumstances (Capital, Vol. III, p. 772).

Like many others, this passage expresses perfectly the theoretical pragmatism which I have been discussing. Reading it literally, one would be perfectly justified in reserving theoretical status for the study of the ‘main conditions’, which coincide with the structure of the mode of production, and saying that the analysis of the empirically given circumstances is itself an empirical analysis.

But what Marx is reflecting here is quite simply the operation I was trying to explain at the beginning, when I said that the first movement of a science of history was to reduce the continuity of history, on which is based the impossibility of sharp ‘breaks’, and to constitute history as a science of discontinuous modes of production, as the science of a variation. He is reflecting this movement by re-establishing continuity as a real reference, a reference to the reality of history, and by making discontinuity a property of the concept in general. Thus the problem of the location of the object whose science is the science of the mode of production is not posed inside the theory itself, which is merely the production of models; this problem is posed on the frontiers of theory, or, more accurately, it makes it obligatory to presuppose that theory has a frontier, which is occupied by a subject of knowledge. ‘Hic Rhodus, hic salta’: theoretical analysis must be abandoned and complemented by ‘empirical’ analysis, i.e., by the designation of the real objects which actually obey the laws expounded. It is then really one and the same problem to collect together the examples which are realizations of the model, despite ‘infinite gradations’, and to designate the transitions from one mode of production to another: to say where the concept of one mode of production is applicable and where it is necessary to apply the concepts of two modes of production in succession. In either case, a residue remains which is given as the irreducibly empirical (in the last analysis, the obviousness of something observed: where its theoretical definition is concerned, on the one hand, the capitalist mode of production is a certain system of relations between labour, means of production, etc., and where its location is concerned, on the other, it is ‘our’ mode of production). But if we refuse to budge, and insist on staying in theoretical discourse, then this residue can be seen as really a lacuna, as something which must be thought and yet cannot be thought with the help of the theoretical concept of the ‘mode of production’ alone.

I have deliberately gone to this extreme conclusion and to the texts which can be used to support it, leaving aside everything in Capital itself which might look like an analysis of the transition from one mode of production to another, i.e., like a solution to the problem of location, namely, an analysis of the formation of the capitalist mode of production and an analysis of its dissolution. I have done so in order to underline straight away that we really do need a second concept at the same theoretical level as that of the mode of production, just as ‘abstract’, if you like, in order to constitute a theory of history as a succession of modes of production. We need it because the concept as we have developed it up to now has precisely left succession in parenthesis. We have only been able to define what a mode of production is by revealing the singularity of its forms, the specific combination that binds together these elements of every combination: labourer, means of production, non-labourers, etc. In order not to pre-judge the issue, let us say that if historical materialism were reduced to this concept alone, it would be unable to think the transition from one combination to another at the same theoretical level.

It follows that we must read all of Marx’s analyses which deal with the formation and dissolution of a mode of production, and look in them for this second attempt, whether it is there explicitly or has to be disengaged But we cannot take these analyses for descriptions pure and simple. However, the fact that Marx let ambiguities survive which allow certain of his terms to have a theory of ‘models’ read into them, is a warning that we shall find more difficulties in this task.

If we return to Capital and try to read in it a theory of the transition from one mode of production to another, we find first of all a concept which seems to be the very concept of historical continuity: the concept of reproduction. The theory of reproduction in fact seems to ensure a triple link or a triple continuity:

– a link between the different economic subjects, in the event, between the; different individual capitals, which really constitute a single ‘inter-twining’ or a single movement. A study of the reproduction of capital is a study of this interlacing and intertwining:

However, the circuits of the individual capitals intertwine, presuppose and necessitate one another, and form, precisely in this interlacing (Verschlingung), the movement of the total social capital (Capital, Vol. II, p. 353).

Therefore, what made it possible to imagine the movement of an individual capital was only an abstraction, and a deforming abstraction, since the movement as a whole is more complex than a mere addition.

– a link between the different levels of the social structure, since reproduction implies the permanence of the non-economic conditions of the production process, notably the legal conditions: in the chapter of Capital on the ‘Genesis of Capitalist Ground Rent’, Marx shows that the institution of a law corresponding to the real relations of production is merely the effect of the repetition of the process of production, of reproduction: see the passage quoted above, Capital, Vol. III, pp. 773-4:

It is in the interest of the ruling section of society to sanction the existing order as law and to legally establish its limits given through usage and tradition. Apart from all else, this, by the way, comes about of itself as soon as the constant reproduction of the basis of the existing order and its fundamental relations assumes a regulated and orderly form in the course of time. And such regulation and order are themselves indispensable elements of any mode of production, if it is to assume social stability and indifference from mere chance and arbitrariness. These are precisely the form of its social stability and therefore its relative freedom from arbitrariness and mere chance. Under stagnant conditions of the production process as well as the corresponding social relations, it achieves this form by mere repetition of its own reproduction. If this has continued on for some time, it entrenches itself as custom and tradition and is finally sanctioned as an explicit law.

– lastly, reproduction ensures the successive continuity of production itself, and this is the basis for all the rest. Production cannot be stopped, and its necessary continuity is inscribed in the identity of the elements as they emerge from one production process and enter another: means of production which have themselves been products, labourers and non-labourers between whom the products and means of production are shared in a certain way. It is the materiality of the elements which supports the continuity, but it is the concept of reproduction which expresses its specific form, because it envelops the different (differential) determinations of the material. Through each of the aspects that I have evoked, the concept expresses merely one and the same pregnancy of the structure which presents a ‘well-bound’ history. At the beginning of her book, The Accumulation of Capital, Rosa Luxemburg writes:

The regular repetition of reproduction is the general sine qua non of regular consumption which in its turn has been the pre-condition of human civilization in every one of its historical forms. The concept of reproduction, viewed in this way, reflects an aspect of the history of civilization (ein kultur-geschichtliches Moment) (trans. Agnes Schwarzschild, London 1951, p. 31).

Thus, the analysis of reproduction seems genuinely to set in motion what has hitherto been seen only in a static form, and to articulate together levels which have hitherto been isolated; reproduction appears to be the general form of permanence of the general conditions of production, which in the last analysis englobe the whole social structure, and therefore it is indeed essential that it should be the form of their change and restructuration, too. That is why I shall dwell on it, for this concept implies more than the previous ones.


In the series of expositions that have the title ‘reproduction’, Marx always prefaced the exposition of the reproduction peculiar to the capitalist mode of production, which is capitalist accumulation (the capitalization of surplus-value) and its peculiar conditions, with a prior exposition of ‘simple reproduction’. Marx calls this simple reproduction an ‘abstraction’, or better, ‘a strange assumption’ (Capital, Vol. II, p. 395). Several explanations of this might be advanced.

It might be thought that this was a matter of an exposition procedure, that ‘simple’ reproduction is only a simplification. At the level of Volume Two (the reproduction schemes), i.e., of the conditions of reproduction which affect the exchanges between the different departments of production, it seems obvious enough why such a simplification should be attempted. It allows the presentation of the general form of these connections in the form of equations, before presenting them in the form of inequalities. The disequilibrium, or disproportion which constitutes the motor of accumulation of the social capital is made intelligible with respect to a simple equilibrium pattern.

It might also be thought that the study of simple reproduction is the study of a particular case, which is partly the same thing, insofar as this particular case is simpler than the general case. But this would not just be an exposition procedure: it would give the knowledge of the movement of reproduction of certain capitals, those which are content to maintain production in certain periods when accumulation temporarily ceases.

Finally, it might be thought that the study of simple reproduction is the study of a part, an always necessary part, of extended reproduction. However much of the surplus-value is capitalized, it is added on over and above an automatic capitalization, which is merely the conservation of the existing capital. The quantity of capitalized surplus-value varies, and it depends on the initiative of the capitalists, in appearance at any rate; simple reproduction cannot be altered, once a capital of a given size is considered, without the. capitalist ceasing to be a capitalist to the precise extent of the decrease. That is why it is important to study simple reproduction in itself (Marx writes ‘As far as accumulation does take place, simple reproduction is always a part of it, and can therefore be studied by itself, and is an actual factor of accumulation’, Capital, Vol. II, p. 395), and only afterwards accumulation or extended reproduction, as a supplement added on to simple reproduction. To be precise, this supplement cannot be added on at will: it has to conform to quantitative conditions which depend on the technical composition of capital; hence it may be intermittent in its actual application. Simple reproduction, on the other hand, is autonomous, continuous and automatic.

None of these explanations are false, nor are they incompatible. But they leave room for a different explanation, one which is more important for us. In Capital, Marx does first present us with the concept of reproduction in the forms of the accumulation of capital, or more accurately, since we want to indicate both ‘simple’ and ‘extended’ reproduction, in the forms of the capitalization of the product, and he first installs us in a quantitative problematic. It is a question of analysing the conditions under which the capitalist or ensemble of capitalists can realize this practical objective: to increase the scale of production, i.e., the scale of exploitation, i.e., the quantity of surplus-value appropriated; which presupposes, in principle at least, the possibility of a practical choice between a simple reproduction and an expansion. But as we know, or are about to discover, this choice is really illusory, a fake, and if we look at the whole of capital, it is a fictive choice. There is no alternative, there are only the real conditions of extended reproduction. Marx tells us that the premiss of simple reproduction is incompatible with capitalist production, ‘although this does not exclude the possibility that in an industrial cycle of ten to eleven years some year may show a smaller total production than the preceding year, so that not even simple reproduction takes place compared to the preceding year’ (Capital, Vol. II, p. 520). Which amounts to saying quite clearly this: the conceptual distinction between simple reproduction and accumulation does not cover the quantitative variations in accumulation, which depend on various circumstances (Marx analyses them) and are the effects of the general law of capitalist accumulation.

Simple reproduction, reproduction on the same scale, appears as an abstraction, inasmuch as on the one hand the absence of all accumulation or reproduction on an extended scale is a strange assumption in capitalist conditions, and on the other hand conditions of production do not remain exactly the same in different years (and this is assumed) ... The value of the annual product may decrease, although the quantity of use-values may remain the same; or the value may remain the same although the quantity of use-values may decrease; or the quantity of value and of the reproduced use-values may decrease simultaneously. All this amounts to reproduction taking place either under more favourable conditions than before or under more difficult ones, which may result in imperfect – defective – reproduction. All this can refer only to the quantitative aspect of the various elements of reproduction, not to the role which they play as reproducing capital or as reproduced revenue in the entire process (Capital, Vol II, pp. 394-5).

When ‘simple’ reproduction such that Iv+s = IIc (which, from the economic point of view, is not the expression of a state of equilibrium anyway, but that of a crisis) occurs during accumulation, this occurrence has precisely only the sense of an occurrence, of a coincidence, i.e., it has no particular theoretical significance. The same is true if we consider the reproduction of an individual capital, which may be extended, simple, or less than simple, and may have a rhythm higher than, equal to or lower than that of the social capital as a whole, etc. These variations make no conceptual difference, in exactly the same way and for the same reason, that variations in the prices of commodities never make them anything but prices: it may be that a commodity is actually sold ‘at its value’ without this being any more than a coincidence. Moreover, it is a coincidence that cannot be registered in a general rule, i.e., cannot be measured: only prices are assessed in the exchange of commodities, not values. In both cases, Marx presents an important conceptual distinction between two levels of the structure, or, better, between the structure and its effects, in the mild form of a ‘provisional assumption’, to be lifted later (‘the prices of commodities coincided with their values’, ‘the conditions of reproduction remain the same’). The assumption ofinvariant conditions’ is not an analysis of the effects, but of the conditions themselves.

We are thus led to look for another explanation for this duplication of the analysis of reproduction, and we find it in a series of indications of Marx’s such as the following:

This illustration of fixed capital, on the basis of an unchanged scale of reproduction, is striking. A disproportion of the production of fixed and circulating capital is one of the favourite arguments of the economists in explaining crises. That such a disproportion can and must arise even when the fixed capital is merely preserved, that it can and must do so on the assumption of ideal normal production on the basis of simple reproduction of the already functioning social capital is something new to them (bei Voraussetzung einer idealen Normalproduktion) (Capital, Vol. II, p. 469).

This ideal ‘normal’ production is obviously production in its concept, production as Marx studies it in Capital, telling us to make it as the ‘norm’ or the ‘ideal average’. Before it is a simplification of the exposition or the study of one particular case, one which we have just seen to be without theoretical significance, even before it makes possible a quantitative analysis of capitalized value and of the origin of its different parts, ‘simple reproduction’ is therefore the analysis of the general formal conditions of all reproduction. And even before it is an exposition of the general forms of the connections between the different departments of production, in the mathematical sense of the term, it is an exposition of the ‘form’ of the reproduction process in the sense in which we have already analysed the ‘capitalist form’ of a mode of production.

This is indeed the sense of the first exposition of ‘simple reproduction’ (Capital, Vol. I, Chapter XXIII). Marx starts from the definition of reproduction as a simple repetition of the immediate production process in the way we have just analysed it, and he writes:

The production process periodically begins again and always passes through the same phases in a given time, but is always repeated on the old scale. Nevertheless, this repetition, or continuity, gives certain new characteristics to the process, or, rather (oder vielmehr,) causes the disappearance of some apparent characteristics which it possessed as an isolated act (die Scheincharaktere seines nur vereinzelten Vorgangs) (Capital, T.III, p. 10; Vol. I, p. 567).

The essential aspect of simple reproduction is not therefore that all surplus-value is unproductively consumed instead of being partially capitalized; it is this uncovering of the essence by the removal of illusions, this virtue of repetition which retrospectively illuminates the nature of the ‘first’ production process (in the manuscript Pre-Capitalist Economic Formations, Marx also writes: ‘the true nature of capital does not appear until the end of the second cycle’).

However, the point of view of repetition itself implies the possibility of an illusion which might conceal the orientation of Marx’s reflection on this point. This would be to wish to follow capital in its successive ‘acts’, to wish to understand what happens when, after a ‘first’ production cycle, capital undertakes to pursue a ‘second’ cycle. In this way, instead of arising as the knowledge of the determinations of the production process itself, reproduction appears as a continuation of production, as a supplement to the analysis of production. Thus the analysis of capital seems to follow in the tracks of the destiny of an object which is capital: at the moment of reproduction, this capital meets others on the market, its freedom of movement is suppressed (it cannot grow in arbitrary proportions because it is in competition with other capitals), and it seems that the movement of social capital is not the sum of the movements of the individual capitals, but a complex movement of its own which has been called an ‘intertwining’. For example, this is the path urged on us at the beginning of Rosa Luxemburg’s Accumulation of Capital, which starts from a literal reading of Marx (‘The literal meaning of the word “reproduction” is repetition ...’) and asks what new conditions reproduction implies with respect to production. The passage of Marx’s which I have quoted shows us that, on the contrary, it is a matter of the same conditions, initially implicit (transposed and deformed in the eyes of the agents of production into ‘apparent characteristics’; and presented in Marx’s exposition of the ‘immediate’ production process in the forms of admitted ‘assumptions’ or ‘presuppositions’).

In reality, it is a matter of a more complex operation than a mere repetition. In Marx’s text, simple reproduction is from the beginning identified with the consideration of the ensemble of social production. The movement that destroys this appearance which arises from the study of the immediate production process and is also what the capitalist and the worker ‘imagine’ (Capital, T.III, p. 13; Vol. I, p. 569: ‘die Vorstellung des Kapitalisten ‘) is at once a ‘repetition’ and the transition to capital as a totality:

The matter takes a quite different aspect, when we contemplate, not the single capitalist, and the single labourer, but the capitalist class and the labouring class, not isolated acts of production, but capitalist production in its full continuous renewal, and on its social scale (Capital, T. III, pp. 14-15; Vol. I, p. 572).

The analysis of Volume Two will show clearly in detail how the analysis of the repetition (of the succession of cycles of production) and that of capital as a form of the ensemble of production are inter-dependent. But this unity is already present here. The ‘isolated act of production’ is twice characterized negatively as something which is not repeated and as something which is done by an individual. Or rather, ‘isolated act’ is a way of saying the same thing twice. Once the isolation has been suppressed, we are no longer dealing with an act, i.e., with a subject, an intentional structure of means and ends, if it is true, as Marx says in the 1857 Introduction, that ‘to treat society as a single subject is to treat it from a false position – speculatively’ (Grundrisse, p. 15). There can therefore be no question in this analysis of following the reproduction process, of attempting effectively – and fictively – to ‘re-new’ the production process.

This analytical operation is in principle the one which the 1857 Introduction to A Contribution to the Critique of Political Economy installed in parallel to the comparative analysis of the modes of production. It is now no longer a question of identifying the variants of the ‘combination’ of the ‘relations of production’ and the ‘productive forces’ on the basis of historical material, but of examining what Marx calls ‘the general determination of production at a given social stage’, i.e., the relation between the totality of social production and its particular forms (branches) in a given synchrony (as this term has been illuminated for us from now on, since the analysis of the ‘repetition’ of production, of the continuity of production in a series of cycles, depends on the analysis of production as a whole, of production as a totality – Totalität). For there is totalization only in the actuality of the social division of labour at a given moment, not in the individual adventures of capitals. This is expressed by Marx when he says that the analysis of reproduction envisages social production exclusively in its result (‘If we study the annual function of social capital ... in its results’, Capital, Vol. II, p. 392 – modified). As we know, this result is production as a whole and its division into different departments: the operation that reveals it is not therefore a section through the movement of the different branches of production, of the different capitals, at a moment chosen with reference to a common external time, and hence dependent both in principle and in actual realization on this movement; it is an operation in which the peculiar movement of the capitals, the movement of production in each of its divisions, is completely set aside, suppressed, without any kind of conservation. Marx bases his whole analysis of reproduction from the first very general exposition of simple reproduction (Volume One) to the system of reproduction schemes (Volume Two) on this transformation of succession into synchrony, into ‘simultaneity’ (in his own term: Gleichzeitigkeit). Paradoxically, the continuity of the movement of production finds its concept in the analysis of a system of synchronic dependencies: the succession of the cycles of individual capitals and their intertwining depend on it. In this ‘result’, the movement which has produced it is necessarily forgotten, the origin isobliterated’ (die Herkunft ist aufgelöscht) (Capital, Vol. II, p. 110).

To move from the isolated act, from the immediate production process, to the repetition, to the ensemble of social capital, to the result of the production process, is to install oneself in a fictive contemporaneity of all the movements, or, to put it more accurately, applying one of Marx’s theoretical metaphors, in a fictive planar space, in which all the movements have been suppressed, in which all the moments of the production process appear in projection side by side with their connections of dependence. It is the movement of this transition that Marx describes for the first time in the chapter of Volume One on ‘Simple Reproduction’.


We can list the ‘appearances’ (Scheincharaktere) which are dissipated in this operation as follows:

First the appearance of the separation and the relative independence of the different ‘moments’ of production in general: the separation of production in the strict sense from circulation, of production from individual consumption, of the production and distribution of the means of production from that of the means of consumption. If we consider an ‘isolated act’ of production or even a plurality of such ‘acts’, all these moments seem to belong to a different sphere from that of production (’sphere’ is a word which Marx very often uses). Circulation belongs to the market on which commodities are presented after ‘leaving’ production, without any certainty that they will actually be sold; individual consumption is a private act which takes place outside the sphere of circulation itself:

The labourer’s productive consumption, and his individual consumption, are therefore totally distinct. In the former, he acts as the motive power of capital, and belongs to the capitalist. In the latter, he belongs to himself, and performs his vital functions outside the process of production. The result of the one is, that capital lives; of the other, that the labourer himself lives (Capital, T.III, p. 14; Vol. I, p. 571).

The distribution of the means of production and consumption appears either as the contingent origin of production, or as revenue (and then it passes into the sphere of consumption).

The introductory act (der einleitende Akt), which constitutes an act of circulation – the purchase and sale of labour-power – itself rests on a distribution of the elements of production which preceded and presupposed the distribution of the social products, namely on the separation of labour-power as a commodity of the labourer from the means of production as the property of non-labourers (Capital, Vol. II, p. 385).

The analysis of reproduction shows that these moments have no relative autonomy or laws of their own, but are determined by those of production. If we consider the ensemble of social capital in its result, the sphere of circulation disappears as a ‘sphere’, since all exchanges are predetermined in the division of the departments of production and in the material nature of their production. The individual consumption of the worker and capitalist, too, is predetermined in the nature and quantity of the means of consumption produced by the total social capital: while one portion of the annual product is ‘destined for productive consumption from the very first’ (T.III, p. 9; Vol. I, p. 566), another is destined from the very first (von Haus aus) for individual consumption. The limits within which individual consumption can oscillate depend on the internal composition of capital and are fixed at each moment.

The individual consumption of the labourer, whether it proceeds within the workshop or outside it, forms therefore an element (Moment) of the reproduction of capital; just as cleaning machinery does, whether it be done during the labour process or at moments when it is interrupted (Capital, T.III, p. 15; Vol. I, p. 572).

Lastly, the distribution of the means of production and consumption, or the division of the different elements, ceases to appear as a contingent factual state: once he has consumed the equivalent of his wages, the worker leaves the production process as he entered it, stripped of property, and the capitalist as he entered it, owner of the products of labour, which include new means of production. Production continually determines the same distribution.

Thus it appears that the capitalist mode of production determines the modes of circulation, consumption and distribution. More generally, the analysis of reproduction shows that every mode of production determines modes of circulation, distribution and consumption as so many moments of its unity.

Further, the analysis of reproduction destroys the appearance involved in the ‘beginning’ of the production process, the appearance of a ‘free’ contract between the worker and the capitalist, which is renewed on each occasion, the appearance which makes variable capital an ‘advance’ from the capitalist to the labourer (on account of the product, i.e., of the ‘end’ of the production process). In a word, all the appearances which seem to reduce to chance the face to face meeting of the capitalist and the worker as buyer and seller of labour power. Reproduction reveals the ‘invisible threads’ which chain the wage-earner to the capitalist class.

The capitalist production process reproduces ... the conditions which force the labourer to sell himself in order to live, and enable the capitalist to purchase him in order that he may enrich himself. It is no longer a mere accident, that capitalist and labourer confront each other in the market as buyer and seller. It is the dilemma (Zwickmühle – ‘double mill’) of the process itself that incessantly hurls back the labourer onto the market as a vendor of his labour-power, and that incessantly converts his product into a means by which another man can purchase him. In reality, the labourer belongs to the capitalist class before he has sold himself to an individual capitalist (Capital, T.III, pp. 19-20; Vol. I, p. 577).

Simultaneously, reproduction destroys the appearance according to which capitalist production merely applies the laws of commodity production, i.e., of the exchange of equivalents. Each sale-purchase of labour-power is a transaction of that form, but the general movement of capitalist production appears as the movement by which the capitalist class continually appropriates a portion of the product created by the working class without giving any equivalent for it. This movement no longer has any beginning or termination (a division duplicated and designated by the legal structure of the contract, precisely a terminal contract or ‘contrat à terme ‘), i.e., there is no longer any isolated structure in which the elements of production meet. In the concept of the elements of production provided by the analysis of reproduction, they no longer need to meet because they are always already together.

Thus simple reproduction destroys in the production process even the appearance of an isolated act: an act whose agents were individuals, transforming things under determinate conditions which eventually obliged them to make these things into commodities and surplus-value for the capitalist. In this appearance the individuals retained their identity, just as capital seemed to be a sum of value which was conserved throughout the succession of acts of production.[20]

And, conversely, these material elements, in the specificity of their material nature, and in the differential distribution of these natural properties through all the branches of production and all the capitals of which they are composed, now express the conditions of the process of social reproduction. Thus reproduction reveals that things are transformed in the hands of the agents of production without their being aware of it, without it being possible for them to be aware of it if the production process is taken for the acts of individuals. Similarly, these individuals change and they really are only class representatives. But these classes are obviously not sums of individuals, which would not change anything: it is impossible to make a class by adding individuals together on whatever scale. Classes are functions of the process of production as a whole. They are not its subjects, on the contrary, they are determined by its form.

Precisely in these chapters of Volume One on reproduction, we find all the images which Marx uses to help us grasp the mode of existence of the agents of the production process as the supports (Träger) of the structure.

On the stage of reproduction, where things ‘come to light’ (Capital, T.III, p. 26; Vol. I, p. 586) and look quite different (ganz anders aussehen), the individuals quite literally come forward masked (‘it is only because his money constantly functions as capital that the economic guise of a capitalist – die ökonomische Charaktermaske des Kapitalisten – attaches to a man’, Capital, T.III, p. 9; Vol. I, p. 566): they are nothing more than masks.

These are therefore the analyses in which Marx shows us the movement of transition (but this transition is a rupture, a radical innovation) from a concept of production as an act, the objectivation of one or more subjects, to a concept of production without a subject, which in return determines certain classes as its peculiar functions. This movement, in which Marx pays retrospective homage to Quesnay (for whom ‘the innumerable individual . acts of circulation are at once brought together in their characteristic social mass movement – the circulation between great functionally determined economic classes of society’, Capital, Vol. II, p. 359), is carried out in exemplary fashion with respect to the capitalist mode of production, but in principle it is valid for every mode of production. As opposed to the movement of reduction and then constitution which characterizes the transcendental tradition of classical philosophy, it directly achieves an extension which excludes any possibility of production being the acts of any subjects. their practical cogito, It embraces the possibility, which I can only suggest here, of formulating a new philosophical concept of production in general.

All the preceding can be summarized by saying that, in a single movement, reproduction replaces and transforms the things, but retains the relations indefinitely. These relations are obviously what Marx calls ‘social relations’; the relations which are drawn, ‘projected’, in the fictive space which I have mentioned,[21] the term itself is Marx’s own:

This natural faculty of labour (to conserve old value, while it creates new) takes the appearance of the self-sustaining faculty of the capital, in which it is incorporated, just as the social productive forces take the appearance of a property of capital, and as the constant appropriation of surplus-labour by the capitalist takes that of a constant self-expansion of capital. All faculties of labour are projected (projektieren sich) as faculties of capital, just as all forms of commodity value are projected as forms of money (Capital, T.III, p. 47; Vol. I, pp. 606-7).

The relations thus revealed mutually imply one another: notably the relations of property and the relations of real appropriation (‘productive forces’) in their complex unity. They comprehend the hitherto disjointed ‘moments’ (production, circulation, distribution, consumption) in a necessary and complete unity. And at the same time, they comprehend everything which appeared during the analysis of the immediate production process as its ‘presuppositions’, as the necessary ‘conditions’ for the process to be able to proceed in the form described: e.g., in capitalist production, the autonomy of the economic instance or of the legal forms corresponding to the forms of commodity exchange, i.e., a certain form of correspondence between the various instances of the social structure. This is what might be called the ‘consistency’ of the structure as it appears in the analysis of reproduction. It might also be said that for Marx the conceptual pair production/ reproduction contains the definition of the structure involved in the analysis of a mode of production.

On the plane instituted by the analysis of reproduction, production is not the production of things, it is the production and conservation of social relations. At the end of the chapter on simple reproduction, Marx writes:

The capitalist production process, therefore, considered in its inter-connection (Zusammenhang) or as reproduction, produces not only commodities, not only surplus-value, but it also produces and eternalizes the social relation between the capitalist and the wage-earner (Capital, T.III, p. 20; Vol. I, p. 578).

This formulation is repeated at the end of the whole work, just as Marx is locating the relationship of the classes to the different forms of

On the other hand, if the capitalist mode of production presupposes this definite social form of the conditions of production, so does it reproduce it continually. It produces not merely the material products, but reproduces continually the production relations in which the former are produced, and thereby also the corresponding distribution relations (Capital, Vol. III, p. 857).

The same goes for any mode of production. Each mode of production continually reproduces the social relations of production presupposed its functioning. In the manuscript Pre-Capitalist Economic Formations, Marx had already expressed this by assigning the production and reproduction of the social relations to the corresponding production as its unity result (instead of a ‘not merely ...’):

Property – and this applies to its Asiatic, Slavonic, Antique and Germanic forms – therefore originally signifies a relation of the working (producing) subject (or self-reproducing subject) to the conditions of his production or reproduction as his own. Hence, according to the conditions of production, property will take different forms. The aim of production itself is to reproduce the producer in and together with these objective conditions of his existence (Grundrisse, p. 395; PCEF, p. 95).

What is the meaning of this double ‘production’?

Let us note first of all that it provides us with a key to a number of formulations of Marx’s which have been regarded, precipitately perhaps, as fundamental theses of historical materialism. For the lack of a complete definition of the terms which they contain, they have lent authority to a number of rather divergent readings. For example, the formulations in the Preface to A Contribution which I discussed at the outset: ‘In the social production their life, men enter into definite relations that are indispensable and independent of their will ... therefore mankind always sets itself only such tasks as it can solve ‘; or the formulations in Engels’s letter to Bloch: ‘We make our history ourselves, but, in the first place, under very definite assumptions and conditions ‘. The whole philosophical interpretation of historical materialism is indeed at stake here: if we take this double ‘production’ literally, i.e. if we think that the objects transformed and the social relations which they support are modified or conserved by the production process in the same way, if, for example, we group them within a single concept ofpractice ‘, we are giving a rigorous foundation to the idea that ‘men make history’. Only on the basis of such a unique, unified concept of practice-production can this formulation have any theoretical meaning, can it be an immediately theoretical thesis (and not simply a moment in the ideological struggle against a mechanistic-materialist determinism). But this concept really belongs to an anthropological conception of production and practice, centred precisely on those ‘men’, who are the ‘concrete individuals’ (notably in the form of the masses) who produce, reproduce and transform the conditions of their former production. In respect to this activity, the constraining necessity of the relations of production only appears as a form which the object of their practice already possesses and which restricts the possibilities of creating a new form. The necessity of the social relations is simply the work of the former production activity, which necessarily leaves to the succeeding one determinate conditions of production.

But our analysis of reproduction has shown us that this double ‘production’ must be taken in two different senses: to take the unity of the expression literally is precisely to reproduce the appearance which makes the production process an isolated act enclosed in the determinations of the preceding and the succeeding. An isolated act, insofar as its only connections with the other acts of production are supported by the structure of linear temporal continuity in which there can be no interruption (whereas in the conceptual analysis of reproduction, these connections are, as we have seen, supported by the structure of a space). Only the ‘production of things’ can be thought as an activity of this kind – it already almost contains the concept of it in its determination of the ‘raw’ material and the ‘finished’ product; but the ‘production of the social relations’ is far rather a production of things and individuals by the social relations, a production in which the individuals are determined so as to produce and the things so as to be produced in a specific form by the social relations. That is, it is a determination of the functions of the social process of production, a process without a subject. These functions are no more men than, on the plane of reproduction, the products are things. Therefore (re)production, i.e., social production in its concept, does not strictly speaking produce the social relations, since it is only possible on condition that these social relations exist; but on the other hand, neither does it produce commodities in the sense of producing things which subsequently receive a certain social qualification from the system of economic relations which invests them, objects which subsequently ‘enter into relations’ with other things and men; production only produces (ever already) qualified things, indices of relations.

Marx’s formulation (‘the process of production does not only produce material objects but also social relations’) is not therefore a conjunction but a disjunction: either it is a matter of the production of things, or else it is a matter of the (re)production of the social relations of production. There are two concepts, the concept of the ‘appearance’ and the concept of the effectivity of the structure of the mode of production. As opposed to the production of things, the production of social relations is not subject to the determinations of the preceding and the succeeding, of the ‘first’ and the ‘second’. Marx writes that ‘every process of social production is at the same time a process of reproduction. The conditions of production are also those of reproduction’; and at the same time they are the conditions which reproduction reproduces: in this sense the ‘first’ process of production (in a determinate form) is always-already a process of reproduction. There is no ‘first’ process of production for production in its concept. All the definitions concerning the production of things must therefore be transformed: in the production of the social relations, what appeared as the conditions of the first production really determines identically all the other productions.

This transaction, which pertains to circulation – the sale and purchase of labour power – not only inaugurates the process of production, but also determines implicitly its specific character (Capital, Vol. II, p. 385).

The concept of reproduction is thus not only the concept of the ‘consistency’ of the structure, but also the concept of the necessary determination of the movement of production by the permanence of that structure; it is the concept of the permanence of the initial elements in the very functioning of is the system, hence the concept of the necessary conditions of production, conditions which are precisely not created by it. This is what Marx calls the eternity of the mode of production:

This incessant reproduction or eternalization (Verewigung) of the labourer, is the sine qua non of capitalist production (Capital, Vol. I, p. 571; retranslated from the German text).

Chapter 4. Elements for a Theory of Transition

Let us return to the question posed above: the question of the transition from one mode of production to another. The analysis of reproduction seems merely to have erected a number of obstacles to its theoretical solution. Really, it enables us to pose the problem in its true terms, for it subjects the theory of transition to two conditions.

First, all social production is a re-production, i.e., a production of social relations in the sense suggested. All social production is subject to structural social relations. The ‘transition’ from one mode of production to another can therefore never appear in our understanding as an irrational hiatus between two ‘periods’ which are subject to the functioning of a structure, i.e., which have their specified concept. The transition cannot be a moment of destructuration, however brief. It is itself a movement subject to a structure which has to be discovered. We can give a strong sense to these comments of Marx’s (reproduction expresses the continuity of production because it can never stop) which he often presented as ‘obvious’, as things ‘every child knows’ (that the labourer can never have lived on ‘the air of time’, that ‘a nation which ceased to work, I will not say for a year, but even for a few weeks, would perish’ – Letter to Kugelmann, 11 July 1868). They mean that the invariant structure of production can never disappear, although it may take a particular form in each mode of production (the existence of a fund for the maintenance of labour, i.e., the distinction between necessary labour and surplus-labour; the division of the product into means of production and means of consumption, a distinction that Marx calls original, or again the expression of a natural law, etc.). They therefore mean that the forms of transition themselves are particular ‘forms of manifestation’ (Erscheinungsformen) of this general structure: they are therefore themselves modes of production. They therefore imply the same conditions as every mode of production, and notably a certain form of complexity of the relations of production, of correspondence between the different levels of social practice (I shall try to suggest what form). The analysis of reproduction shows that if we can formulate the concept of the modes of production which belong to periods of transition between two modes of production, at the same stroke the modes of production are no longer suspended in an indeterminate time (or site): the problem of their location has been resolved once we can explain theoretically how they follow one another, i.e., once we know the moments of their succession in their concepts.

But on the other hand (second consequence), the transition from one mode of production to another, e.g., from capitalism to socialism, cannot consist of the transformation of the structure by its functioning itself, i.e., of any transition of quantity into quality. This conclusion follows from what I have said about the double sense in which the term ‘production’ has to be understood in the analysis of reproduction (the production of things, and the ‘production’ of social relations). To say that the structure can be transformed in its functioning itself is to identify two movements which manifestly cannot be analysed in the same way with respect to it: on the one hand, the very functioning of the structure which, in the capitalist mode of production, takes the particular form of the law of accumulation; this movement is subject to the structure, it is only possible on condition that the latter is permanent ; in the capitalist mode of production it coincides with the ‘eternal’ reproduction of capitalist social relations. On the contrary, the movement of dissolution is not subject in its concept to the same ‘presuppositions’, it is apparently a movement of a completely different kind, since it takes the structure as the object of transformation. This conceptual difference shows us that where a ‘dialectical logic’ would quickly solve the problem, Marx holds firmly to non-dialectical logical principles (obviously, non-Hegelian-dialectical principles): what we have recognized as distinct in essence shall not become a single process. And more generally, the concept of the transition (from one mode of production to another) can never be the transition of the concept (to one other-than-itself by internal differentiation).

And yet we do have a text where Marx presents the transformation of the relations of production as a dialectical process of the negation of the negation. This is the passage on the ‘Historical Tendency of Capitalist Accumulation’ (Capital, Vol. I, Chapter XXXII). It groups into a single schema those of Marx’s analyses which deal with the origin of the capitalist mode of production (‘primitive accumulation’), those which deal with its peculiar movement of accumulation, and those which deal with its end, which Marx here calls its ‘tendency’, using this term in the way he does in Volume Three. I shall be obliged to take each of these three moments separately, according to the aggregate of the analyses that Marx devotes to them in Capital. But first of all, I should like to demonstrate the remarkable form of this passage, which already determines certain conclusions.

In principle, Marx’s reasoning in this text implies that the two transitions are of the same nature. First transition: from the individual private ownership of the means of production, based on personal labour (‘the pygmy property of the many’) to capitalist private ownership of the means of production, based on the exploitation of the labour of others (‘the huge property of the few’). First transition, first expropriation. Second transition: from capitalist ownership to individual ownership, based on the acquisitions of the capitalist era, on co-operation and the common possession of all the means of production, including the land. Second transition, second expropriation.

These two successive negations are of the same form, which implies that all the analyses Marx devoted to primitive accumulation on the one hand (origin), to the tendency of the capitalist mode of production on the other, i.e., to its historical future, are similar in principle. But as we shall see, in Capital these analyses in fact present a remarkable disparity: the analysis of primitive accumulation seems to be relatively independent of the analysis of the mode of production strictly speaking, or even to be an enclave of ‘descriptive’ history in a work of economic theory (on this opposition I refer the reader to the preceding paper by Louis Althusser); on the contrary, the analysis of the historical tendency of the capitalist mode of production seems to be one moment of the analysis of the capitalist mode of production, a development of the intrinsic effects of the structure. It is this last analysis which suggests that the (capitalist) mode of production is transformed ‘by itself’, through the play of its own peculiar ‘contradiction’, i.e., through its structure.

In the passage on the ‘Historical Tendency of the Capitalist Mode of Production’, the two transformations are reduced to the second type, which is all the more surprising in that the text constitutes the conclusion to the analysis of the forms of primitive accumulation. The capitalist mode of production, too, appears in these formulations to be the result of the spontaneous evolution of the structure:

This industrial regime of small independent producers ... engenders by itself the material agents for its dissolution which are contained in its peculiar contradiction (it prevents the advance of production) (Capital, T.III, pp. 203-4; Vol. I, pp. 761-2).

The second movement, ‘This expropriation is accomplished by the action of the immanent laws of capitalist production itself, which lead to the concentration of capitals ... The socialization of labour and the concentration of the means of production at last reach a point where they become incompatible with their capitalist integument (Hülle) ... Capitalist production begets its own negation with the fatality that presides over the metamorphoses of nature’ (T.III, pp. 204-5; Vol. I, p. 763).

Thus, while summing up the analyses that Marx devoted to the formation and dissolution of the capitalist mode of production, these formulations claim to give the very concept of the transition that we are looking for. They must therefore be compared with these analyses themselves. But the apparent disparity between these analyses must not be allowed to prevail over the unity postulated by the text on the ‘Historical Tendency’ via the forms of the ‘negation of the negation’: on the contrary, it must be reduced if it is to be possible to formulate the concept of the transition. (Obviously, there can be no question of maintaining that all transitions from one mode of production to another have the same concept: the concept is specified: each time, like that of the mode of production itself. But just as all historical modes of production have appeared as forms of a combination of the same nature, historical transitions must have concepts of the same theoretical nature. This is what is strictly implied by the preceding quotations, even if they do go on to suggest that this nature is that of an external dialectical supersession.) Let us look at these ‘transitions’ again, one by one.


The chapters which Marx devotes to ‘so-called primitive accumulation’ (die sogennante ursprüngliche Akkumulation) are presented as the solution to a problem which arose in the study of reproduction (capitalist accumulation) and which was provisionally left on one side. The movement of accumulation of capital is only possible because a surplus-value susceptible to capitalization exists. This surplus-value itself can only be the result of a previous production process, and so on, apparently indefinitely. But in given technical conditions the minimum sum of value capable of functioning as capital and its division into constant and variable capital are also given, and condition every extraction of surplus-value. The production of this original capital therefore constitutes a threshold and crossing this threshold cannot be explained by the action of the law of capitalist accumulation alone.

But it is not really just a question of measuring a sum of value. The movement of reproduction is not only continually the origin of a capitalizable surplus-value, it implies the permanence of capitalist social relations, and it is only possible on condition that they exist. The question of primitive accumulation therefore simultaneously involves the formation of capitalist social relations.

What characterizes the myth of primitive accumulation in classical economics is the retrospective projection of the forms of capitalist production, and of the forms of exchange and law which correspond to it: by pretending that the original minimum capital was saved by the future capitalist out of the product of his labour before being advanced in the form of wages and means of production, classical economics gave some retroactive validity to the laws of exchange between equivalents and of the ownership of the product based on the legitimate disposal of the set of factors of production. This retrospective projection does not lie in the distinction between a necessary labour and a surplus-labour, and hence between a wage and a profit, with respect to a hypothetical individual production (for these distinctions can serve conventionally to distinguish between various portions of the product even in non-capitalist modes of production, even in modes of production without exploitation where these portions do not constitute the revenues of different classes: Marx himself uses this convention, for example, in the chapter on the ‘Genesis of Capitalist Ground Rent’ in Volume Three); the retrospective projection lies precisely in the idea that the formation of capital and its development are part of a single movement subject to common general laws. The basis for the bourgeois myth of primitive accumulation is therefore, in a reading of absolute reversibility, the formation of capital by the movement of an already potentially capitalist private production, and the self-generation of capital. But it would be even more accurate to say that the entire movement of capital (the movement of accumulation) thus appears as a memory: the memory of an initial period in which, by his personal labour and saving, the capitalist acquired the possibility of indefinitely appropriating the product of others’ surplus-labour. This memory is inscribed in the form of the bourgeois rights of property which base the appropriation of the product of labour indefinitely on the previous ownership of the means of production:

At first the rights of property seemed to us to be based on a man’s own labour. At least, some such assumption was necessary since only commodity-owners with equal rights confronted each other, and the sole means by which a man could become possessed of the commodities of others, was by alienating his own commodities; and these could be replaced by labour alone. Now, however, property turns out to be the right, on the part of the capitalist, to appropriate the unpaid labour of others or its product, and to be the impossibility, on the part of the labourer, of appropriating his own product. The separation of property from labour has become the necessary consequence of a law that apparently originated in their identity (Capital, Vol. I, pp. 583-4).

If we adopt the view-point of classical economics we must retain both faces of this ‘law of appropriation’ at once, the universally equal commodity right (and the hypothetical personal labour which it presupposes and induces through its own consistency) on the one hand, and on the other the exchange without equivalence which is an expression of the essence of the process of capitalist accumulation. It is in the constantly present space of these two forms that the memory of the mode of production is inscribed, the continuing present of an origin homogeneous with the current process.

As we know, this is a myth. Marx sets himself the task of proving that, historically, things did not happen like that. At the same stroke, what he calls the ‘apologetic’ function of the myth is exposed, expressed in the perenneal nature of the economic categories of capitalism. I shall presume that the reader has this study in mind and draw attention to its very remarkable form.

Both a history and a pre-history are involved in the study of ‘primitive accumulation’ (the name has been retained, but it now designates a quite different process). A history: we have discovered that the bourgeois theory of primitive capital is no more than a myth, a retrospective construction, very precisely the projection of a current structure which is expressed in the ‘law of appropriation’ and depends on the capitalist structure of production. It has therefore become clear that the ‘memory’ inscribed in this law of appropriation is a purely fictive one: it expresses a current situation in the form of a past whereas this situation’s real past had another form, a completely different one, demanding an analysis. The study of primitive accumulation is this replacement of memory by history. A pre-history: this study reveals to us a different world at the origin of capital. Knowledge of the laws of the development of capitalism is useless to us here because this is a completely different process, not subject to the same conditions. Thus a complete rupture appears, a rupture reflected in theory, between the history of the formation of capital (of capitalist social relations) and the history of capital itself. Thus the real history of the origins of capitalism is not just different from the myth of origins; by the same token it is different in its conditions and principles of explanation from what has appeared to us to be the history of capital; it is a pre-history, i.e., a history of a different age.

But in their turn, these determinations are in no sense vague or mysterious to us, for we know that a different age is precisely a different mode of production. Let us call it the feudal mode of production, following Marx’s historical analysis, but without asserting any law of necessary and unique succession of these modes of production, an assertion which nothing in the concept of a ‘mode of production’ allows us to make immediately, if the nature of the latter really is that of a varied combination. We see that to recognize the history of the origins of capital as a real pre-history is at the same time to pose the problem of the relationship between this pre-history and the history of the feudal mode of production, which, just like the history of the capitalist mode of production, can be known by the concept of its structure. In other words, we must ask ourselves whether this pre-history is identical with the history of the feudal mode of production, simply dependent on it or distinct from it. The set of conditions for this problem is summed up by Marx as follows:

The capitalist system is based on the radical separation of the producer from the means of production. As soon as capitalist production is once established, it reproduces this separation on a continually extending scale; but as the latter is the basis of the former, it could not have been established without it. In order that the capitalist system should come into existence it is therefore necessary that the means of production have already, at least in part, been seized absolutely from the producers who had been using them to realize their own labour, and that they are already held by commodity producers who use them to speculate on the labour of others. ‘primitive’ accumulation, therefore, is nothing else than the historical movement which divorces labour from its external conditions, and it is called ‘primitive’ because it forms the prehistoric stage of the bourgeois world. The capitalist economic order emerged from the entrails of the feudal economic order. The dissolution of the latter set free the constitutive elements of the former (Capital, T.III, pp. 154-5; Vol. I, pp. 714-15).

Marx returned to this problem several times, using the same method on each occasion, and the texts in which he did so should be assembled for an analysis of their content: in Capital, besides Part 8 of Volume One (‘The So-called Primitive Accumulation’), the chapters in Volume Three devoted to the ‘Historical Facts about Merchant’s Capital’, to ‘Pre-Capitalist Relationships’, and to the ‘Genesis of Capitalist Ground Rent’. We shall find that this dispersion is not accidental. Marx himself calls Part 8, on so-called primitive accumulation, a ‘sketch’ (T. III, p. 156; Vol. I, p. 716, but we have various preparatory manuscripts on the same subject to which to refer, above all the already cited text on Pre-Capitalist Economic Formations.

All these studies have a common retrospective form, but in a sense which we have to specify, since we have just been criticizing the form of retrospective projection of the bourgeois myth of primitive accumulation. It is very clear from the preceding text that the study of primitive accumulation takes as its guiding thread precisely the elements which were distinguished by the analysis of the capitalist structure: these elements are grouped together here under the heading of the ‘radical separation of the labourer from the means of production’. The analysis is therefore retrospective, not insofar as it projects backwards the capitalist structure itself, presupposing precisely what had to be explained, but insofar as it depends on knowledge of the result of the movement. On this condition it escapes empiricism, the listing of the events which merely precede the development of capitalism: it escapes vulgar description by starting from the connections essential to a structure, but this structure is the ‘current’ structure (I mean that of the capitalist system insofar as it has currently come into its own). The analysis of primitive accumulation is therefore, strictly speaking, merely the genealogy of the elements which constitute the structure of the capitalist mode of production. This movement is particularly clear in the construction of the text on Pre-Capitalist Economic Formations, which depends on the action of two concepts: that of the presuppositions (Voraussetzungen) of the capitalist mode of production, thought on the basis of its structure, and that of the historical conditions (historische Bedingungen) in which these presuppositions happen to be fulfilled. The outline history of the different modes of production in this text, rather than being a true history of their succession and transformation, is a historical survey (sondage) of the routes by which the separation of the labourer from his means of production and the constitution of capital as a sum of disposable value were achieved.

For this reason, the analysis of primitive accumulation is a fragmentary analysis: the genealogy is not traced on the basis of a global result, but distributively, element by element. And notably, it envisages separately the formation of the two main elements which enter into the capitalist structure: the ‘free’ labourer (the history of the separation of the producer from the means of production) and capital (the history of usury, of merchant capital, etc.). In these conditions, the analysis of primitive accumulation does not and never can coincide with the history of the previous mode or modes of production as known from their structures. The indissoluble unity possessed by the two elements in the capitalist structure is suppressed in the analysis, and it is not replaced by a comparable unity in the previous mode of production. That is why Marx writes: ‘The capitalist economic order emerged from the entrails of the feudal economic order. The dissolution of the latter set free the constitutive elements of the former.’ The dissolution of the latter, i.e., the necessary evolution of its structure, is not identical to the constitution of the former in its concept: instead of being thought at the level of the structures, the transition is thought at the level of the elements. This form explains why we are not dealing with a true history in the theoretical sense (since, as we know, such a history can only be produced by thinking the dependence of the elements with respect to a structure), but it is also the condition on which we can discover a very important fact: the relative independence of the formation of the different elements of the capitalist structure, and the diversity of the historical roads to this formation.

The two elements necessary for the constitution of the structure of capitalist production each have their relatively independent history. In the text of Pre-Capitalist Economic Formations, after running through the history of the separation of the labourer from the means of production, Marx writes:

These, then, on the one hand, are the historical presuppositions for the labourer to be found as a free labourer, as objectiveless, purely subjective labour-power, confronting the objective conditions of production as his non-property, as someone else’s property, as value existing for itself, as capital. On the other hand, we must now ask what conditions are necessary for him to find a capital confronting him (Grundrisse, pp. 397-8;

We ought to be even more precise, and say: for him to find a capital confronting him in the form of money-capital. Marx then goes on to the history of the constitution of the second element: capital in the form of money-capital, and he returns to this second genealogy in Capital after the chapters devoted respectively to merchant’s capital and interest-bearing capital, i.e., once the elements necessary to the constitution of the capitalist structure have been analysed within that structure. The history of the separation of the labourer from the means of production does not give us money-capital (‘the question remains: whence came the capitalists originally? For it is clear that the expropriation of the agricultural population creates, directly, none but great landed proprietors’, Capital, T.III, p. 184; Vol. I, p. 742); for its part, the history of money-capital does not give the ‘free’ labourer (Marx notes this twice in Capital, vis-à-vis merchant’s capital – Vol. III, pp. 321-3 – and vis-à-vis finance capital – Vol. III, p. 582 – and in Pre-Capitalist Economic Formations he writes:

The mere existence of monetary wealth, even its conquest of a sort of supremacy, is not sufficient for this dissolution to result in capital. If it were, then ancient Rome, Byzantium, etc., would have concluded their history with free labour and capital, or rather, they would have begun a new history. There the dissolution of the old relations of property was also tied to the development of monetary wealth – of commerce, etc. However, in fact the result of this dissolution was not industry, but the domination of the countryside over the city ... Its [capital’s] original formation occurs simply because the historical process of the dissolution of an old mode of production allows value existing in the form of monetary wealth to buy the objective conditions of labour on the one hand, and to exchange the living labour of the now free workers for money on the other. All these moments are already in existence. What separates them out is a historical process, a process of dissolution, and it is this which enables money to turn into capital – Grundrisse, pp. 405-6; PCEF, pp. 109-10).

In other words, the elements combined by the capitalist structure have different and independent origins. It is not one and the same movement which makes free labourers and transferable wealth. On the contrary, in the examples analysed by Marx, the formation of free labourers appears mainly in the form of transformations of agrarian structures, while the constitution of wealth is the result of merchant’s capital and finance capital, whose movements take place outside those structures, ‘marginally’, or ‘in the pores of society’.

Thus the unity possessed by the capitalist structure once it has been constituted is not found in its rear. Even when the study of the pre-history of the mode of production takes the form of a genealogy, i.e., when it aims to be explicitly and strictly dependent, in the question that it poses, on the elements of the constituted structure, and on their identification, which requires that the structure is known as such in its complex unity – even then the pre-history can never be the mere retrospective projection of the structure. All it requires is that the meeting should have been produced and rigorously thought, between those elements, which are identified on the basis of the result of their conjunction, and the historical field within which their peculiar histories are to be thought. In their concepts, the latter have nothing to do with that result, since they are defined by the structure of a different mode of production. In this historical field (constituted by the previous mode of production), the elements whose genealogy is being traced have precisely only a ‘marginal’ situation, i.e., a non-determinant one. To say that the modes of production are constituted as combination variants is also to say that they transpose the order of dependence, that they make certain elements move in the structure (which is the object of the theory) from a place of historical domination to a place of historical subjection. I am not saying that the problematic is complete in this form, that it leads us to the threshold of a solution: but at any rate, this is how we can disengage it from the way in which Marx practices the analysis of primitive accumulation, explicitly closing all the roads ideology might take.

But already at this point, we can introduce a different consequence: it is the fact that the analysis of primitive accumulation, in its genealogical form, is adequate for one basic characteristic of the process of formation of the structure: the diversity of the historical roads by which the elements of the structure are constituted, by which they lead to the point at which they can join together and constitute that structure (the structure of a mode of production) by coming under its jurisdiction, becoming its effects (thus the forms of merchant’s capital and finance capital only become forms of capital in the strict sense on the ‘new bases’ of the capitalist mode of production – see Capital, Vol. III, pp. 322-3 and 583-4). Or again, to return to the terms mentioned above: the same set of presuppositions corresponds to several series of historical conditions. Here we are touching on a point which is all the more important in that Marx’s analyses in Volume One of Capital have led to misunderstandings, despite all his precautions: these analyses are explicitly the analyses of certain forms, certain ‘methods’ among others, of primitive accumulation, found in the history of Western Europe and mainly in that of England. Marx explained his position on this very point very clearly in his letter to Vera Zasulich of 8 March 1881 (the different drafts of which need to be read). There are therefore a plurality of processes of constitution of the structure which all reach the same result: their particularity depends on each occasion on the structure of the historical field in which they are situated, i.e., on the structure of the existing mode of production. The ‘methods’ of primitive accumulation which Marx describes in the English example must be related to the specific characteristics of the mode of production which is dominant in that particular case (the feudal mode of production), and notably to the systematic utilization of extra-economic (legal, political and military) power, which, as I recalled briefly above, was founded in the specific nature of the feudal mode of production. More generally, the result of the transformation process depends on the nature of the historical environment, of the existing mode of production: Marx shows this for merchant’s capital (Capital, Vol. III, pp. 326-7). In a text such as Pre-Capitalist Economic Formations, Marx describes three distinct forms of constitution of the free labourer (of the separation of the producer from his means of production), which constitute different historical processes, correspond to specific earlier forms of property, and are designated as so many different forms of ‘negation’ (Grundrisse, pp. 398-9, PCEF, pp. 99-101). Further on, and this list is referred to again in Capital, he similarly describes three distinct forms of the constitution of money-capital (which obviously have no one-to-one correspondence with the three forms of constitution of the free labourer):

There is, consequently, a three-fold transition: First, the merchant becomes directly an industrial capitalist. This is true in crafts based on trade, especially crafts producing luxuries and imported by merchants, together with raw materials and labourers from foreign lands, as in Italy from Constantinople in the fifteenth century. Second, the merchant turns the small masters into his middlemen, or buys directly from the independent producer, leaving him nominally independent and his mode of production unchanged. Third, the industrialist becomes a merchant and produces directly for the wholesale market (Capital, Vol. III, p. 330).

(We should also add the forms of usury which constitute the pre-history of interest-bearing capital and one of the processes of constitution of capital.)

The relative independence and historical variety of the constitution processes of capital are gathered together by Marx into a single word: the constitution of the structure is a ‘find’ (trouvaille); the capitalist mode of production is constituted by ‘finding already there’ (vorfinden) the elements which its structure combines (Grundrisse, p. 407; PCEF, p. 111). This find obviously does not imply chance: it means that the formation of the capitalist mode of production is completely indifferent to the origin and genesis of the elements which it needs, ‘finds’ and ‘combines’. Thus it is impossible for the reasoning whose movement I have retraced to be looped into a circle: the genealogy is not the other side of a genesis. Instead of re-uniting the structure and the history of its formation, the genealogy separates the result from its pre-history. It is not the old structure which itself has transformed itself, on the contrary, it has really ‘died out’ as such (‘All in all, the entire guild system – both master and journeyman – dies out, where the capitalist and the labourer emerge’, Grundrisse, p. 405; PCEF, p. 109). The analysis of primitive accumulation thus brings us into the presence of the radical absence of memory which characterizes history (memory being only the reflection of history in certain pre-determined sites – ideology or even law – and as such, anything but a faithful reflection).


Here I shall set aside this analysis of primitive accumulation, although I have not drawn every consequence from it, and turn to the study of the second movement, that of the dissolution of the capitalist mode of production (which I am using here as a paradigm). This second analysis deals with everything Marx tells us about the historical tendency of the capitalist mode of production, the peculiar movement of its contradiction, the development of the antagonisms implied by the necessity of its structure, and all that can be revealed in that structure of the exigency of a new organization of social production. If, as I have said, it is true that these two analyses have, by right, an object of the same nature (the transition from one mode of production to another) – which identity of object is perfectly clear in the text on the ‘Historical Tendency of Capitalist Accumulation’ (Capital, T.III, pp. 203-5; Vol. I, pp. 761-5) – it is no less clear that Marx treats them differently. The difference lies not just in the literary realization (on the one hand – for primitive accumulation – a historical analysis which is fairly extensive and detailed, but dissociated from the body of the exposition and apparently less systematic; on the other – the dissolution of capitalism – insights only, but formulated in general terms and organically linked to the analysis of the capitalist mode of production), it is the expression of two complementary theoretical situations: on the one hand, we have identified the elements whose genealogy has to be retraced, but we do not have in concept the knowledge of the historical field which is the theatre of this genealogy (the structure of the previous mode of production); on the other, we do have the knowledge of this historical field (the capitalist mode of production itself) and nothing else. Before formulating a complete problematic, we must therefore carry out a second preliminary reading.

In the first place, we can establish the strict theoretical equivalence of a number of ‘movements’ analysed by Marx at the level of the aggregate social capital: the concentration of capital (of the ownership of the means of production), the socialization of the productive forces (by the application of science and the development of co-operation), the extension of capitalist social relations to all branches of production and the formation of one world market, the constitution of an industrial reserve army (relative over-population), and the progressive decline in the average rate of profit. The ‘historical tendency’ of capitalist accumulation is identical in principle with the ‘tendential law’ analysed in Volume Three, which Marx calls the ‘real tendency of capitalist production’, and of which he writes:

The progressive tendency of the general rate of profit to fall is, therefore, just an expression peculiar to the capitalist mode of production of the progressive development of the social productivity of labour ... It is thereby proved a logical necessity proceeding from the nature of the capitalist mode of production, that in its development the general average rate of surplus-value must express itself in a falling general rate of profit (Capital, Vol. III, p. 209 – modified).

In fact, the tendency for the average rate of profit to fall is merely the immediate effect of the rise in the average organic composition of capital, of the constant capital expended as means of production compared with the variable capital expended as labour-power, which is the expression of the peculiar movement of accumulation. To say that all these movements are theoretically equivalent is therefore to say that they are different expressions in a single tendency, dissociated and expounded separately merely in the interests of the order of exposition (proof) of Capital. But their separation expresses no succession: from the view-point of the system of concepts, we are dealing with the same movement of analysis of the structure.

This movement is none other than the movement that Marx calls the development of the contradiction peculiar to the capitalist mode of production. Defined first in a very general way as the ‘contradiction’ between the socialization of the productive forces (which defines their development in the capitalist mode of production) and the character of the relations of production (private ownership of the means of production), it is then specified in the forms peculiar to the capitalist mode of production as the contradiction between an increase in the mass of value produced, and hence of profit, and a decrease in the rate of profit. But the search for profit is the sole motor of the development of production in the capitalist mode of production.

But what kind of movement is this? It seems that we could define it as a dynamics of the system, whereas the analysis of the complex combination which constitutes the structure of the mode of production fulfils the function of a statics, This pair of concepts does enable us to account for the movement insofar as it depends solely on the internal connections of the structure, insofar as it is the effect of that structure, i.e., its existence in time. Knowledge of this movement implies no other concepts than those of production and reproduction, in the form peculiar to the historical mode of production considered. Thus the ‘contradiction’ is not something different from the structure itself, it is indeed ‘immanent’ to it, as Marx says: but inversely, the contradiction by itself includes a dynamics: it is only given as a contradiction, i.e., it only produces contradictory effects in the temporal existence of the structure. It is therefore perfectly accurate to say, as Marx also does, that the contradiction ‘develops’ in the historical movement of capitalism.

The question we must examine can then be formulated as follows: is the dynamics of the structure at the same time – in the same ‘time’ – its history? In other words, is this movement at the same time a movement towards the historical future of capitalism? (and more generally towards the future of the mode of production considered, since each mode has its own specific ‘contradiction’, i.e., ‘an expression peculiar to’ it ‘of the progressive development of the social productivity of labour’). And since the relationship between the statics and the dynamics allows us to make the development of the contradiction the very movement which produces the effects of the structure, can we also say that it constitutes the ‘motor’ of its supersession? The identity – or difference – which we are looking for between this dynamics and this history is obviously the unity of the concepts, and cannot be satisfied by the coincidence provided ipso facto by a merely empirical temporality: if the development of the contradiction is inscribed in the chronology of a succession, it is quite simply that history. Since, on the contrary, we want to construct the relationship between these two concepts, Marx’s text forces us here to start from the most explicit concept (the dynamics of the development of the structure) in order to go on to, or attempt to go on to, the other (its historical future).

If we try to define more accurately what Marx meant by the ‘contradictory’ nature and ‘tendency’ of the mode of production, his repeated formulations confront us with the problem of the relationship between the structure and its effects. The ‘tendency’ is defined by a restriction, a diminution, a postponement or a travesty of effectivity. Tendency is a law ‘whose absolute action is checked, retarded and weakened, by counteracting causes (entgegenwirkende Ursachen)’ (Capital, Vol. III, p. 229 – modified), or even one whose effects (Wirkung, Verwirklichung, Durchführung) are annulled (aufheben) (p. 227) by these opposed causes. The tendency character thus appears first of all as a failure of the law, but an extrinsic failure, caused by the obstacle of external circumstances which do not depend on it, and whose origins are not explained (for the time being). The exteriority of the opposed causes is enough to justify the fact that their effectivity is purely negative: the result of their intervention is not to alter the result of the law itself, the nature of its effects, but merely the chronology of their production; we are thus led to define a tendency as something which is only realized in the long run, and the retarding causes as a set of empirical circumstances which merely mask the essence of the process of development. ‘Thus’, writes Marx, ‘the law acts only as a tendency. And it is only under certain circumstances and only after long periods that its effects become strikingly pronounced’ (Capital, Vol. III, p. 233).

But this definition is unsatisfactory, because, in its empiricist and mechanistic character, it is a return precisely to what Marx criticized in the economists, particularly Ricardo: the study of ‘factors’ called ‘independent’ because of an inability to find their common origin in the unity of a structure, a study which belongs to the ‘exoteric’ or ‘vulgar’ side of political economy. It also ignores Marx’s systematic use of the term tendency to designate the laws of production themselves, or else the laws of the movement of production insofar as this movement depends on its structure. In the Preface to the first German edition of Capital, Marx wrote:

It is not a question of the more or less complete development of the social antagonisms that result from the natural laws of capitalist production, but of these laws themselves, of these tendencies manifesting and realizing themselves with iron necessity (Capital, T.I, p. 18; Vol. I, p. 8).

And also in Volume One, to formulate the law of the production of relative surplus-value:

The general result is treated, here, as if it were the immediate result directly aimed at. When a capitalist cheapens shirts by increasing the productivity of labour, he does not necessarily aim thereby to reduce the value of labour-power and shorten the portion of the day in which the worker works for himself. But it is only insofar as he ultimately contributes to this result, that he contributes to raising the general rate of surplus-value. The general and necessary tendencies (Tendenzen) of capital must be distinguished from the forms in which they appear (Erscheinungsformen).

It is not our intention to consider here the way in which the immanent laws of capitalist production (immanente Gesetze) appear in the external movements of capitals, assert themselves as coercive laws of competition, and thereby impose themselves on the capitalist as the motives of their operations (Capital, T.II, p. 10; Vol. I, p. 316).

Here it seems that Marx’s term ‘tendency’ designates not a restriction on the law due to external circumstances, which necessarily belong to the sphere of ‘appearances’, of ‘surface’ phenomena, but the law itself, independently of any extrinsic circumstance. If Marx’s vocabulary is rigorous here, we may think that it is only as a first appearance that the law of the development of production (expressed in the fall in the rate of profit, etc.) is externally limited.

But if we examine the ‘causes’ hindering the realization of the tendency one by one, we find that they are all either the immediate effects of the structure or determined by the structure which sets limits (Grenzen) on the variation of their effects. Under the first heading we can list the increasing intensity of exploitation, the depreciation of existing capital, relative over-population and its restriction to less developed branches of production, the increase in the scale of production (and the creation of an external market); under the second, the depression of wages below the value of labour-power. Now, it is peculiar to all causes which are immediate effects of the structure that they are ambivalent: so much so that all the causes that counteract the action of the law are at the same time the causes which produce its effects:

But since the same causes which raise the rate of surplus-value (even a lengthening of the working-time is a result of large-scale industry) tend to decrease the labour-power employed by a certain capital, it follows that they also tend to reduce the rate of profit and to retard this reduction (Capital, Vol. III, p. 229 – modified).

Similarly, the depreciation of the existing capital is linked to the increase in the productivity of labour, which cheapens the elements of constant capital, and thus prevents the value of constant capital from increasing in the same proportion as is material volume, etc. In a general way, if the aggregate social capital is considered, ‘the same causes which produce a tendency in the general rate of profit to fall, also call forth counter-effects’ (Capital, Vol. III, p. 233 – modified). This is a crucial point, for it enables us to establish the fact that the reduction of the law of development to the status of a tendency is not a determination external to that law, influencing only the chronology of its effects, but an intrinsic determination of the production of its effects. The effect of the opposed causes, i.e., of the law itself, is not to delay the historical effects of capitalist production, but to determine a specific rhythm for the production of those effects, a determination which only appears negatively (as a ‘restriction’, etc.) with reference to the ahistorical absolute of a ‘free’, ‘unlimited’ growth of the productivity of labour (leading to an increase in the organic composition of capital and a fall in the rate of profit). Moreover, the definition of the mode of action peculiar to the structure, which includes the reduction of the apparent exteriority of the opposing causes, is once again linked to the consideration of the social capital (or what comes to the same thing, of the ‘individual capital as an aliquot part of the total capital’ – Vol. III, p. 216), which is the theoretical support for Volume One and the first half of Volume Two, i.e., the consideration of a capital in the theoretical ‘synchrony’ which I discussed with respect to reproduction. All the reasoning that enables Marx to establish the existence and level of a general average rate of profit depends on such a synchrony (Marx calls it a simultaneity) in which the addition together of the capitals portion by portion is possible by definition; if we were obliged to ask to what extent does the cheapening of the means of production one by one hinder the value of constant capital from increasing with respect to that of the corresponding variable capital, it would become impossible to establish such a law. The impure theoretical status of the ‘causes’ which ‘counteract’ the fall in the general rate of profit merely reveals, in a number of formulations (which I have cited), Marx’s difficulty in thinking this ‘synchrony’ explicitly, insofar as it was a matter of a law of development of the structure. But in fact he closes the circle nevertheless, since it is the tendential fall in the rate of profit which arouses the competition of capitals, i.e., the mechanism by which the equalization of profits and the formation of the general rate of profit are actually achieved (Capital, Vol. III, p. 250). (At the same stroke, this clarifies and limits the place of competition, for Marx excludes the analysis of its mechanism from the analysis of capital in general, since it merely ensures this equalization without determining the level at which it is established, just as it did for the market price of a particular commodity.) The development of the structure according to a tendency, i.e., a law which does not only (mechanically) include the production of effects, but also the production of effects according to a specific rhythm, therefore means that the definition of the specific internal temporality of the structure is part of the analysis of that structure itself.

It is now clear what is ‘contradictory’ about tendency, which enables us to illuminate the true status of contradiction in Marx. Marx defines the terms between which there is a contradiction as the contradictory effects of a single cause:

Thus, the same development of the social productiveness of labour expresses itself with the progress of capitalist production on the one hand in a tendency of the rate of profit to fall progressively and, on the other, in a progressive growth of the absolute mass of the appropriated surplus-value, or profit; so that on the whole a relative decrease of variable capital and profit is accompanied by an absolute increase of both. This two-fold effect (doppelseitige Wirkung), as we have seen, can express itself only in a growth of the total capital at a pace more rapid than that at which the rate of profit falls ... To say that the mass of profit is determined by two factors – first, the rate of profit, and, secondly, the mass of capital invested at this rate, is mere tautology. It is therefore but a corollary of this tautology to say that there is a possibility for the mass of profit to grow even though the rate of profit may fall at the same time. It does not help us one step further ... But if the same causes which make the rate of profit fall, entail the accumulation, i.e., the formation of additional capital, and if each additional capital employs additional labour and produces additional surplus-value; if, on the other hand, the mere fall in the rate of profit implies that the constant capital and with it the total old capital, have increased, then this process ceases to be mysterious (Capital, Vol. III, pp. 219-21).

(Obviously, it is one and the same thing to say that the fall in the rate of profit is slowed down by the growth in the scale of production, as above, or to say, as here, that the mass of accumulation is relatively diminished by the fall in the rate of profit.) This very important definition includes both the refutation of an empirical notion of contradiction (which Marx links to Ricardo’s name – Capital, Vol. III, p. 243), and the limitation of its role. The empiricism of classical economics could only reveal contradictory terms as in ‘peaceful coexistence’, i.e., in the relative autonomy of distinct phenomena, e.g., successive ‘phases’ of development dominated inversely by one or other of the contradictory tendencies... . Marx, on the contrary, produced the theoretical concept of the unity of the two contradictory terms (which he calls a ‘combination’ here too: ‘the tendency of the rate of profit to fall is combined with – ist verbunden mit – a tendency of the rate of surplus-value to rise, hence with a tendency for the rate of labour exploitation to rise’ – Capital, Vol. III, p. 234 – modified), i.e., he produces the knowledge of the contradiction’s foundation in the nature of the structure (of capitalist production). Classical economics reasons from independent ‘factors’ whose interaction ‘may’ induce such and such a result: the whole problem is therefore to measure these variations and relate them empirically to other variations (the same is true of prices, and of the values of commodities, which are supposed to depend on the variation of certain factors: wages, average profit, etc.). Marx does not regard the law (or tendency) as a law of variation in the size of the effects, but as a law of production of the effects themselves: it determines these effects on the basis of the limits within which they can vary, and which do not depend on this variation (the same is true of wages, the working day, prices, and the different fractions into which surplus-value is divided); it is these limits alone which are determined as effects of the structure, and in consequence they precede the variation instead of being its average resultant. It is by the law of its production from a single cause that contradiction is given us here, and not in the variation of its result (the level of accumulation).

But this definition also includes the limitation of the role of contradiction, i.e., its situation of dependence with respect to the cause (the structure): there is only a contradiction between the effects, the cause is not divided in itself, it cannot be analysed in antagonistic terms. Contradiction is therefore not original, but derivative. The effects are organized in a series of particular contradictions, but the process of production of these effects is in no way contradictory: the increase in the mass of profit (and hence the scale of accumulation) and the decrease in its rate (hence the peculiar speed of accumulation) are the expression of one and the same increasing movement of the quantity of means of production set to work by capital. That is why only an appearance of contradiction is found in the knowledge of the cause: ‘this law’, says Marx, ‘this inner and necessary connection between two seeming contradictions’ (Vol. III, p. 220); the inner and necessary connection which defines the law of production of the effects of the structure excludes logical contradiction. From this point of view, the ‘two-fold effect’ is thus merely the ‘double-edged’ (zwieschlächtig – Vol. III, p. 215) nature of the law. It is particularly noteworthy that here, in order to express the derivative and dependent character of the contradiction between certain effects of the structure, we find Marx returning to the same term that he used at the beginning of Capital to designate the false contradiction ‘in adjecto’ of the commodity (on this point see Pierre Macherey’s paper).[22] For their part, the effects present a simple contradiction (a term-by-term contradiction: relative over-population and relative over-production, etc.) and one distributed into several contradictory aspects or component contradictions which, for all that, do not constitute an overdetermination, but simply have inverse effects on the scale of accumulation.

Just as the cause which produces the contradiction is not itself contradictory, so the result of the contradiction is always a certain equilibrium, even when this equilibrium is attained by way of a crisis. Thus it seems that contradiction has a status analogous to that of competition in the movement of the structure: it determines neither its tendency nor its limits, rather it is a local, derivative phenomenon, whose effects are pre-determined in the structure itself:

These different influences may at one time operate predominantly side by side in space, and at another succeed one another in time. From time to time the conflict of antagonistic agencies finds vent in crisis. The crises are always but momentary and forcible solutions of the existing contradictions. They are violent eruptions which for a time restore the disturbed equilibrium... . The periodical depreciation of existing capital – one of the means immanent in the capitalist mode of production to check the fall of the rate of profit and hasten accumulations of capital-value through formation of new capital – disturbs the given conditions, within which the process of circulation and reproduction of capital takes place, and is therefore accompanied by sudden stoppages and crises in the production process... The ensuring stagnation of production would have prepared – within capitalistic limits – a subsequent expansion of production. And thus the cycle would run its course anew (Capital, Vol. III, pp. 244 and 250 – modified).

Thus the only intrinsic result of the contradiction, which is completely immanent to the economic structure, does not tend towards the supersession of the contradiction, but to the perpetuation of its conditions. The only result is the cycle of the capitalist mode of production (the crisis is cyclical because the reproduction of the aggregate capital depends on the turnover of fixed capital – see Capital, Vol. II, p. 186 – but it is possible to say metaphorically that the crisis manifests the circle in which the whole mode of production moves with an immobile movement).

Marx also says that the crisis reveals the barriers (Schranken) of the mode of production:[23]

Capitalist production seeks continually to overcome these immanent barriers (immanenten Schranken), but overcomes them only by means which again place these barriers in its way and on a more formidable scale. The real barrier (die wahre Schranke) of capitalist production is capital itself (Capital, Vol. III, p. 245).

The ‘limits’ towards which the movement of the mode of production tends (its dynamics) are not therefore a question of a ladder, of a threshold to attain. If the tendency cannot pass these limits, it is because they are inside it, and as such never reached: in its movement it carries them with it, they coincide with the causes which make it a ‘mere’ tendency, i.e., they are simultaneously its actual conditions of possibility. To say that the capitalist mode of production has internal limits is quite simply to say that the mode of production is not a ‘mode of production in general’, but a delimited, determinate mode of production:

The capitalist mode of production meets in the development of its productive forces a barrier which has nothing to do with the production of wealth as such: and this peculiar barrier testifies to (bezeugt) the limitations (Beschränktheit) and to the merely historical, transitory character of the capitalist mode of production; testifies that for the production of wealth, it is not an absolute mode, moreover, that at a certain stage (auf gewisser Stufe) it rather conflicts with its further development (Capital, Vol. III, p. 237).

(The term wealth should always be regarded as strictly synonymous with use-value.)

These limits are therefore the same as those whose effects we have already met in the determination of the tendency: a mode of production of wealth in itself does not exist, i.e., there only exists a determinate type of development of the productive forces, depending on the nature of the mode of production. The rise in the productivity of labour is limited by the nature of the relations of production which make it into a means of formation of relative surplus-value. For its part, the extorsion of surplus-value is limited by the productivity of labour (within the limits of variation of the working day, the relationship between necessary labour and surplus-labour is given by this productivity at each moment). What we see here therefore is not the contradiction, but the complexity of the mode of production, which was defined at the beginning of this exposition as a double articulation of the mode of production (‘productive forces’, relations of ownership of the means of production): the internal limits of the mode of production are none other than the limitation of each of these two connections by the other, i.e., the form of their ‘correspondence’ or of the ‘real subsumption’ of the productive forces beneath the relations of production.

But if the limits of the mode of production are internal ones, they only determine what they affirm and not what they deny (i.e., via the idea of an ‘absolute mode of production’, a mode of production ‘of wealth in itself’, the possibility of all the other modes of production which have their own peculiar internal limitations). Only in this sense do they imply the transition to a different mode of production (the historical, transitional character of the existing mode of production): they designate the necessity for a way out and a different mode of production whose delimitation is absolutely absent from them; and since the limits consist of the ‘correspondence’ which articulates the two connections within the complex structure of the mode of production, the movement suppressing these limits implies the suppression of the correspondence.

But it is also clear that the transformation of these limits does not simply belong to the time of the dynamics. Indeed, if the effects within the structure of production do not by themselves constitute any challenge to the limits (e.g., the crisis, which is ‘the mechanism [with which] capitalist production spontaneously gets rid of the obstacles that it happens on occasion to create’ – Capital) there may be one of the conditions (the ‘material basis’) of a different result, outside the structure of production: it is this other result which Marx suggests marginally in his exposition when he shows that the movement of production produces, by the concentration of production and the growth of the proletariat, one of the conditions of the particular form which the class struggle takes in capitalist society. But the analysis of this struggle and of the political social relations which it implies is not part of the study of the structure of production. The analysis of the transformation of the limits therefore requires a theory of the different times of the economic structure and of the class struggle, and of their articulation in the social structure. To understand how they can join together in the unity of a conjuncture (e.g., how, if other conditions are fulfilled, the crisis can be the occasion for a – revolutionary – transformation of the structure of production) depends on this, as Althusser has shown in an earlier study (‘Contradiction and Overdetermination’, in For Marx).


The preceding analyses constitute a number of still disjointed moments of the problematic within which it is possible to think theoretically the transition from one mode of production to another. It will not be possible to articulate this problematic effectively, i.e., to produce the unity of the questions which have to be answered, until we succeed in situating with respect to one another the concepts that we have proposed up to this point (history, genealogy, synchrony – diachrony, dynamics, tendency) and in defining differentially their peculiar objects.

All these concepts, which are still largely descriptive and will remain so precisely so long as they are not articulated, seem to be so many conceptualizations of historical time. In an earlier paper, Althusser showed that, in any theory of history (whether scientific or ideological), there was a rigorous and necessary correlation between the structure of the concept of history peculiar to that theory (a structure itself dependent on the structure of the concept of the social totality peculiar to that theory), on the one hand, and the concept of temporality in which that theory of history thinks the ‘changes’, ‘movements’, ‘events’, or, more generally, the phenomena which appertain to its object, on the other. The fact that this theory is usually absent as such, that it is reflected in the form of a non-theory, i.e., of empiricism, does not contradict such a demonstration. The structure of temporality is then quite simply that provided by the ruling ideology, and it is never reflected in its function as a presupposition. We have even found that in Hegel the structure of historical temporality, being dependent, from the point of view of the articulation of the system, on the structure of the simple Hegelian totality, i.e. of the expressive totality, merely took up on its own accord the very form of the empiricist ideological conception of time, providing it with its concept and theoretical foundation.

At the same time, we have found that the form of this time was not only the continuous linearity, but also, by way of consequence, the uniqueness of time. Because time is unique, its present has the structure of contemporaneity, and all the moments whose chronological simultaneity can be established must also necessarily be determined as the moments of one and the same current whole, they must necessarily belong to the same history. Here we should note that, in this ideological conception, the peculiar form of time precedes the determination of historical objects in relation to it: the order and duration of this time always precede any determination of a phenomenon as ‘unfolding over time’ and thereby as a historical phenomenon. Of course, the effective estimation of order or duration always presupposes a connection with or reference to the temporality of certain objects, but the form of their possibility is always already given. In reality, this is to move in a circle, since it is to admit the structure of a time which is merely the effect, either of a perception, or of an ideological conception of the social totality. But this movement of real dependence, before the location of ‘historical’ phenomena in time, is not thought as such in the representation of time which serves as its premiss, and it is possible to take the short cut of discovering (in reality, rediscovering) in the determinations of history the presupposed structure of this time. From this movement we get the determination of the historical object as an event, present even when it is doubted, i.e., in the idea that there are not only events, i.e., not only ‘short’-term phenomena, but also non-events, i.e., long events, long-term permanences (which are wrongly christened ‘structures’).

If we then remember the problematic within which Marx originally thought his theoretical undertaking, but which was not peculiarly his problematic, the problematic of periodization, we can draw several conclusions. If we pose the problem of the transition from one mode of production to another solely in the framework of this problematic, it is impossible for us to escape the form of unique linear time: we must think the effects of the structure of each mode of production on an equal footing with the phenomena of transition, situating them in the unique time which serves as a framework or common support for every possible historical determination. We have no right to establish differences in principle or method between analyses of the effects of a mode of production and analyses of the transition between two modes of production which succeed one another or coincide with one another in the framework of this time, and we can only distinguish the movements by determinations of the ‘structure’ of this time: long-term, short-term, continuity, intermittency, etc. The time of periodization is therefore a time for which any true diversity is impossible: the supplementary determinations which are inserted in the course of a historical sequence, e.g., during transitions from one mode of production to another, are part of the same time as them, and have the movement of their production in common with them.

Moreover, a superficial reading of Marx is more than likely not to dissipate the forms of this illusion, if it is content to take the different ‘times’ implied by the analysis in Capital as so many descriptive aspects or subordinate determinations of time in general. It would then be possible to try practising the fundamental operation implied by the ideological theory of time: the insertion of the different times one within another. It would be possible to inscribe the segmented times (labour times, production times, circulation times) in cycles (the cyclical process of capital); these cycles themselves would necessarily be complex cycles, cycles of cycles, because of the different turnover speeds of the different elements of capital, but as a whole they in their turn could be inserted in the general movement of capitalist reproduction (accumulation), which Marx, following Sismondi, describes as a spiral ; and finally this ‘spiral’ would manifest a general tendency, an orientation – precisely that of the transition from one mode of production to another, of the succession of the modes of production and of periodization. In such a reading the harmonization of the different ‘times’ and the imbrication of their forms would obviously raise no difficulties in principle, for their possibility would already be inscribed in the uniqueness of time in general which serves as a support for all these movements. The only difficulties would be difficulties of application, difficulties in identifying the phases and in forecasting the transitions.

What is most noteworthy in such a reading – which is, as we shall see, not just a purely polemical expository device on my part – is that it necessarily implies that each ‘moment’ of time is thought simultaneously as a determination of all the intermediate times which have been inserted into one another in this way – whether this determination is immediate, or, on the contrary, merely mediated. And to draw the most extreme consequence straightaway, it is absolutely consistent with this conception to determine a given time during which the worker expends his labour-power as a certain quantity of social labour, as a moment of the cycle of the production process (in which capital exists in the form of productive capital), as a moment of the reproduction of social capital (of capitalist accumulation) and finally as a moment of the history of the capitalist mode of production (which tends towards its transformation, however distant the latter may be).

Such an ideological reading provides the base from which it is possible to characterize the whole Marxist theory of the economic structure as a dynamics. The concept has been re-introduced in this way in order to oppose Marx to classical and modern political economy, while situating both on the same terrain, and assigning them the same ‘economic’ object: Marx thus becomes one of the innovators, perhaps the main one, who have introduced ‘dynamic’ theory into political economy (see for example Granger’s Méthodologie économique). This has made it possible to present classical and neo- classical economics as theories of economic equilibrium, i.e., of a ‘statics’ of the connections of the economic structure; while Marx, on the contrary, is supposed never to have seen the study of equilibrium as anything more than a provisional moment, operational in scope, an expository simplification; the essential object of Marx’s analysis is the time of the evolution of the economic structure, analysed in its successive components, the different ‘times’ of Capital:

As for the particular object of Marx’s study, capitalist production, it is necessarily presented as a dynamic process. Capitalist accumulation is the object of Volume One of Capital. The notion of a static equilibrium is obviously a priori incorrect as a description of this phenomenon. The ‘simple reproduction’ of capital is already a temporal process ; but it is little more than a first abstraction. The system is characterized precisely by ‘reproduction on an extended scale’, the growth and continuous qualitative metamorphosis of capital through the accumulation of surplus-value. The various forms of crisis appear as a chronic disorder of the system, not as accidents. The general picture of economic reality is thus made totally dynamic (G. G. Granger: Méthodologie économique, p. 98).

Such an interpretation, in which the dynamics of the capitalist system is itself a moment, a local aspect of the ‘claim that the laws of the economy are relative and evolutionary in character’, is really an example of the structure of temporal insertion that I outlined above. The concepts of history and dynamics then become twins, one popular history), the other learned (dynamics), since the second expresses very accurately the determination of the historical movement on the basis of a structure. This makes it possible to add a third term to these two: diachrony, which does not produce any new knowledge here, since it simply expresses the form of unique linear temporality which is implied by the identification of the first two concepts.

But in reality, such a reading of Marx completely ignores the mode of constitution of the concepts of temporality and history in the theory of Capital. It may have been possible to adopt (or interpret) these concepts in their normal sense, i.e., in their ideological use, in a text such as the Preface to A Contribution, from which we started: there they merely have the function of registering and designating a theoretical field which has not yet been thought in its structure. But in the analysis of Capital, as our studies of primitive accumulation and of the tendency of the mode of production have shown, they are produced separately and differentially: their unity, instead of being presupposed in an always already given conception of time in general, must be constructed out of an initial diversity which reflects the complexity of the whole which is analysed. On this point it is possible to generalize from the way Marx posed the problem of the unity of the different cycles of the individual capitals in a complex cycle of the social capital: this unity must be constructed as an ‘intertwining’ whose nature is initially problematic. On this, Marx writes:

Therefore, the manner in which the various component parts of the aggregate social capital, of which the individual capitals are but constituents functioning independently, mutually replace one another in the process of circulation – in regard to capital as well as surplus-value – is not ascertained from the simple intertwinings of the metamorphoses in the circulation of commodities – intertwinings which the acts of capital circulation have in common with all other circulation of commodities. That requires a different method of investigation. Hitherto one has been satisfied with uttering phrases which upon closer analysis are found to contain nothing but indefinite ideas borrowed from the intertwining of metamorphoses common to all commodity circulation (Capital, Vol. II, p. 115).

We know that this ‘different method of investigation’ which peculiarly constitutes the analysis of the reproduction of the total social capital, leads to a paradoxical result: a synchronic structure of the relation between the different sectors of social production, in which the peculiar form of a cycle has completely disappeared. But this method alone allows us to think the intertwining of the different individual production cycles. In the same way, the complex unity of the different ‘times’ of historical analysis, those which depend on the permanence of the social relations and those in which is inscribed the transformation of the social relations, is initially problematic: it must be constructed by a ‘different method of investigation’.

The relationship of theoretical dependence between the concepts of time and history is thus inverted with respect to the preceding form, which belongs to empiricist or Hegelian history, or to a reading of Capital which implicitly reintroduces empiricism or Hegelianism. Instead of the structures of history depending on those of time, it is the structures of temporality which depend on those of history. The structures of temporality and their specific differences are produced in the process of constitution of the concept of history, as so many necessary determinations of its object. Thus the definition of temporality and its various forms becomes explicitly necessary; similarly, the necessity of thinking the relationship (the harmony) between the different movements and the different times becomes a basic necessity for theory.

In Marx’s theory, therefore, a synthetic concept of time can never be a pre-given, but only a result. The preceding analysis in this paper allows us to anticipate this result to a certain extent, and to propose a differential definition of concepts which have been confused until now. We have seen that the analysis of the relations which appertain to a determinate mode of production and constitute its structure must be thought as the constitution of a theoretical ‘synchrony’: this is reflected with respect to the mode of production by Marx in the concept of reproduction. The analysis of all the peculiar effects of the structure of the mode of production is necessarily part of this synchrony. The concept of diachrony will therefore be reserved for the time of the transition from one mode of production to another, i.e., for the time determined by the replacement and transformation of the relations of production which constitute the double articulation of the structure. Thus it appears that the ‘genealogies’ contained in the analyses of primitive accumulation are elements of diachronic analysis: and thus the difference in problematic and methods between the chapters of Capital devoted to primitive accumulation and all the others, irrespectively of their degree of theoretical perfection, has been established as more than a mere difference in style or literary form. This difference is a consequence of the strict distinction between ‘synchrony’ and ‘diachrony’, and we have met; another example of this in what goes before, an example to which I shall return: when I analysed the forms of the two connections (property, ‘real appropriation’) peculiar to the capitalist mode of production and the relationship between them, we observed a ‘chronological dislocation’ in the constitution of these two forms, the capitalist form of property (‘capitalist relations of production’) chronologically preceding the capitalist form of real appropriation (‘capitalism’s productive forces’); this dislocation was reflected by Marx in his distinction between the ‘formal subsumption’ of labour to capital and its ‘real subsumption’. At the time, I remarked that this chronological dislocation was suppressed as such in the synchronic analysis of the structure of the mode of production, that it was then indifferent to the theory. In fact, this dislocation, which then purely and simply disappears, can only be thought in a theory of the diachrony; it constitutes a relevant problem for diachronic analysis. (Here we should note that the expressions ‘diachronic analysis’ and ‘diachronic theory’ are not absolutely rigorous; it would be better to say ‘analysis – or theory – of the diachrony ‘. For if the terms ‘synchrony’ and ‘diachrony’ are taken in the sense which I have proposed here, the expression ‘diachronic theory’ has no meaning, strictly speaking: all theory is synchronic insofar as it ex- pounds a systematic set of conceptual determinations. In an earlier essay, Althusser has criticized the synchrony-diachrony distinction insofar as it implies a correlation between objects or aspects of a single object, showing how it was, in fact, a version of the empiricist – and Hegelian – structure of time, in which the diachronic is merely the development (devenir) of the present – the ‘synchronic’. It is clear straightaway that this cannot be the case in the usage which I have proposed here, since the synchrony is not a real self-contemporaneous present, but the present of the theoretical analysis in which all its determinations are given. This definition therefore excludes any correlation between the two concepts, one of which designates the structure of the thought process, while the other designates a particular relatively autonomous object of analysis, and only by extension the knowledge of it.)

For its part, the synchronic analysis of the mode of production implies that we stress several concepts of ‘time’ which differ in function. All these times are not directly, immediately historical: they are not in fact constructed out of the general historical movement, but quite independently of it, and independently of one another. Thus, the time of social labour (which measures the value produced) is constructed on the basis of the distinction between socially necessary labour and socially unnecessary labour, which depends at each moment on the productivity of labour and the proportions in which social labour is divided among the different branches of production (see Capital, T.I, pp. 59ff.; Vol. I, pp. 44ff.: Theories of Surplus Value, Vol. I, pp. 225-6). Thus it does not coincide at all with the empirically observable time during which a labourer works. In the same way, the cyclical time of the turnover of capital, with its different moments (production time, circulation time) and its peculiar effects (regular disengagement of money-capital, change in the rate of profit), is constructed on the basis of the metamorphoses of capital and the distinction between fixed and circulating capital.

In the same way, finally, the analysis of the tendency of the capitalist mode of production produces the concept of the dependence of the advance of the productive forces in relation to the accumulation of capital, and therefore the concept of the peculiar temporality of the productive forces in the capitalist mode of production. Only this movement can be called a dynamics as I have proposed, i.e., a movement of development inside the structure and sufficiently determined by it (the movement of accumulation), proceeding according to a peculiar rhythm and speed determined by the structure, with a necessary and irreversible orientation, and indefinitely retaining (reproducing) the properties of the structure on a different scale. The peculiar rhythm of capitalist accumulation is inscribed in the cycle of crises, while its peculiar speed expresses the ‘limitation’ of the development of the productive forces, as Marx says, simultaneously accelerated and decelerated, i.e., the reciprocal limitation of the two connections articulated in the structure (capitalist ‘productive forces’, relations of production). The necessary orientation of the movement consists of the increase in constant capital with respect to variable capital (in the production of means of production with respect to the production of means of consumption). The retention of the properties of the structure is particularly clear in the expansion of the market: one of the means employed by the capitalist or by an ensemble of capitalists to counter-act the fall in the rate of profit being to expand the field of his or their market (by ‘external’ trade):

This internal contradiction [between production and consumption] seeks to resolve itself through expansion of the out-lying fields of production. But the more productiveness develops, the more it finds itself at variance with the narrow basis on which the conditions of its consumption rest (Capital, Vol. III, p. 240).

In this ‘out-lying’ adventure, therefore, capitalist production always meets its own peculiar internal limitation, i.e., it never escapes being determined by its own peculiar structure.

Only in the ‘time’ of this dynamics can the ‘age’ of capitalist production, of one of its branches, or of a set of branches of production, be determined: this age is measured precisely by the level of the relation between constant capital and variable capital, i.e., by the internal organic composition of capital:

It goes without saying that the more advanced the age of capitalist production, the more money is accumulated in all hands, and therefore the smaller the quantity annually added to this hoard by the production of new gold, etc. (Capital, Vol. II, p. 473 – modified).

This is a very important point, for it shows that only in the ‘time’ of the dynamics – which, as I have said, is not immediately the time of history[24] – is it possible to determine and assess the forwardnesses or backwardnesses of development ; indeed, only in this internal orientated time can historical unevennesses of development be thought simply as temporal dislocations:

What is true of different successive stages of development in one country, is also true of different coexisting stages of development in different countries. In an undeveloped (unentwickelt) country, in which the former composition of capital is the average, the general rate of profit would = 66 2/3 per cent, while in a country with the latter composition and a much higher stage of development it would = 20 per cent. The difference between the two national rates of profit might disappear, or even be reversed, if labour were less productive in the less developed country ... The labourer would then spend more of his time in reproducing his own means of subsistence, or their value, and less time in producing surplus-value (Capital, Vol. III, p. 210).

The consequences of this differential determination of time, and of the distinction between the time of the dynamics and the time of history in general for the contemporary problem of ‘under-development’ (which is a favourite haunt for every theoretical confusion) cannot be expounded here; at least what we have said gives us a foretaste of its critical importance.

Like the preceding ones, this ‘time’ of the dynamics (of the tendency) is determined in the synchronic analysis of the mode of production. The distinction between dynamics and diachrony is therefore a strict one, and the former cannot appear as one determination in the field of the latter, in which it is not relevant in the form in which Marx analyses it. It is easy to cast light on this distinction by borrowing a paradox from the analysis of the societies ‘without a history’ (strictly speaking a meaningless expression, for it designates social structures in which the dynamics appears in the peculiar guise of a non-development, as in the Indian communities which Marx discusses in Capital, T.II, pp. 46-8; Vol. I, pp. 356-9): the event constituted by the meeting between these societies and ‘Western’ societies in transition to capitalism (in conquest, colonization, or the various forms of commercial connection) is obviously part of the diachrony of those societies, since it determines – more or less brutally – a transformation of their modes of production: but it is no part of these societies’ dynamics. This event in their history is produced in the time of their diachrony without being produced in the time of their dynamics: a limit-case which brings out the conceptual difference between the two times, and the necessity of thinking their articulation.

We must therefore finally situate the concept of history with respect to these different concepts: should we for example assimilate it to the concept of diachrony, remembering the old problematic of periodization? Can we say that ‘history’ is this diachrony, the basic theoretical problem of which is the analysis of the modes of transition from one structure of production to another? No, obviously, for this old problematic has now been transformed. It is no longer defined by the necessity of ‘cutting up’ linear time, which would presuppose this reference time as an a priori. The question now is to think theoretically the essence of the transition periods in their specific forms and the variations of these forms. The problem of periodization in the strict sense has therefore been suppressed, or rather it has ceased to be part of the moment of scientific proof, of what Marx called the order of exposition (only exposition is science): periodization as such is at most a moment of the investigation, i.e., a moment of the preliminary critique of the theoretical materials and their interpretations. Here the concept of history is therefore not identical with any of the particular moments produced in theory in order to think the differential forms of time. The concept of history in general, unspecified, is simply the designation of a constitutive problem of the ‘theory of history’ (of historical materialism): it designates that theory as a whole as the site of the problem of the articulation of the different historical times and the variants of this articulation. This articulation no longer has anything to do with the simple model of the insertion of one time into another; it accepts coincidences not as obviousnesses, but as problems: for instance, the transition from one mode of production to another may seem to be the moment of a collision, or collusion, between the times of the economic structure, of the political class struggle, of ideology, etc. The question is to discover how each of these times, e.g., the time of the ‘tendency’ of the mode of production, becomes a historical time.

But if the general concept of history has the peculiar function of designating a constitutive problem of the theory of history, then, as opposed to the preceding concepts, it does not belong to that theory of history. And indeed, the concept of history is no more a concept of the theory of history than the concept oflife’ is a concept of biology. These concepts belong to the epistemologies of these two sciences, and, as ‘practical’ concepts, to the practice of the scientists, locating and staking out the field of that practice.


Here I can only outline a number of the concepts that belong to the theory of the ‘diachrony’ and enable us to think the nature of periods of transition from one mode of production to another. Indeed, as we have seen, Marx devoted far less theoretical effort to this second moment of the theory of history than he did to the first. On this point, I have no other aim here than to draw up a balance-sheet of results.

The analysis of Primitive Accumulation is part of the field of diachronic study, but not in itself part of the definition of the periods of transition (to capitalism). In fact, the analysis of primitive accumulation, of the origin of the capitalist mode of production, gives an element by element genealogy which passes through the transition period, but which in the same movement ascends to the heart of the previous mode of production. The outline definitions which can be borrowed from it must therefore be related to a different analysis which is not an analysis of the origins but one of the beginnings of the capitalist mode of production, and which in consequence does not proceed element by element, but from the point of view of the whole structure. In the study of manufacture we have a notable example of this analysis of the beginnings. The forms of transition are in fact necessarily modes of production in themselves.

In the first part of this paper, when I examined manufacture as a certain form of the connection of real appropriation, a certain form of the ‘productive forces’, I set aside the problem posed by the chronological dislocation in the constitution of the structure of capitalist production between the formation of its specific property relations and that of its specific ‘productive forces’. As I showed, this was not part of the examination of the structure of the mode of production. In contrast, this dislocation constitutes the essence of manufacture as a form of transition. The concepts which Marx uses to designate this dislocation are those of ‘real subsumption’ and ‘formal subsumption’ (of labour to capital). The ‘formal subsumption’ which begins with the form of out-work on behalf of a merchant capitalist and ends with the industrial revolution includes the whole history of what Marx calls ‘manufacture’.

In the ‘real subsumption’ of modern industry, the labourer’s belonging to capital is doubly determined: on the one hand he does not possess the material means to work on his own behalf (the ownership of the means of production); on the other, the form of the ‘productive forces’ takes away his ability to set the social means of production to work on his own, outside an organized and inspected process of co-operative labour. This double determination reveals a homology in the form of the two connections constituting the complex structure of the mode of production: they can both be characterized as the ‘separation’ of the labourer from his means of production. Which amounts to saying that they divide up their ‘supports’ in the same way, that they determine coincident forms of individuality for the labourer, the means of production and the non-labourer. The labourers who are in a relationship of absolute non-ownership with the means of production, constitute a collective in the production process which coincides with the ‘collective labourer’ who can set to work the ‘socialized’ means of production of modern industry, and thereby really appropriate nature (the objects of labour). What is here called ‘real subsumption’ is what Marx introduced in the Preface to A Contribution as a ‘correspondence’ between the relations of production and the level of the productive forces. We can therefore specify the sense in which the term ‘correspondence’ is to be understood. Since the two connections between which there is a homology both belong to the same level, constituting the complexity of the structure of production, this ‘correspondence’ cannot be a relation of translation or reproduction of one by the other (of the form of the productive forces by that of the relations of production): it is not one of the two which is ‘subsumed’ beneath the other, it is labour which is ‘subsumed’ beneath capital, and this subsumption is ‘real’ when it is thus doubly determined. The correspondence therefore lies completely in the unique division of the ‘supports’ of the structure of production and in what I called above the reciprocal limitation of one connection by the other. At the same time, it is clear that this correspondence is in its essence completely different from any ‘correspondence’ between different levels of the social structure: it is established in the structure of one particular level (production) and depends completely on it.

In ‘formal subsumption’, on the other hand, the labourer’s belonging to capital is only determined by his absolute non-ownership of the means of production, but not at all by the form of the productive forces, which are still organized according to craft principles. It seems not impossible that each labourer might return to handicrafts. That is why Marx says that the labourers’ belonging to capital is still ‘accidental’ here:

In the early stages of capital, its command over labour has a purely formal and almost accidental character. The worker at this time only works under capital’s orders because he has sold it his labour-power: he only works for it because he does not have the material means to work on his behalf (Capital, T.II, p. 23; Vol. I, p. 330).

However, this absence of ownership of the means of production for the direct labourer is by no means ‘accidental’: it is the result of the historical process of primitive accumulation. In these conditions, there is not strictly speaking any homology between the forms of the two connections: in manufacture the means of production continue to be set to work by individuals in the strict sense, even if their component products have to be assembled to constitute a useful object on the market. We can therefore say that the form of ‘complexity’ of the mode of production may be either the correspondence or the non-correspondence of the two connections, of the productive forces and the relations of production. In the form of non-correspondence, which is that of the phases of transition such as manufacture, the relationship between the two connections no longer takes the form of a reciprocal limitation, but becomes the transformation of the one by the effect of the other: this is shown by the whole analysis of manufacture and the industrial revolution, in which the capitalist nature of the relations of production (the necessity of creating surplus-value in the form of relative surplus-value) determines and governs the transition of the productive forces to their specifically capitalist form (the industrial revolution arises as a method of formation of relative surplus-value beyond any predetermined quantitative limit). The ‘reproduction’ of this specific complexity is the reproduction of this effect of the one connection on the other.

It thus seems that, neither in the case of correspondence, nor in that of non-correspondence, can the relationship between the two connections ever be analysed in terms of a transposition or translation (even a distorted one) of one into the other, but only in terms of an effectivity and a mode of effectivity. In one case we are dealing with the reciprocal limitation of the effectivities of the two connections, in the other with the transformation of one by the effectivity of the other:

We now see that a certain minimum amount of capital in the hands of individuals is the concentration of wealth necessitated for the transformation of individual labour into combined, social labour; it becomes the material base for the changes which the mode of production will undergo’ [here ‘mode of production’ should be understood in the restricted sense of ‘form of the productive forces’] (Capital, T.II, p. 23; Vol. I, p. 330)

What has occasionally been called the ‘law of correspondence’ between the productive forces and the relations of production should therefore rather be named, as Charles Bettelheim has proposed, ‘the law of necessary correspondence or non-correspondence between the relations of production and the character of the productive forces’ (‘Les cadres socio-économiques et l’organization de la planification sociale’, Problèmes de Planification, V, École Pratique des Hautes Études, Paris 1965). This would express the fact that the ‘law of correspondence’ has as its peculiar object the determination of effects within the structure of production and the varying mode of this determination, and not a connection of expression which would merely be the inverse of a mechanical causality.

The mode of ‘correspondence’ between the different levels of the social structure, which has more strictly been called the mode of articulation of these levels, depends in turn on the form of the internal correspondence of the structure of production. We have already encountered this articulation above in two forms: on the one hand, in the determination of the determinant ‘last instance’ in the social structure, which depends on the combination peculiar to the mode of production considered; on the other, with respect to the form of the productive forces peculiar to capital and to the mode of intervention of science in their history, as the determination of the limits within which the effect of one practice can modify another practice from which it is relatively autonomous. Thus the mode of intervention of science in the practice of economic production is determined by the peculiar new form of the ‘productive forces’ (unity of means and object of labour). The particular form of correspondence depends on the structure of the two practices (practice of production, theoretical practice): here it takes the form of the application of the science, in conditions determined by the economic structure.

We can generalize this kind of relationship between two relatively autonomous instances; it recurs, for example, in the relationship between economic practice and political practice, in the forms of class struggle, law and the State. Marx’s indications here are much more precise, although Capital does not contain any theory of the class struggle as such, or of law or the State. Here, too, the correspondence is analysed as the mode of intervention of one practice within limits determined by another. This is the case with the intervention of the class struggle within limits determined by the economic structure: in the chapters on the working day and on wages, Marx shows us that the sizes of these are subject to a variation which is not determined in the structure and depends purely and simply on the balance of forces. But the variation only takes place between certain limits (Grenzen) which are set by the structure: it thus possesses only a relative autonomy. The same is true of the intervention of law and of the State in economic practice, which Marx analyses in the example of factory legislation: the State intervention is doubly determined, by its generalized form, which depends on the particular structure of the law, and by its effects, which are dictated by the necessities of economic practice itself (family and education laws govern child labour, etc.).

In this case, too, there is therefore no relationship of simple transposition, translation or expression between the various instances of the social structure.

Notes for Part 3.

1. Grundrisse der Kritik der politischen Ökonomie (Rohentwurf 1857-8), Dietz Verlag, Berlin 1953. Notable among these manuscripts is the one called Formen, die der kapitalistischen Produktion vorhergehen, pp. 375-413. References below are to this text and to the English translation by Jack Cohen, edited and introduced by E. J. Hobsbawm, Pre-Capitalist Economic Formations, Lawrence and Wishart, London 1964.

2. Lenin: ‘What the Friends of the People Really Are and How They Fight the Social-Democrats’, Collected Works, Vol. 1.

3. Louis Althusser: ‘A Complementary Note on Real Humanism’, For Marx, pp. 242-7.

4. ‘Artificial but not arbitrary.’ Here I have adopted Auguste Comte’s very words in the Cours de philosophie positive (First Lecture, Vol. I, p. 24) about the division of science into several branches. The problem of the ‘break’ between the different states of a single science is of the same nature: ‘It is impossible to assign a precise origin to this revolution . . . It is constantly more and more complete. . . . However, . . . it is convenient to fix an epoch in order to prevent our ideas from straying’ (ibid., p. 10). Bacon, Descartes and Galileo thus determine the transition of physics to positivity, and at the same time the beginning of the general preponderance of the positive state. With his double articulation of the sciences and the law of the three states, Comte is the most rigorous thinker so far of this general theoretical problem: how the distinct practices which constitute a ‘division of labour’ are articulated together, and how this articulation varies with the mutations in these practices (‘breaks’).

5. Here we should note a serious difficulty for our reading, not only where the Contribution is concerned, but also Capital : the term ‘social formation’ which Marx uses, may be either an empirical concept designating the object of a concrete analysis, i.e., an existence : England in 1860, France in 1870, Russia in 1917, etc., or else an abstract concept replacing the ideological notion of ‘society’ and designating the object of the science of history insofar as it is a totality of instances articulated on the basis of a determinate mode of production. This ambiguity includes, first, philosophical problems of a theory of science and of the concept, which are not explicitly solved, and the empiricist tendency to think the theoretical object of an abstract science as a mere ‘model’ of existing realities (see Althusser’s paper on this point). But, secondly, it also includes an objective omission from historical materialism itself, which can only be imputed to the inevitably gradual character of its development: Capital, which expounds the abstract theory of the capitalist mode of production, does not undertake to analyse concrete social formations which generally contain several different modes of production, whose laws of coexistence and hierarchy must therefore be studied. The problem is only implicitly and partially contained in the analysis of ground rent (Volume Three); it is only present practically in Marx’s historical and political works (The Eighteenth Brumaire, etc.); Lenin alone, in The Development of Capitalism in Russia and the works of the period of the transition to socialism, begins to treat this problem theoretically.

And we should also note that the insufficient elaboration, in this first draft, of the concepts which designate the articulation of the instances of the social formation, is in itself the (negative) cause of a constant confusion in Marxist literature between the social formation and its economic infra-structure (which is itself often related to one mode of production). Many of the contemporary discussions of non-capitalist or pre-capitalist modes of production bear witness to this.

6. Periodization, thought of as the periodization of the modes of production themselves, in their purity, first gives form to the theory of history. Thus the majority of the indications in which Marx assembles the elements of his definition are comparative indications. But behind this descriptive terminology (men do not produce in the same way in the different historical modes of production, capitalism does not contain the universal nature of economic relations), there is the indication of what makes the comparisons possible at the level of the structures, the search for the invariant determinations (for the ‘common features’) of ‘production in general’, which does not exist historically, but whose variants are represented by all the historical modes of production (of the 1857 Introduction to A Contribution ).

7. Louis Althusser: ‘On the Materialist Dialectic’, For Marx, op. cit., Chapter 6.

8. It is not my aim to give a theory of ‘fetishism’, i.e., of the ideological effects directly implied by the economic structure, nor even to examine in detail what Marx himself tells us about it, but merely to retain and use the index he provides by explicitly linking the problem of fetishism with that of the place of the economy in the structure of various social formations.

9. First of all, since it is always necessary at the theoretical level to begin with what is determinant ‘in the last instance’. The reason is clear: the very names of the problems depend on it.

10. Pierre Vilar writes of the feudal mode of production: ‘In general, growth seems to depend on a re-occupation of waste lands, on an investment in labour rather than in capital, and the owning classes’ levy on production is legal and not economic ‘ (Première Conférence Internationale d’Histoire Économique, Stockholm 1960, p. 36). To this point we should add the oft-repeated comment that it is difficult to find specifically economic crises outside capitalism.

11. On this point, see particularly the work of Claude Meillassoux: ‘Essai d’interprétation des phénomènes économiques dans les sociétés d’auto-subsistence’, Cahiers d’Études Africaines, 1960, No. 4; Anthropologie Économique des Gouro de Cote d’Ivoire, Mouton, The Hague, 1964.

12. Obviously, we are here using a general concept of ‘manual labour’, one not restricted to actions performed by the hands, although the hands are the dominant organs, but extended to the world of the whole psycho-physiological organism. Similarly, ‘machine’ should not be understood in the restricted sense of machines which are mechanical.

13. In the text of Capital, this determination is followed by a second one, which notes that in the capitalist mode of production the description ‘productive labourer’ is at the same time restricted to the wage labourer, the labourer who corresponds to an advance of variable capital for the capitalist. These two inverse movements (extension-limitation) are not mutually exclusive or contradictory. Each corresponds to one of the two internal connexions of the mode of production, or more exactly to the determination of one element -- the direct labourer -- with respect to each of the two connexions, according to the specific form that the latter take in the capitalist mode of production. In the one that we have taken as the object of our study, the element (the labourer) which has the ability actually to set to work the social means of production is constituted not only by wage labourers and non-wage labourers (intellectual workers), but also by the capitalists themselves, insofar as they have the technical function of supervision and organization. The same double movement (extension- limitation) will recur later in this exposition, when I analyse the specific types of development of the productive forces in the capitalist mode of production and the historical tendency of that mode of production.

14. The function of ownership of the means of production may be performed by individuals, collectivities, real or imaginary representatives of the collectivity, etc.; it may appear in a unique form, or, on the contrary, be duplicated -- ‘property’ and ‘possession’, etc.

15. ‘The means of labour acquire in mechanization a material form of existence (materielle Existenzweise ) which is the condition for the substitution of natural forces for human force, and the conscious application of science instead of empirical routine’ (Capital, T.II, p. 71; Vol. I, p. 386).

16. References to the ‘Three Essays on the Theory of Sexuality’ are to The Standard Edition of the Complete Works of Sigmund Freud, translated and edited by James Strachey, Vol. VII, London 1953.

17. In reality, these questions are necessarily posed to any theory of development, notably in its original domain: the biological (whether individuals or species are concerned). The Darwinian revolution can be situated in a history of theories of development as a new way of posing them, which introduces a new answer (‘evolution’, restricted to the species and distinct from individual development). On this point, it has been possible to write: ‘Originally such a development was understood as applying to a unique and qualified individual. No doubt, around the middle of the [nineteenth] century, it became hard to tell what was the subject of this development (what developed). This invariant behind the embryological transformations could not be assimilated to surface and volume (as in an unfolding), nor to the adult structure (as in a maturation) ... Other than [a] pseudo-unity in instantaneity (ecological, etc.), the only universe left for Darwin was a unity in a succession reduced almost to a minimum: that of a continuous lineage (filiation ), both in the genealogical sense (all species deriving from the same stock) and in an almost mathematical sense (tiny elementary variations) This lineage explained the relative persistence of types and plans of organization: it was not the substratum or foundation of the history: it was merely a consequence of it ‘ (G. Canguilhem, G. Lapassade, J. Piquemal and J. Ulman: ‘Du développement à l’évolution au XIXe siècle’, Thales, T.XI, 1962). In Freudian (and Marxist) pseudo-development, we do not even find this minimum - we are dealing with the radical absence of any pre-existing unity, i.e., any germ or origin.

18. Althusser has proposed the term ‘technical relations of production’, which clearly marks the distinction. But we should remember that ‘relations’ in itself implies their social character.

19. Particularly in La Nasssance de la Clinique, op. cit.

20. ‘The capitalist imagines that he is consuming the surplus-value, and is keeping intact the capital-value; but what he thinks cannot change the fact that after a certain period, the capital-value he then possesses is equal to the sum total of the surplus-value that he has acquired for nothing during this period, and the value he has consumed is equal to that which he advanced. Of the old capital which he advanced out of his own pocket, not an atom remains. It is true, he has in hand a capital whose amount has not changed, and of which a part, viz., the buildings, machinery, etc., were already there when the work of his business began. But what we have to do with here, is not the material elements, but the value, of that capital’ (Capital, T.III, pp. 12-13; Vol. I, pp. 569-70).

21. In Volume One, Marx defines them in their concept (but not in all their effects) by the analysis of the abstract object which he calls a ‘fraction of the total social capital promoted into autonomy’ (Capital, Vol. II, p. 353). By which we are obviously to understand, as Establet notes (Lire le Capital, first edition, Vol. II, p. 343), not a real firm or enterprise which is capitalist in form, but a fictive capital which is necessarily a productive capital and yet carries out all the functions historically assumed by different types of ‘capital’ (merchant’s capital, interest-bearing capital, etc.). The division of social capital is an essential property: it is therefore possible to represent capital in general by one capital.

For their part, only the analyses of reproduction in Volume Two, Part 3 (‘Reproduction and Circulation of the Aggregate Social Capital’), which make way for the establishment of schemes of reproduction, and thus allow the mathematical formalization of economic analysis, explain by what mechanism the reproduction of the social relations is assured, by subjecting the qualitative and quantitative composition of the total social product to invariable conditions. But these structural conditions are not specific to the capitalist mode of production: in their theoretical form they imply no reference to the social form of the production process, to the form of the product (‘value’), to the type of circulation of the social product which it implies (‘exchange’), or to the concrete space which supports this circulation (‘market’). On this point, I refer the reader in particular to the various recent works of Charles Bettelheim, and to his critical comments in Problèmes de la Planification, No. 9 (École Pratique des Hautes Études) -- Note of 1967.

22. Lire le Capital, first edition, Vol. I.

23. These limits must not be confused with the limits of variation (Grenzen) which we discussed above.

24. Not even the time of economic history, of course, if by that is meant the relatively autonomous history of the economic base of the mode of production. This is for two main reasons: firstly, such a history, dealing as it does with concrete-real social formations, always studies economic structures dominated by several modes of production. It therefore has nothing to do with the ‘tendencies’ determined by the theoretical analysis of isolated modes of production, but with the compounded effects of several tendencies. This considerable problem lies outside the field of the present analysis, and it is only touched on incompletely in the next section (on the ‘phases of transition’). Secondly, the ‘age’ of production which we are discussing here is not, clearly, a chronological feature, it does not indicate how old capitalist production is: for it is an age compared between several economic zones (or ‘markets’) subject to the capitalist mode of production, which is important because of the effects which lead to an unevenness in the organic composition of capital from one region to another or from one department to another. According to the closeness of the analysis, it will be a matter of an average organic composition or of a differentiated analysis of the organic composition of capital from branch of production to branch of production: this is the beginning of a study of the effects of domination and uneven development implied by the unevenness of the organic composition between competing capitals. Obviously, this is not our object here. I am only suggesting it as a possibility.