Self Criticism: 
Answer to Questions 
from 
ŒTheoretical Practice¹ 
Etienne Balibar 
Theoretical Practice 7-8, 1973 
pp. 56-72 
© Digital Reprints 
May 2002

ETIENNE BALIBAR 
Allow me to thank you for your letter and for the interest 
you have brought to my text of 1965. I am particularly grateful 
to you for having so vigorously emphasized a number of the 
weaknesses or contradictions that appeal in it. It is the 
greatest service I could have wished for, and one that 
experience shows is not always easy to obtain. I shall try 
to answer you on the main points, not because I wish to 
'defend' my text, which certainly is confused or wrong in 
several essential points, but because these very errors are 
highly instructive. 
I. ON THE THEORY OF FETISHISM 
Self Criticism: An 
Answer to Questions 
from ŒTheoretical 
Practice¹ 
To get things clear, three aspects of the problem must be 
very carefully distinguished: 
­ what Marx thought about these two themes, which should 
be considered separately; 
­ What I was trying to do in the passage from Reading 
Capital: 
­ lastly what we should think about these themes, or the 
questions they suggest, in the current state of the problematic 
of historical materialism. 
Just a few words on the question of fetishism. 
In fact, in my text in Reading Capital, the question of 
fetishism has an accessory role: I simply wanted to use 
Marx's formulations as indices enabling me to ascend to 
the structural characteristics of 'determination in the last 
instance'. Thus I certainly did not think that the examination 
of the latter point had necessarily to proceed via a 
theory of fetishism. I made use of the fact that, precisely 
in the section of Capital on the 'fetishism of commodities', 
Marx proceeds to draw up a comparative table of the manifestation 
of social relations in different real or even merely 
possible modes of production. I did not intend to include 
the phenomenon of 'fetishism' in the very mechanism of 
'determination in the last instance'. 
On reflection, and here you are perfectly right not to be 
satisfied by such an empirical position, this proximity is, 
however, not completely the result of accident in Marx 
himself. Nor in consequence the considerable part played 
by the whole question of fetishism in the history of investigations 
into the dialectic after Marx. It is not by accident 
that it intervenes simultaneously with a characterization 
I broadly subscribe to your criticism of my formulations on 
'fetishism' and 'determination in the last instance by the 
economic (Chapter 1, Section 3 in my text in Reading Capital). 
It is indisputably a bad passage although on one essential 
point I draw opposite conclusions front this to your own. 
56

effect (i.e., an affect of allusion/illusion, of recognition/ 
misrecognition in social practice) can only be explained by 
a positive cause, the existence and functioning of ideological 
social relations historically constituted in the class 
struggle. Specific social relations really distinct from 
the relations of production although they are determined 
by the latter 'in the last instance'. Really distinct means 
realised, materialised in specific practices, depending on 
special ideological apparatuses, etc.[1] Of course, such 
a theoretical representation finds its verification in practice, 
the practice of the class struggle, in which the proletariat 
discovers the existence, the necessity of ideological 
social relations, the necessity and the means to 
transform them. Moreover, such a representation is 
already outlined in Marx, not so much in Capital, but as a 
function of certain conjunctures of the political struggle 
(cf. already the third section of the Communist Manifesto). 
But it does not begin to take a general form (and it cannot 
become a theory strictly speaking) until Mao Tse-tung or 
his period (on the practical basis of the 'cultural revolution'). 
The 'theory' of fetishism has a completely different character: 
­ on the one hand it makes the misrecognition/recognition 
a 'structural effect' (or 'formal effect') of the circulation 
of commodities, a (subjective) effect of the place occupied 
by individuals in the structure of exchange with respect to 
the commodity. 
­ on the other hand it makes the commodity itself the 
'object' of this misrecognition (in so far as the 'substance' 
of value is social labour) the source or subject of its own 
of the capitalist mode of production as a historical structure 
compared with other possible ones in a kind of typology. 
And first of all because it is essential to the definition 
of the 'economic' in its relationship with the 'commodity-
form' (the 'commodity categories'): i.e., insofar 
as the 'representation' of the economic is essential to the 
economic itself, to its real functioning and hence to its 
conceptual definition. 
The theory of 'fetishism' (and with it all the theory of Part 
I of Capital Volume One) thus really is the index of a 
fundamental problem, one which will only become clear 
when we go further into the historical study of the different 
aspects of the reproduction of capitalist relations of production 
(since the commodity-form is realised at the level 
of the circulation of the products of labour and of the functioning 
of the legal and ideological 'superstructures'). 
Only, as a theory (I insist on this specification), it is 
totally idealist. On this particular, but decisive point, the 
rupture with idealism has not taken place. In fact there has 
only been a change in the form of idealism, the discovery 
of a form which is, certainly, 'critical' and has played 
a necessary part in the process of constitution of historical 
materialism, but which remains ideological and 
thereby demonstrates to us the dialectical, i.e., contradictory, 
uneven and uninterrupted character of this 
process, as is the case for every scientific theory. 
Why totally idealist? Because it prevents a materialist 
theory of ideology, it is an obstacle to it precisely where 
that theory is required: in the explanation of an ideological 
effect. As we are now beginning to realise, an ideological 
57

misrecognition, which results from the 'auto-development' 
of its-form. This is a direct consequence of the way, 
throughout Part One, Marx has 'logically' developed the 
abstract (universal) and simultaneously concrete (immediately 
present in 'no matter what' everyday exchange of 
'no matter what' product of labour) form of the commodity, 
represented as a subject. 
It is thus a genesis of the subject, comparable to others 
that can be found in classical philosophy but with the 
following 'critical' variant: it is a genesis of the subject 
as an 'alienated' subject (a genesis or theory of cognition 
as misrecognition). 
That is why, after having been stated by Marx in a Hegelian- 
Feuerbachian problematic, it has been possible for this 
theory to be adopted and developed enthusiastically in a 
structuralist or formalist problematic (as in Godelier, the 
editors of Cahiers pour l'Analyse, etc. ). For 'structuralism' 
is the strict theoretical equivalent of this Hegel- 
Feuerbach combination, very precisely, as Althusser 
suggests, 'Hegel in Feuerbach', elaborated by Marx at 
the time of the constitution of historical materialism ­ 
1844-6 ­ and which in this case he did not renounce in this 
philosophical combination, 'Hegelianism' does mean 
process, but process of the manifestation of a subject, in 
this case an alienated subject ­ in Feuerbach's sense ­ in 
which the 'real' relation between essence and attribute is 
'inverted'. That is why ultimately structuralism equals 
humanism: for the question of the (structural) place is 
equivalent to the question of the (human) subject, if the fact 
of occupying a place in the system of social relations in- 
58 
stitutes in addition a point of view, a representation, a 
consciousness of this system, and explains it of itself alone. 
In consequence, not only does the theory of the fetishism 
of commodities prevent the scientific explanation of the 
special ideological effects implied by commodity circulation, 
it also prevents our thinking their revolutionary transformation: 
it makes it seem that the 'transparency' of social 
relations is an automatic effect (even if not immediately, 
which is anyway not very easy to explain) of the suppression 
of 'commodity categories', i.e., of the commodity. 
It is a theory of ideology in general, of the historically 
transitory role of ideology in general: if it is to be 
believed, one fine day not only will there no longer be class 
ideologies, but, as there will no longer be any commodities, 
there will no longer be any ideology at all. Alienation, 
then suppression of alienation. 
I do not think that fundamentally one leaves this ideological 
circle by replacing the structure of the 'commodity form' 
with the more general structure of systematically varied 
modes of production, i.e., the 'place' in commodity 
circulation by the place in the structure of the 'whole' and 
with respect to that whole, and by thus introducing the 
possibility of allowing the point of application of 'fetishism' 
to vary. For what then remains unintelligible (and fundamentally 
useless) is a social practice of the material transformation 
of ideological relations (as a specific revolutionary 
practice), and hence the distinct reality of these 
relations. If the effect of illusion is the effect for the individual 
of the place in the 'whole' that constitutes him as 
a subject, then the lifting of the illusion is still no more

you like, and teleological, i.e., pre-determined, despite 
all the 'materialist inversions' you like), it is the real 
dialectic of the 'class struggle', the material forms of 
which are irreducible to the form of linear development, 
progress and teleology. It is thus the dialectic of the 
different aspects of the class struggle, really distinct from 
one another in their unity, as the practice of the workers' 
movement teaches (and not apparently distinct, like an 
essence' and its 'phenomenon' or 'phenomena'). The 
economic aspect (the 'economic' class struggle) is only 
one of these aspects, unevenly developed, unevenly decisive 
according to historical conjunctures, and never 
capable of producing revolutionary effects by itself. Which 
by no means rules out, but on the contrary demands, that 
in all historical periods, whatever the dominant mode of 
production and whatever the conjuncture, the ensemble of 
the class struggle is still determined by its economic 
material conditions. For social classes themselves, or 
rather the class struggle in and by which alone classes 
exist, have no historical reality except as presuppositions 
and results of the process of material production and reproduction 
of the material conditions of production Thus 
to define and study, for each historical period, the specific 
way in which each really distinct aspect of the class strugsle 
('economic', 'political' and 'ideological') depends on 
its material conditions, is precisely the object of historical 
materialism. 
I insist on the role of this text of Althusser's (and of the 
next 'On the Materialist Dialectic' which complements 
it), for, reading it closely, one can derive from it an 
essential (perfectly correct) thesis: the only real histhan 
a subjective, individual matter, however much it is 
socially conditioned by the structure of the whole, and however 
much it is repeated 'millions of times over' for 
millions of individuals occupying similar places it is only 
the effect of a different place or of a coming to consciousness 
in one place 
In other words, the theory of fetishism can never truly think 
that subject is an ideological notion (elaborated first of all 
within juridical ideology). On the contrary, it seems to 
make the notion of 'subject' the scientific' concept of 
ideology. Assuming that these schematic suggestions are 
correct. Iit is on this point, too, that Althusser's text 
cited above should be examined, for in it the problem 
is not perhaps solved in an absolutely clear fashion. 
II. DETERMINATION IN THE LAST INSTANCE 
I now return to the question of 'determination in the last 
instance' and to your question. About my text in Reading 
Capital, the following can be schematically said the direct 
object of the section in question was not fetishism, but it 
was 'determination'. And the leading idea was to take up 
and generalize an argument already expounded by Althusser, 
in his article 'Contradiction and Overdetermination' (in 
For Marx) on the subject of the historical conjuncture. This 
article is the source of my terminology: determination, 
domination, displacement of domination. In this reference 
there is something correct that we should try to retain: 
this is the fact that Althusser's text, however provisional 
the character of some of its formulations, does show the 
following: the 'dialectic' of history is not the pseudodialectic 
of development (linear, despite all the 'negations' 
59

graphy of the instances of the 'complex' social 'whole'), 
but also it transforms anew the object whose dialectic 
this topography should make it possible to analyse. 
Instead of social formations, it is now (and anew) a 
question only of modes of production, i.e., of a still 
'abstract' generality, of which, in practice, the social 
formations will appear only as particular and concrete 
'realizations'. 
Is there a 'General Theory of Modes of Production' ? 
First of all, indeed, the idea emerges that the theory of 
modes of production itself derives from a general theory 
of modes of production, which can only be a theory of 
the mode of production in general and of its possible 
'variations': in short, a theory of typologistic or 
structuralist inspiration, however consistent. 
Naturally, this idea of 'theory' (this temptation of theory) 
must be taken in its strongest sense: i.e., in the sense 
of a theory providing real explanations of real history 
(so as not to revert to a positivist position). It must be 
understood that in such a perspective, the variation 
(the varied combination) of the play of the 'elements' is 
capable by itself of explaining historical effects. Such is 
undoubtedly one of the main ambiguities of the highly 
unsatisfactory expression 'structural causality'. The 
article by Badiou to which you refer ('Le (Re)commencement 
du Matérialisme Dialectique,' Critique, May 1967) 
avoids a whole series of intellectual confusions or slides 
torical dialectic is the process of transformation of each 
concrete 'social formation' (implying the real interdependence 
of the different social formations); in other words, 
the 'social formations' are not the 'concrete' site (or 
environment) in which a general, abstract dialectic is 
'realised' (for example the transition from capitalism in 
general to socialism in general, or from one stage to 
another in general in the 'development' of capitalism), in 
reality they are the only object that is transformed, because 
they are the only one that really contains a history 
of class struggles. This point is decisive. I add that it is 
by no means an accident that Althusser was able to advance 
in this direction on the basis of an analysis at the texts 
and practice of Lenin, for not only is Lenin clearer on 
this point than Marx, he even makes a true and as time 
passed more and more conscious rectification of certain 
of Marx's formulations. It is on this basis that we 
in our turn must take up, develop and possibly rectify 
the whole ensemble of the theory of historical materialism. 
I shall return to this. 
To get back to my text, it is clear that in my attempt 
to 'generalize' Althusser's idea, I incautiously changed 
its point of application: what he had used to deal with the 
historical 'conjuncture', I applied to the comparison of 
modes of production; I made the variation or displacement 
of the 'dominant' the principle of a comparative 
analysis (even a theory) of the forms (even the types of 
modes of production. This displacement has serious 
consequences: not only does it introduce an ambiguity 
which is from then on found in every use of the 'topographical' 
concepts introduced by Althusser (the topo- 
60

idea is confused from the standpoint of historical 
materialism (in which it is useless: historical materialism 
cannot but encounter relatively persistent social forms 
­ e.g., Roman' law, money, certain aspects of the 
State apparatus, etc. ­ despite the transformation of the 
dominant mode of production, and it will then try to 
explain such a persistence by the very characteristics 
of the unique historical process in which the mode of 
production has changed). From a theological or structuralist 
viewpoint, on the contrary, it is perfectly 
logical, or at least can hope to obtain a precise formulation. 
In this vein, the idea of analogy appears in my 
text not only vis-à-vis the place and nature of 'fetishism' 
but also vis-à-vis the structure of the modes of production 
as a whole (cf. the rather risky comparison 
between the forms of the appropriation of the surplus 
product in the capitalist mode of production and in the 
so-called 'Asiatic' mode of production, the craze of 
French Marxism at the time). 
Do not Confuse the Real Object and the Formal Concepts 
of its Analysis 
But there is something more basic and serious: from 
such a viewpoint, the very denomination of the 'instances' 
in the social formation cannot but tend to designate anew 
essentially invariant elements of historical analysis, 
contradicting what was postulated at the beginning of that 
part of my text. In clear language, this means that there 
is an essence of 'economic' phenomena and also essences 
because it takes this tendency consistently as far as it 
will go, quite rightly thinking that a formalist theory of 
this type should be capable of formalization. At least 
this extreme attempt has the advantage of showing how 
one might (and how one obviously should not) transform 
historical materialism into a totally idealist 'theory of 
historical sets'.[2] But the same thing is just as present, 
though less apparent, in others (not to speak of myself) 
who discuss 'concrete' historical examples, for example 
Poulantzas. Each time the principle is the same: to 
endow the 'topographic' characterizations of the social 
formation with a real historical effectivity: for example, 
to explain the functioning of the capitalist State and its 
role in the reproduction of the relations of production 
by its 'relative autonomy' as an instance or else to 
explain the transformation of the social formation insofar 
as it is a 'transition from one mode of production to 
another' by the very 'displacement' or variation in the 
relationship of the instances which serves to compare 
them (to define them differentially). Or even, if this were 
not enough, to repeat this operation with elements 'of a 
higher order' ­ the modes of production themselves ­ 
merely by thinking the social formations as 'complex 
wholes in dominance' of several modes of production. In 
short, the idea of structural difference, suitably 
elaborated, is supposed to function ipso facto as the 
origin of the historical differentiation: a formula which, 
I think, shows well enough the idealism of the 'theory' 
in question. 
You are quite right to emphasize and criticize the 
appearance in this chapter of the idea of analogy. This 
61

of 'political' and 'ideological' phenomena, pre-existing 
the process of their historical transformation (and thus 
of their historical definition). In other words, preexisting 
the process of the determination by the action of 
the class struggle, itself having a determinant overall 
historical structure. That means that at a sufficient 
level of generality, but one capable nonetheless of 
explaining a historical causality, and definite effects, 
'economic' has the same meaning in the feudal mode of 
production and in the capitalist mode of production, and 
in fact in any mode of production. In short, it is a return 
to the ideological presuppositions of bourgeois political 
economy and history, in the very sense that 
Marx called it 'metaphysical'. There is no doubt that 
this temptation was induced in our work by our concern 
to avoid any 'historicist' interpretation of Marx's 
criticism, and in consequence, in Lenin's words, to 
'bend the stick in the other direction'. But the stick 
cannot be bent indiscriminately, or, if you like, the 
space in which it is bent is not a mere plane. Of course, 
this relapse is no accident, and I think that I am able to 
state that, in this and other analogous forms, it is the 
index of a real difficulty. I shall return to it. 
Finally, the result of all this is that the section in question 
here partly contradicts the theses of the two preceding 
divisions ('Mode of Production: Manner of Producing' 
and 'The Elements of the System of Forms') and of the 
following Chapter Two ('The Elements of the Structure 
and their History'). Or at any rate, it helps to orient 
their interpretation in a direction which is not the only 
possible one, and precisely not the right one. As I 
62 
recalled a moment ago, my own text suggests that the 
'general' concepts it is dealing with are not 'the atoms of 
a history', the given elements out of which varying 
combinations are to be constructed, but only the 
'pertinent' categories of the differential analysis of social 
forms. These concepts only indicate and in some sense 
formally orient the general problematic (I say problematic 
and not theory) of 'historical materialism' at work 
in certain definite theoretical analyses of Marx's. They 
cannot anticipate their content. Logically this means that 
at most I can suggest the following, when the (social) 
form of the combination that characterizes the mode of 
production in the strict sense (a combination of determinate 
'relations of production' and 'productive forces') 
changes, then the conditions in which an 'economic', 
'political' or ideological' instance intervenes historically, 
i.e., the conditions in which effects, themselves combined, 
of specifically 'economic', 'political' and 'ideological' 
class struggles are constituted and occur, necessarily 
change too in a determinate manner. That is why, in 
opposition to all economism, the concept of mode of production 
in Marx, even at an abstract level, really 
designates a complex unity of determinations that derive 
from the base and from the superstructure. But we can 
in no way deduce the mode of this constitution, the process 
of functioning and the historical tendencies of the 
social relations considered, nor the laws of the combination 
of the different aspects of the class struggle, 
from the mere presentation of this combination in its 
formal characteristics, i.e., on the basis of a comparison 
between the different possible forms. That is why it is

impossible to invent 'possible' historical modes of production. 
To the question, what is responsible for a theoretical 
slide on this point (which I made myself), it can be 
answered that it is notably the double sense in which the 
term 'combination' (Verbindung) is taken here, on two 
quite different levels. 
In a first sense, it has to be said that the object of historical 
materialism when it is analysing a determinate 
mode of production is first to define and explain a particular 
combination (better: process of combination) of 
the social 'factors' of production, which can be described 
as a 'combination of the relations of production and the 
productive forces' so long as it is pointed out - better 
than I did at this time - that this combination is always 
made, on a given basis, in the (social) form and under 
the influence of the relations of production themselves. 
In other words, that the 'productive forces', although 
they have to be distinguished from the relations of 
production to which they cannot be reduce, do not however 
exist as such (as a system of the material transformation 
and appropriation of nature) except under the 
influence of their own combination with (in) determinate 
relations of production.[3] Such, abstractly and briefly, 
is the object of Capital, notably in Volume One. 
But alongside this first sense there is a second, quite 
different one: this is the idea that the theory of historical 
materialism proceeds, in different circumstances, as 
much vis-à-vis the mode of production itself, as, later, 
is-à-vis concrete social formations, via a combination 
63 
of distinct aspects ('a synthesis of many determinations' 
said Marx). Strictly speaking, this last formulation, if 
it is correct, can only be provisional, precisely because 
this formal characteristic of the theory is not independent 
of the characteristics of the material object of which it 
gives us the knowledge. At any rate the too senses 
cannot be confused without playing on words, or rather 
without surreptitiously moving from historical materialism 
to a kind of 'meta-history'. Here too what has to be 
respected is the rigorous distinction between the real 
object and the concept, or object of knowledge, so as to 
stay on the razor's edge, without leering off 'to the left' 
into empiricist or 'to the right' into formalism. 
III. IDEOLOGICAL PROBLEMATIC OR SCIENTIFIC 
PROBLEMATIC OF 'PERIODIZATION' 
But by saying this I am touching on something much more 
profound that is capable of clarifying us as to the root, 
in the history of Marxism, of the preceding confusions 
and difficulties. You are quite right to stress that the 
concept of 'mode of production' as used in two practically 
distinct senses, as a historical 'unit of periodization' 
and as a 'particular combination of elements', too senses 
which seem to be confused later. One might say squarely: 
the only object of part of the text was precisely to 
identify the two senses, deliberately 
This position should be modified. But, to my mind, not 
at all for the reason you suggest: not because there is

only appears in the fourth chapter, can be said to be to 
show that 'transition' (speaking plainly, social revolution) 
cannot be explained in an evolutionist way, even by translating 
an evolution into the language of the 'negation of 
the negation', 'qualitative change', etc. It is easy to 
understand what practical, political reasons (reasons 
that are more actual than ever) drove us to work in this 
direction and to demonstrate that transition is not, is 
never, for reasons of principle, mere supersession, an 
'internal' results of 'tendencies' observable in the mode 
of production itself and responsible for the development 
of its characteristic relations of production, even if this 
development is simultaneously a development of contradictions.[
4] Hence, from the theoretical point of view, 
it was essential to demonstrate that transition requires 
the analysis of other material conditions and other social 
forms than those implied in the concept of mode of production 
alone (in this case: of the capitalist mode of 
production alone). Or else the analysis of the material 
results and social forms (re)produced by the development 
of the capitalist mode of production in another respect 
than the capitalist relation of production alone. 
But paradoxically, this 'demonstration' consisted 
essentially of accepting for the mode of production itself, 
and at the same stroke, to coin a phrase, for periods 
of 'non-transition', a purely 'internal', linear and 
hence predetermined development or dynamic. In other 
words, if it was essential to recognize in 'transition' 
the characteristics of a history' in the strongest sense 
(unforeseeable in the reality of its concrete forms), 
this was in the event because 'non-transition', for its 
any question of a fallacy, of a confusion between two 
levels of the discourse. If this confusion exists, it is 
only an effect. Besides, the movement of the argument 
was in fact as follows: to show how the construction of 
the concept of 'mode of production' by Marx in fact 
transforms radically the not specifically Marxist problematic 
of historical 'periodization'. Marx thus transforms 
a formal ideological a priori into a scientific 
problem whose solution is precisely the knowledge of 
tendencies implied in definite systems of social relations, 
and of equals definite processes of transition or transformation 
of the social relations. There are therefore 
two notions of 'periodisation' or rather two uses of the 
notion of 'periodization', one ideological and the other 
Marxist and scientific. But that is not the fundamental 
mistake, which lies in the very way 'mode of production' 
is considered as a 'basic concept' of historical 
materialism. 
Matters are especially delicate here: great care must be 
taken not to fall back behind what was correct. And above 
all care must be taken not to fall back behind what is 
indeed in Marx a theoretical revolution on which depends 
the whole construction of historical materialism: the 
definition, vis-à-vis capitalism, of the concept of mode 
of production (material mode of production in the necessary 
form of exploitation) and of its historical tendencies. 
Schematically, let me put it as follows: taken as a whole, 
my text contains a curious paradox. Leaving on one side 
the section devoted to the critical examination of the 
notion of 'productive forces', the main objective, which 
64

of itself so long as the capitalist relations of production 
'remain unchanged', and, on the other, that this tendency 
cannot of itself exceed, 'explode' the very limits of its 
existence. 
Behind this 'argument' there is an old philosophical 
representation, and it is no accident that throughout this 
work I was guided approximately by certain reminiscences 
of Spinozist formulae.[5] There is the idea that 
identity with itself, persistance (including in the form 
of the persistance of relations implied in a cyclical 
process) needs no explanation since it explains itself 
by itself, needs no cause (or production) since it is its 
own cause. Only 'change', as 'real' change, i.e., abolitiontransformation 
of the essence, could need a cause and 
an explanation. Let me say that this is a survival of the 
philosophy of the 'principle of inertia', of substance and 
the ontological argument. . . . But what also explains my 
'relapse' is the power of an old economic idea, an old 
idea of the economists, which enabled them to define 
their object as a set of natural laws, against narrowly 
'political' and institutional representations, and which 
is preserved even in the 'Marxist' thesis of the so-called 
autonomy of the 'economic' process (with respect to the 
'rest' of the social practices, institutions, etc.).[6] I mean 
the old idea that the economic process can be assimilated 
to an automatic, 'self-regulating' mechanism, assuming, 
of course, that it remains within the 'natural' limits of 
its functioning. An idea that the economists have 
attempted to verify, with temporary successes, at the 
level of the market, of price equilibrium, etc. But it 
must also be said an 'economistic' idea which Marx 
part, was not to be a history (in the strongest sense). 
Let me say that this conception inevitably remained 
captive to the very ideology that underlies the ordinary 
practice of 'periodization' which, according to my initial 
project, was what was to be superseded. Indeed, it 
comes down to identifying the notions of history and 
'transition'; simply, instead of saying: everything is 
always transition or in transition since everything is 
historical (ordinary historicism), I said more or less: 
there is only real history if there is (revolutionary) 
transition, and every period is not a period of transition. 
Which, let it be said in passing, is a fine example of the 
setting to work of the empiricist-linear representation 
of time as an a priori form presupposed by periodization. 
The Persistent Ambiguity of the Concept of 'Reproduction' 
But above all this means that, despite certain epistemological 
subtleties, I did not succeed in escaping from 
a basic ambiguity in the notion of the 'reproduction' of 
social relations (of course, I was neither the first or the 
last in this). I still thought within this concept both the 
social form of the (re)production of the conditions of 
production modified and partly destroyed by production 
itself, and on the other hand the identity with themselves, 
the persistence of the given relations of production. To 
sum up, I said: since, in Marx's analysis, the tendency 
towards the accumulation of capital (and all the secondary 
tendencies resulting from it, including the tendency for 
the rate of profit to fall) is identical with the process of 
the reproduction of the relations of production itself, 
this means that, on the one hand, this tendency exists 
65

sequence that all the other (non 'economic') social 
processes can only be expressions or inessential 
phenomena of that form. 
Let it be said in passing that a close examination of the 
texts shows convincingly that this residual 'economism' 
of certain isolated formulations of Marx's is directly 
linked to the idea he had of the object of Capital, which 
gave us so many problems from the beginning because of 
its plainly empiricist-formalist character: the idea that 
Capital studies 'the capitalist system in its ideal average', 
i.e., ultimately, the model of the capitalist system. An 
idea that must be totally abandoned in order to think the 
object of historical materialism as the process of transformation 
of concrete social formations, which are 
unique as such and in consequence absolutely incapable 
of appearing as the variants of a single 'model', even if, 
what is very different, the history of social formations 
in the modern period is basically the history of the 
development and of the effects of their transformation 
by a single dominant mode of production, first constituted 
'locally' but necessarily extending itself globally. 
There is no 'General Theory' of Historical Transition 
I return to my text in Reading Capital. In it it is visible 
that the paradoxical attempt I have been discussing, and 
for which, I repeat, Marx (or at least the isolated 
reading of certain of Marx's texts) is partly responsible, 
had one rather 'logical' effect, come to think of it, where 
the 'theory of transition' is concerned. 
After having posed the problem in this way, I had to ask 
seems never quite to have escaped, even when he shifted 
his object from the 'superficial' sphere of the market to 
the sphere of the production and reproduction of the 
conditions of production 'as a whole'. In Capital this is 
revealed by the fact that the 'economic' theory of 
Volume Two, which is the site at which the scientific 
concept of the reproduction af social relations emerges 
from the 'critique of political economy', can and must 
nevertheless, if it is isolated, inevitably appear incorrectly 
as a complete theory of the reproduction of the conditions 
of production 'as a whole'. And as the theory of 
Volume Two does nothing but show how the different 
immediate production processes 'intertwine' at the social 
level and through the intermediary of the commodity 
circulation of their products, this means that the process 
of production does not reproduce only a part (means of 
production, means of consumption) of the conditions of 
production (which is indeed one of Marx's fundamental 
discoveries), but also reproduces the totality of them 
or reproduces them all 'in potentia'. Which is manifestly 
false, if only because the reproduction of the means 
of consumption is not yet, of itself alone, the reproduction 
of labour power (the process of which obligatorily includes 
the practices of the superstructure), but only its preconditional 
basis.[7] In other words, this means that the 
form of the immediate process of production is thus not 
only held to be what materially determines the ensemble 
of the process of reproduction of the social relations, 
but also what constitutes it completely, by the mediation 
of the market (a point with far-reaching consequences 
which unfortunately I cannot develop here, and in con- 
66

prevented Marx from treating the transition from 
capitalism to socialism in the same mode. It is precisely 
this form that thus led him (in Capital, let me make it 
clear) to treat the transition from feudalism to capitalism 
in a 'historical' mode, though incompletely, and the 
transition from capitalism to communism in a 'logical' 
mode, i.e., not to treat it at all. And this even though 
the necessary historical connection between the development 
of capitalism and the proletarian revolution is 
from beginning to end the very object of Marxist theory. 
What, for my part, did I do to remedy this state of 
affairs? I postulated that these two problems were, ought 
to be, formally the same in nature. And as the 'genealogy' 
can only be a provisional theoretical form I sought to 
conceptualize what it involved. Now to explain the 
necessity and the causality of a historical process the 
only concept I had at my disposal was that of 'mode of 
production'. I therefore suggested that the analysis of 
transition consists of the definition of a new mode of 
production, different from the capitalist mode of production 
itself, although 'complex', or 'contradictory', 
and therefore 'unstable' (characterized by a fundamental 
'non-correspondence'). But by this fact not only did I 
cancel out part of my own earlier presuppositions, since, 
logically, a new 'mode of production' cannot be anything 
but a new tendential process of reproduction, like the 
capitalist mode of production itself. But above all, (1) I 
introduced the germ of an insoluble problem: what is the 
specificity of the relations of production defining such a 
mode of production'? ­ a problem that is unavoidable 
once one has recognised the primacy of the relations of 
myself whether Marx himself had, in Capital, broached 
the problems of 'transition', given that it was insufficient 
for its analysis to 'extend' the tendency defined in the body 
of the work. What presented itself naturally enough was 
the set of texts on primitive accumulation, the genesis of 
ground rent, the origins of merchants' capital, etc. I 
have no reason to modify the essential orientation of this 
analysis. It is fundamental to compare these different 
texts and to draw conclusions from their comparison. 
One specification only, since you raise it, concerning 
the expression 'genealogy': obviously I did not mean that 
the transition process was itself a 'genesis' or 
'genealogy' (which anyway would not be very meaningful), 
but that Marx had had, and could not but have had to 
treat this problem in the (provisional) theoretical form 
of a genealogy, i.e., of a series of retrospective 
historical 'soundings' starting from the elements of the 
capitalist mode of production taken one by one. Which 
led me to state simultaneously: 
­ that it is precisely this theoretical form that enabled him 
to discover and expound to us the relative independence and 
real distinctness of the historical processes in which the 
elements of the capitalist structure (labour power as a commodity 
on the one hand and money-capital on the other in the 
hands of a mercantile bourgeoisie) are constituted. Hence the 
fact that the constitution of capitalist relations of production 
is not pre-determined, i.e., that the constitution process 
is not teleological. 
­ and at the same time that it is precisely this form (or 
rather the precondition that imposed it on Marx: his 
incomplete theory of the reproduction of capital) that 
67

burden in the history of Marxism in that they suggest, 
by different means, the existence of a universal 
mechanism of the transformation of social formations, 
deducible from the structural schema of 'the' social 
formation in general (cf. the famous 'narrowness' of the 
envelope of relations of production with respect to the 
development of the productive forces). This comparison 
simultaneously illuminates a basic epistemological fact, 
the necessary solidarity and even interdependence of 
evolutionist representations and relativist representations 
(typological or structuralist representations) of history, 
which seem to be opposed but are both non-dialectical. 
These two representations both arise from the fact of 
posing separately (1) the analysis of the historically 
relative character of a mode of production (the capitalist 
mode of production is not a mode of production of wealth 
'in itself', but only, 'neither more nor less' than 
feudalism or slavery, a mode of appropriation of the 
unpaid labour of others which is only distinguished by a 
'different way' of extorting it), and (2) the analysis of 
the role of the class struggle in history (arising on the 
basis of very ancient material conditions ­ the 'scarcity 
of products, the 'non-development' of the productive 
forces ­ and destined to be abolished on the basis of 
new conditions: the 'impetuous development' of the productive 
forces, 'abundance'). Once these two problems 
are separated, it is no longer possible to pose in scientific 
terms the question of why no new form of relations of 
exploitation is possible beyond capitalist relations of 
production: the social revolution that destroys capitalist 
relations of production appears merely as a particular 
production over the productive forces in their 'combination'; 
(2) I introduced an indefinitely renewable 
aporia concerning the formation of this new mode of 
production, or, if you like, the 'transition' to this 
'transitional mode of production': (3) I introduced the 
possibility of a 'general theory of transition' or 
'transitions', itself conceived as an aspect of a 'general 
theory of the combination ­ or articulation ­ of modes 
of production'. Such a theory is in fact the substitute for 
a real elaboration of the dialectic of the history of social 
formations in the sense in which Althusser had outlined 
it in his text 'Contradiction and Overdetermination'. But 
this could not be clearly perceived so long as we did not 
think distinctly the two concepts of 'social formation' and 
'mode of production', and the nature of their relationship 
(which is still far from having been perfectly worked 
out). That is why it is particularly interesting that some 
of us initially attempted, following the line of Reading 
Capital, definitions of the social formation as a 
'combination of several modes of production', or as 
an articulation of modes of production', i.e., as a 
'complex' mode of production, or else as a 'higher 
ranking' mode of production in a kind of scale of types.[8] 
I cannot here go into great length on the difficulties of 
such a theory, of which the idea of a 'general theory of 
transition' is a particular application. Let me say that 
such a theory is substantially equivalent, though more 
complicated in detail, to the formulations of the Preface 
to A Contribution to the Critique of Political Economy 
(1859) (which are themselves a reprise of themes from 
The German Ideology), which have been a very heavy 
68

as the bourgeoisie 'represented' that movement within 
feudal society, or by explaining that at a certain period 
the bourgeoisie becomes a 'superfluous' class because it 
is a class 'outside production', in the same way as the 
feudal aristocracy had become a 'superfluous' class of 
rentiers, etc.[10] It is essential to pose the general thesis 
that historical materialism is not only a theory of the 
necessity of the (revolutionary) transformation of the 
social relations, but also a theory of the transformation 
of the mode of transformation of social relations. Such 
that two revolutions' never have the same concept. 
In order to Analyse the Real History of the Capitalist 
Mode of Production, all Evolutionism must be Removed 
from the Concept of 'Tendency' 
Finally it turns out that this thesis is closely linked to a 
rectification of what I have thought elsewhere as the 
development or evolution of a mode of production (in 
particular capitalism). In order to explain this, it is 
enough to say the following: that it is impossible to 
account for the specificity of each revolutionary 
'transition' unless it is related not only to the specific 
general form of the preceding social relations (e.g., the 
form capital/wage-labour, and the type of combination 
of the relations of production and the productive forces 
it implies), but also to the specific history of the preceding 
mode of production i.e., to the history of the 
social formations that depend on the development of that 
mode of production. 
case of the general mechanism of contradiction/readjustment 
between the relations of production and the productive 
forces, and its specific result, the abolition of 
all the forms of class rule, remains inexplicable. It is then 
possible to leave the field, together or separately, to 
relativism in the definition of the relations of production 
and to evolutionism in the analysis of the development of 
the productive forces.[9] 
To go right to the point, let me say that one of the basic 
theses of historical materialism seems to me to be the 
following: there is a general problematic (to borrow an 
expression of Duroux's) of 'transition' in social 
formations, i.e., of 'revolution in the relations of production'. 
This pertains to the fact that the concept of 
'class societies', resting on modes of production which 
are at the same time modes of exploitation, cannot be 
constituted without reference to the historical transformation 
of modes of exploitation (in other words, there 
is no such thing as exploitation in general, only determinate 
forms of exploitation). But for all that there is no 
such thing as a general theory of transition, in the strong 
sense of an explanation of the causality of a process. On 
the contrary, it emerges that each historical 'transition' 
is different, materially and therefore conceptually. This 
point is of fundamental political importance if it is true 
that Marxist theoreticians, starting with Engels himself, 
have occasionally tended to consider the 'transition' from 
feudalism to capitalism and the 'transition' from 
capitalism to socialism as analogous processes, e.g., by 
representing the modern proletariat as the 'representative' 
of the movement of the productive forces in the same way 
69

recognized its nature. In particular it presupposes going 
back, even at the most abstract level, to the question of 
reproduction and of the 'tendencies' of the capitalist 
mode of production. From this point of view, I should 
no doubt invert my normal formulation: it should not be 
said that there is in the mode of production a tendency to 
reproduction of the relations of production, or rather a 
tendency (to accumulation, the concentration of capital, 
a rising organic composition, etc.) which realises the 
reproduction of the relations of production. On the contrary, 
it should be asked how a 'single' tendency can 
turn out to be reproduced as a tendency, in a repetitive 
fashion, such that its efforts are cumulative according 
to an apparent continuity. It should be asked in what form 
a tendency can be realized (produce historical effects), 
taking into account the conditions of its own reproduction. 
It should be asked how this reproduction is possible even 
when in the social formation, the sole real 'site' of the 
process of reproduction, its material conditions have 
been historically transformed. 
In other words, it is essential to break once and for all 
with the ideological illusion I have been discussing 
according to which the existence of a historical 'tendency' 
seems simultaneously to be the tendency of this 'tendency' 
to persist, hence to be realised, etc. And for that it is 
essential to understand that it is not the mode of production 
(and its development) that 'reproduces' the social 
formation and in some sense 'engenders' its history 
but quite the contrary, the history of the social 
formation that reproduces the mode of production on 
which it rests and explains its development and its trans- 
In fact, the first person to realise all the theoretical 
importance of this fact, as a result of the constraint of 
circumstances, was not Marx, but Lenin, and this discovery 
implies ultimately a rectification of certain of 
Marx's formulations. It was Lenin, insofar as he 
demonstrated that the process of 'socialist' revolutionary 
'transition' was not linked to the existence of capitalist 
relations of production in general, but to the existence 
of a determinate stage in the history of capitalism: 
imperialism, i.e., of determinate 'transformed forms' 
of capitalist relations of production (and not only, of 
course, of productive forces inside the 'framework of 
unchanged relations of production). Such that the problem 
of an analysis of the socialist (proletarian) revolution 
and of what 'socialism' itself is as a historical epoch 
becomes inseparable from the problem of the analysis 
of imperialism, and hence of the determinate historical 
phases (or periods) of the history of capitalism. It has 
to be stated that one of the orientations of my text in 
Reading Capital led precisely to making these phases, 
i.e., these historical transformations, strictly unthinkable, 
except in the economistic-evolutionistic sense of 
developmental phases, linear stages in the realisation 
of a tendency unchanged in itself. 
Still very schematically, it is clear therefore that the 
examination of the problem of the socialist transition 
presupposes among other things an overall critical 
review of the problem of the history of capitalism, and 
a recasting of our 'reading' of Capital as a function of 
this problem, a review and a recasting made all the 
more difficult by the fact that Marx himself partly mis- 
70

4 Given that it is absolutely excluded that it could be a 
matter of an 'external' effect, since there is no exterior 
to the historical process. As Mao Tse-tung explains in 
On Contradiction: 'The fundamental cause of the development 
of a thing is not external but internal' (Selected 
Works, Vol. I, p. 313). But it is precisely the structural 
modality of this internal contradiction that has to be made 
explicit. From the Communist Manifesto on Marx took 
as his object its principal aspect: the 'internal/external' 
position of the proletariat as a class in the structure of 
capitalist relations of production. 
5 I say formulae, for Spinoza's materialist dialectic is 
quite the opposite of this bad application of it. 
6 Note that this 'autonomy' could also he stated in our 
jargon as 'the economic is both determinant and 
dominant', or at least this is one of the possible interpretations 
of that obscure formula. 
7 Cf for the time being on this point Althusser's article 
'Ideology and Ideological State Apparatuses', op. cit. 
8 This was the case for myself, for Althusser, for 
Terray (Le marxisme devant les sociétés "primitives", 
Maspero, Paris 1969) and above all for Pierre-Philippe 
Rey whose text Sur l'articulation des modes de production 
has been published in Problèmes de Planification, dir. 
Charles Bettelheim, Ecole Pratique des Hautes Etudes, 
VIe Section, no. 13-14. Note that in the same issue there 
is an already old and unfortunately very elliptical 'note' 
by Duroux, who had seen these difficulties very clearly. 
9 This ideological complementarity of relativism and 
evolutionism, which is only apparently surprising, has 
been clearly pointed out recently by Claudia Mancina: 
formations. The history of the social formation, i.e., 
the history of the different class struggles of which it 
is composed, and of their 'resultant' in successive 
historical conjunctures, to use a metaphor frequently 
employed by Lenin. That is, the history of class struggles 
and of their results in successive materially determined 
historical conjunctures. In this I shall perhaps be able 
to make an effective contribution to Marxism-Leninism: 
not to Marxism followed by Leninism, but, if I may say 
so, to Marxism in Leninism. 
January - October 1972 
NOTES 
1 Cf. on this point Louis Althusser: 'Ideology and 
Ideological State Apparatuses,' June 1970, in Lenin and 
Philosophy and Other Essays, NLB, London 1971, 
pp. 121-73. 
2 Let me make it clear that here I am aiming at this 
particular point in Badiou's work and absolutely not at 
all its aspects. None of us, particularly at this period, 
succeeded in being perfectly 'consistent'. 
3 I have attempted to make this materialist thesis of the 
primacy of the Relations of Production inside the combination 
of relations of production and productive forces 
more clearly explicit in my forthcoming article: 
'Capitalisame et theorie des formations sociales' in 
L'Economie, dir. A. Vanoli, coll. 'Les Sciences de 
l'Action', CEPL Paris. 
71

'Strutture e contradizzione in Godelier', Critica 
Marxista, 1971 no.4. 
10 Marx pronounced such formulae in determinate conjunctures 
as a reminder of the fact that the proletariat 
had in its turn to make 'its' revolution, as the bourgeoisie 
had made its. But Engels almost theorized this 
analogy, or rather made this analogy the very basis for 
the exposition of historical materialism, notably in 
Socialism utopian and scientific and more clearly still, 
if that is possible, in his article 'Social Classes ­ 
Necessary and Superfluous', (The Labour Standard 6th 
August 1861, in Karl Marx and Friedrich Engels 
Articles on Britain, Progress Publishers, Moscow 1971 
pp. 384, ff. ), which provided Kautsky with much of his 
inspiration. Lenin, though, never said it. 
I have recently attempted to analyse Marx's formulations 
on the analogy between the bourgeois revolution and the 
proletarian revolution where the State is concerned 
from the Communist Manifesto to The Civil War in France 
('La rectification du Manifeste Communiste', La Pensée, 
August 1972). 
72