This article was downloaded by: [University of Oxford]

On: 17 February 2010

Access details: Access Details: [subscription number 905765479]
Publisher Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK

Economy and Society

Publication details, including instructions for authors and subscription information:

http://www.informaworld.com/smpp/title~content=t713685159

Louis Althusser and G. A. Cohen: a confrontation

Grahame Lock

To cite this Article Lock, Grahame(1988) 'Louis Althusser and G. A. Cohen: a confrontation', Economy and Society, 17: 4,

499 517

To link to this Article: DOI: 10.1080/03085148800000021

URL: http://dx.doi.org/10.1080/03085148800000021

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.

Louis Althusser and G. A.

Cohen: a confrontation
Grahame Lock

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Abstract
The paper compares and confronts the work of two of the most distinguished
living Marxist philosophers: G. A. Cohen from the English-speaking world, and
Louis Althusser from France. It develops a critique of certain of Cohen's theses
from the standpoint of ideas present in the work of Althusser. But it also
problematizes certain presuppositions common to the work of both - in particular,
the notion that historical development (transition or revolution) should be
explained in terms of some general theory of non-correspondence between
productive forces and production relations: the difference being simply that,
within this scheme, Althusser accords explanatory primacy to the latter, Cohen to
the former. Cohen's and Althusser's accounts of technological innovation and
development are also compared, in connexion with the contrasting place which
they attribute to the notice of human rationality on the one side and class struggle
on the other.
During the decades following the Second World War the organized Marxist
labour movement slowly but surely lost ground in most countries of the
western world. Yet Marxism, nevertheless, retained some hold on the
intellectuals of those lands, a hold usually proportionately greater than the
attraction which it exercized on the wider population, or even in many cases on
the working class. In the 1970s, however, Marxist theory entered what
Althusser called a 'crisis'' - a crisis obviously linked to, but not directly
reducible to, the more or less simultaneous crisis not only of the abovementioned Marxist labour movement, but of the whole of the workers'
movement, extending to the various national labour or socialist parties and to
the trade unions. It is therefore a somewhat paradoxical fact that two of the
most impressive achievements of post-war Marxist philosophy - indeed,
perhaps of the whole history of Marxist philosophy - were produced more or
less contemporaneously with the ripening of this crisis. I refer to the work of
Louis Althusser in France, and of G. A. Cohen in England.
In the present paperZI attempt to construct a confrontation between certain
Economy and Society Volume 17 Number 4 N a m b e r 1988
0Routledge 1988 0308-5147/88/1704/0499 $3.00

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

500

Graharne Lock

central aspects of their respective doctrines. This attempt, by its very nature it brings together and compares work ofvery different kinds, written in starkly
contrasting styles and for quite different audiences -, leaves much in the work
of both thinkers undiscussed. But it may well help to focus attention on just
what kinds of matters are -both explicitly and tacitly - at stake in their work.
The principal theses presented by Cohen, to be found in his book Karl
Mum's Theory ofHistory: a Defence (Cohen 1978), are by now well-known. I
shall nevertheless rehearse some of them, in order to be able clearly to draw
the contrast with the comparable Althusserian claims. I shall ignore, for the
purposes of this article, later qualifications and retractions by Cohen of
arguments presented in his book.3
Cohen characterizes his purpose in the following manner: to defend
historical materialism by offering arguments in its favour (some of which have
been overlooked, even by its warmest protagonists), thus blunting or defusing
a number of its adversaries' instruments of attack. This aim is to be achieved in
particular by 'presenting the theory . . . in an attractive form' (p. ix). Unlike
most latter-day Marxists, Cohen defends in the book a traditional or
'old-fashioned' orthodoxy: that is, the writings of Marx himself, and more
especially, as far as exegesis is concerned, the well-known 1859 Vorwort to the
Zur Kritik hpolitischen Oekonomie. His goal is then, further, to 'straighten
Marx out', to tidy up his thoughts, and thus to provide a 'less ambiguous'
version of his ideas. Yet it is precisely these last ambitions, as I shall suggest,
which turn out to be most directly problematic.
The Marxism which he offers - in clear contrast, as will become apparent,
to that of Althusser - is a technological explanation of history. It is a conception
according to which 'history is, fundamentally, the growth of human productive
power', where 'forms of society rise and fall according as they enable or
impede that growth' (p. 2). For Cohen, explanatory primacy is attributed to
the productive forces. Their tendency to develop through history is explained
in terms of the historical situation of scarcity in which men live, together with
their possession of a rational and intelligent nature. It is in turn the
development of the productive forces in history which explains the emergence
and disappearance of economic structures (otherwise known as 'production
relations'); and it is, finally, the latter which explain the nature of social
superstructures. Conversely, social superstructures (that is, non-economic
institutions) are said to 'consolidate' economic structures, while these
structures stimulate - where they do not hinder - the reproduction andlor
development of the productive forces.
I shall say more shortly about these themes. But first I want to make some
comments on the political and theoretical situation, not so much in which the
book was written, as into which it fell. It was a situation which, as I already
noted, has been characterized as one of a m'sis of Marxism; one provoked
perhaps by political events, at least to some extent, and reflected within
political organizations by doctrinal hesitation or backsliding - as for example
in the 'emergency' abandonment by western Communist parties of central

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Lo~risAlthusserand G. A. Cohen: a cotfrontation

501

categories of traditional Marxist doctrine, like that of the dictatorship of the

proletariat (cf. Balibar 1977). But it was also a properly theoretical crisis, a
crisis within the theory of Marxism: a moment when the whole structure
looked (and still looks) as if it might fall apart. This state of crisis is of course
not such a matter of public debate in the Anglo-American world, for obvious
reasons connected with the weakness of the Marxist labour movement there.
Be that as it may: Cohen's book shows no signs of having been written in this
context of crisis. Rather, it was written as, and has been received as, a more or
less purely philosophical treatise, in a quite traditional sense of this phrase: as
a work to be read and assessed in abstraction from any consideration of its
possible roots in or impact on the political and ideological situation. Its
reception was moreover overwhelmingly favourable.
On the one hand Cohen rendered Marxism if not respectable, then at least
a to-be-respected opponent in English-language philosophical and socialscientific circles. On the other hand he provided less theoretically-minded (or
less theoretically handy) Marxists with an unimpeachably serious philosophical backstop. On the way he disposed of many of the philosophical objections
to Marxism made by Anglo-Saxon experts such as H. B. Acton and John
Plamenatz (though not Karl Popper, whose rebuttals were, for whatever
reason, left aside), as well as providing trenchant criticism of various positions
defended in the philosophy of science, sociology and political science, by Carl
Hempel, Larry Wright, Robert Merton and others.
Yet at the same time Cohen makes hardly any reference to nonAnglophone Marxist writing - except to that of Marx and Engels. In his
Foreword he does however confess himself 'bound to say a word' about his
attitude to the work of Louis Althusser. It was, he writes, 'Althusser's Pour
Mum [which]persuaded me that the abidingly important Marx is to be found
in Capital and the writings preparatory to it'. But he was disappointed by
Althusser's essays in Lire le Capital, which he qualified as vague and evasive.
The question may therefore be posed as to whether - on Cohenian criteria Althusser's work is vague and evasive because it is not yet suficiently 'worked
out' (clarified) or because, like certain other 'continental Marxisms', it cannot
be thus 'worked out' and would not even aim to be. Cohen does not imply the
latter, and may not believe it. I suspect moreover that his own book, in certain
of its most fundamental aspects, is some kind of answer to Althusser. For he
knows his rival's work, being to my personal knowledge one of the few
English-speaking intellectuals to have read PourMamand Lire le Capital in the
original French, before the translations were available - together, for
purposes of textual control, with the complete French edition of Marx's
Capital. It is for these reasons too that it seems not absurd to attempt a
confrontation between the most celebrated of recent Anglophone and
Francophone Marxist philosophers.
Cohen's manner of writing is marked by a profound honesty, a quality all
too often absent from Marxist work. This of course I admire. But my
admiration for the book derives further in part from an interpretation of its

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

502

Grahame Lock

character and effects which differs from that given by Cohen himself when he
talks about his purpose. He states, as we saw, that he is aiming among other
things at the formulation of a less untidy presentation of 'the' theory originally
elaborated by Marx. But his goal is also to identify exactly what Mam said, as
opposed to what many interpreters have claimed that he said. And the -very
successful -result of this second enterprise as a matter of fact tends to hinder
the satisfaction of the first aim. For what Cohen reveals in his many careful
exegeses and treatments - whether or not this was his intention - are the
enormous ambivalences, gaps and contradictions in Marx's work; enormous
enough to make it unlikely that a reasonably comprehensive defence of Mam's
writings, taken as a whole, can in fact be provided. Of course, Cohen's
intention is often to reveal these internal difficulties in Marx. But his aim is
different: it is to get rid of them as quickly as possible, in order to produce a
'least ambiguous' defence of Marxism. My suggestion is that the problem lies
deeper.
An example would be Mam's contention, as interpreted by Cohen, that
class struggle is a 'theoretically derivative battle'. The sense in which he uses
the term 'derivative' is admittedly only to be understood in the context of his
principal theses it is not for example that he believes class struggle to be a
secondary process in the political sense.
Yet I believe that a different reading of Mam is also possible here. One
would need in this connexion to distinguish between (i) an explanatory theory
which aims to account for the general development of history, and thus for its
'line': e.g. the progression from tribal society to ancient society to feudalism to
capitalism to socialism and finally to communism; and (ii) what one might call
a 'special' theory, which would allow one to attempt to account for particular
transitions as tokens of one given type of transition: say, the transition from
feudalism to capitalism, or from capitalism to communism. Now Mam, it
seems to me, offered both general and special theories, which in this case were
mutually incompatible. A further complication is moreover that he also now
and again suggested that the set of special theories at the same time
constituted another general theory, namely a general theory of class struggle,
as stated in the opening words of the CommunistManifesto, ch. 1: 'The history
of all hitherto existing society is the history of class struggle.'
With regard to (i), Cohen is right to assert that Mam did try to elaborate a
general account of history which relies on the attribution of explanatory
primacy to the development of the productive forces. T o the extent that
Althusser tended to deny that this thesis was to be found in Marx, he was
wrong. With regard to (ii), however, I think that Mam was inclined to attribute
explanatory primacy to class struggle, especially in regard to the analysis of
capitalist society and of its tendencies. And in insisting on this fact, Althusser
was right.
Now this same fact (if it is one) would still not embarrass Cohen, who
accepts that 'for Mam the immediate explanation of major social transformations is often found in the battle between classes' (p. 48). What would

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Louis Althtrsser and G. A. Cohen: a corflrot~ratiott 503

however embarrass him is the combination of this proposition with a
demonstration of an incompatibility between the two kinds of theories, general
and special. On Althusser's reading, indeed, any general theogr of history is
speculative, and therefore a kind of metaphysics of history. He believes that
Marx demonstrated the need to avoid any such account, and in particular to
avoid any kind of evolutionism. An evolutionist theory is a teleological
account; and a teleology requires that the principle of historical development
be present at the point or moment of origin of the process. Cohen's theory
possesses such a principle in the three factors mentioned above: economic
scarcity, human rationality and human intelligence.
Let us now turn to the central concepts of what I called the special theory of
capitalism: those of class and class struggle. Andrew Levine and Erik Olin
Wright (1980) are among those who have challenged Cohen from a standpoint
which lays a good deal of emphasis on a version of the category of class, and
which throws doubt on the relevance of his reference to the three factors just
named. But, as we shall see, the basis of their critique is less far-reaching than
that of Althusser.
They suggest that Cohen does what Marx condemned in the classical
political economists (though himself also attempted), namely 'building an
economic theory out of a logically prior notion of individual (instrumental)
rationality in a milieu of (relative) scarcity' (Levine and Wright p. 57). They
believe that his position is thus faithful to (an aspect of) Marx, but wrong. For
it ignores the fact that 'the realization of human interests (in the development
of productive forces) can be blocked by social constraints'. In other words, it
ignores what they call the problem of class capacities. These they define as the
organizational, ideological and material resources available to classes in class
struggle. Going further than Marx himself, many later Marxists have, they
note, questioned the assumption that the development of class capacities
automatically follows on from the emergence of 'revolutionary interests'; and
they have tried to investigate the 'systematic processes' at work in capitalist
society tending to disorganize the working class. The authors conclude that a
rational interest in the transformation of an economic system is, on this view,
not a suficient condition of its transformation. But since the development of
the productive forces does not (necessarily) lead to any corresponding
development of class capacities, it is 'arbitrary' of Cohen to attribute
explanatory primacy to the productive forces.
They consequently propose to 'supplement' the account of historical
materialism found in Marx's 1859 V o m o ~on
, which Cohen leans heavily,
with such an inquiry into class capacities. That they talk only about
supplementing this text - though they do indicate that this is a minimum
requirement (Levine and Wright p. 68) - may be an index of their own
unwillingness to abandon certain categories of 'orthodox Marxism' which
Althusserian Marxists have rejected: for instance that of the rational (or
objective) interest of a class and that of rational action, predicated of a class,
conceived of as some kind of 'subject of history'.'

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

504

Grahatne Lock

Levine and Wright appear in fact to accept Cohen's aim as proper: to derive
an 'adequate, substantive picture of the general contours of human history';
they disagree only on the question of whether this aim is realized in his book. I
doubt however, for reasons sketched out below, whether the aim itself is in
order, or whether its acceptance does not lead to intolerably many further
dificulties. I shall give a first example: their retention of the above-mentioned
categories of the 'rational interest' and 'rational action' of a class is, as we saw,
paralleled by their instrumental conception of class capacities; for these latter
refer, as we also saw, to the 'resources' available to a given class. But Althusser
and Balibar have suggested that the proletariat only exists as a class to the
extent that it is organisationallyand ideologically united, which unity is assured
only in proportion to the division of the bourgeoisie; and vice versa.'
Consequently, one can on the Althusserian view hardly speak of or measure
the extent of the 'resources available to a class in struggle', as if the class
already unambiguously existed but was unfortunately deprived of the
organizational, material and ideological 'instruments' necessary to realize its
rational interest. How does this argument bear on Cohen's view?
Cohen argues for a purely structural definition of class (thus, implicitly,
against the view of Althusser and Balibar) at pp. 73-7 of his book. His
argument is framed in terms of a critique of the historian E. P. Thompson,6
whose objections to the structural definition are shown to be unsound. This
definition asserts that class may be defined purely in terms of production
relations, and indeed must be so defined, since the exclusion of factors and
'consciousness, culture and politics' is required to 'protect the substantive
character' of the Marxian thesis that class position strongly conditions
consciousness, culture and politics.
In my view this is not quite right. One may, I think, usefully distinguish
between two relations of 'conditioning': the conditioning of an individuafs
consciousness etc., and the conditioning of the ideology of a class. What Cohen
is principally talking about, it seems to me, is the former: the conditioning of
an individual's consciousness (and, one might add, of his unconscious) by
so-called 'external' factors. This process is the one studied by political
socialization theory in its own way. But I believe that the core of the Marxist
claim concerns rather the causation and transformation of ideologies and of
the contradictions inside and between them, in their relation to class struggle.
I see of course that it is unilluminating to talk about the class conditioning of
the ideology of a class if the relevant class is not identified independently of the
ideology in question. If such a notion is to make sense at all it would therefore
be necessary to differentiate two senses of 'class'. Cohen himself notes that
Marx was forced to distinguish between a 'class-in-itself (corresponding to
Cohen's own use) and a 'class-for-itself, even quoting Marx to the effect that
'the proletariat can act as a class only by constituting itself a distinct political
party' (p. 7). And I am prepared to accept that there is a sense in which a
divided working class, in Cohen's sense, may and indeed must be called a
class. In this case one might propose a terminological stipulation, in order to

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Louis Alfhusser and G. A. Cohen: a confion~afio~ 505

avoid confusion: for instance by calling such a class, in this sense of the term,
'the working class', and then calling what I am talking about, and what
Althusser talks about - namely the class as constituted by its (tendency to)
political and ideological unity - the 'proletariat'. (Similarly one could
distinguish the 'capitalist class' and the 'bourgeoisie'.)
The above remarks do not imply that I think that a 'pure' structural
definition of class is possible. They mean only that I believe that a distinction
such as that just sketched out would permit a better formulation of acounts of
the dynamic of class struggle under capitalism, of the splintering and
unification of classes and so on. It would permit a better formulation of
accounts of the uneven development of the economic, political and ideological
class struggles in their relation one to another. Besides, one might invoke here
the argument of Balibar to the effect that the term 'class' does not have the
same meaning when applied to the working class on the one hand and the
capitalist class on the other (cf. Balibar 1974: 188-9). The precise sense of
this argument has however yet to be sufficiently elaborated.
The argument for the primacy of the productive forces for which Cohen
argues is of a quite novel kind (to my knowledge at least) within Marxist
theory. Malxists have long been dissatisifed with Mam's own account of the
causal primacy of the economy (however this term is interpreted), or of the
relation between 'base' and 'superstructure'. Engels's famous letter of 21/22
Septembrr. 1890 to Bloch is one early expression ofthis unease.
In 1965 Althusser and Balibar, in Pour Mum and Lire le Capital, produced
an ingenious extension and development of Engels's unexplained notion of
'determination (by the economy) in the last instance'. Determination was
distinguished from domination, and so 'the economy' appeared twice in the
characterizing formula of the capitalist mode of production: it was now the
structure of the capitalist mode of production (i.e. its production relation +
productive forces, in which conjunction the 'primacy' of the former over the
latter was asserted) which was said to determine, in the last instance, the
domination of the economic instance in any capitalist social formation. This
may sound a little unclear. The point was that, although the economy, in one
sense, is determinant in every type of society, what it determines is which
'instance' -the economic, the ideological or the political - is dominant in any
given type of society. So the domination of the economic instance in capitalism
is a feature which it does not share with other kinds of society. For instance,
the determination by the economy results in classical antiquity in the
domination of the political instance. In the Middle Ages it results in the
domination of the ideological (religious) instance. (Balibar, in Althusser and
Balibar 1970: 217.)
Cohen attacked the problem in a quite different way. He produced in fact a
second -only the second in many years -properly new and interesting schema
intended to deal with this problem of economic determination. For him the
'real basis of society' is indeed constituted by the set of production relations;
and these production relations make up what is also called the 'economic

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

506

Graharne Lock

structure' of society. But this means that theproductivefirces are not part of the
economic structure ofsociety; nevertheless they enjoy explanatory primacy over the
production relations (pp. 28-9). This assertion thus involves the rejection of
the widely-held notion that if the productive forces are explanatorily primary,
then they must form part of the economic basis or foundation of society; for
Cohen denies the consequent of this proposition. The productive forces are
said to be the 'basis' of society only in the sense in which a basis may be
something external to that of which it is the basis (like the pedestal on which a
statue stands, but which is not a part of the statue). Or in other words, the
productive forces 'are indeed the foundation of the economy but they do not
belong to the economic foundation' - they occur 'below' the latter (p. 30).
They 'strongly determine the character of the economic structure, while
forming no part of it' (p. 31).
Their determining and explanatory role in this respect is then elucidated by
reference to functional-explanatory forms, to which I shall return later. The
advantage of Cohen's formula (filled in with his further, elegant and extensive
account ofwhat the determination and explanation in question come to) is that
it at one and at the same time respects Mam's own words in the Vomort and yet
is methodologically exceptionally rigorous.
Now there is, in spite of the differences between them, a similarity between
the two above-mentioned schemas, those of Althusser/Balibar and of Cohen;
for both attempt to explain historical development (transitions or revolutions)
in terms of a general theory of non-correspondence between productive forces and
production relations. The difference is that in Althusser/Balibar the latter are
accorded explanatory primacy, and in Cohen the former. But both parties
recognize that such a general theory, if it were to be recognizably Marxist,
would require such a notion of non-correspondence. Just here, however, is
where in my opinion the problem lies. My doubts concern, as I have indicated,
the project itself. Why was it supposed that such a general theory is possible at
all?
Interestingly, the Althusserian version - rich in contradictions - already
contained within itself the seeds of its own dissolution, and of an alternative.
The reason is roughly the following. (The argument is a little technical, but
seems to me to make sense.) Correspondence and non-correspondence were
treated by Althusser/Balibar in terms of the 'subjection' or 'subsumption' of
labour to c a ~ i t a lProductive
.~
forces, for these authors, were constituted by the
relation of real appropriation of nature (producing things in a particular way).
Production relations on the other hand were defined as the relations of
expropriation of the producer (e.g. those characteristic of capitalist exploitation). Now if the correspondence between the production relations and the
level (or type) of the productive forces is, as Balibar argued, to be understood
in terms of a correspondence between the real and the formal subjection of
labour, then a non-correspondence may be understood as involving a failure in
the reproduction of one at least of these relations of subjection (Balibar, in
Althusser and Balibar 1970: 303-4). But one type of failure is a much better

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Louis Althusser and C. A. Coheti: N confrot~ra~ion 507

explanatory candidate than the other. For it is unlikely (to take the case of a
putative transition from capitalism to socialism) that the workers, with their
allies, could overnight modifjr the relation of 'real' subjection - i.e., could
overnight modifjr the structure of the productive forces, the type and level of
technologies in use etc. So it would seem that transition (revolution) must
rather be, at least in the first instance, the result of a refusal or rejection of
'formal' subjection: that is (on a free reading) of a rebellion of workers - not
yet necessarily motivated by any clear idea of either the benefits or the costs of
socialism - against their so-called 'absolute non-ownership' of the means of
production, where 'ownership' is to be understood in the sense of appropriation rather than in the superstructural, merely juridical sense.
Thus, to put it bluntly, the motor of transition would lie in class struggle.
Yet questions of the superiority of socialism to capitalism are hardly at issue
here (whatever sense is given to the idea of superiority). There is very little
room in this account for any 'rational calculation' by the working class of its
'rational interest'. Nor is such an account obviously applicable to the analysis
of social forms other than capitalism. We have nothing more than a 'special'
theory. But these conclusions were not drawn by the authors ofLire le Capital
until later."
Now the above points touch on comments made by Cohen, but which he
would not consider central to the main themes of his book. My opinion is
nevertheless that similar critical considerations apply to any attempt to back up
a general theory of historical development with a general account of relations
of correspondence and non-correspondence between productive forces and
production relations. Cohen's position depends for example on an implicit
rejection of Balibar's arguments to the effect that all general theories of the
kind which Cohen's book seems to propose rest on the notion of some sort of
'essential definition' of economic - or political or ideological - phenomena
which pre-exists the process of their 'historical definition'; whereas the sense
of the term 'economy' and related terms changes with the transition to
capitalism, and must change again with any transition from capitalism to
socialism or communism (Balibar 1974). Similarly, the transition from
feudalism to capitalism would be of another kind (in more than the obvious
sense) than the transition from capitalism to communism; and the two would
in consequence not be explained as two instances of a single kind of
contradiction, namely the non-correspondence between productive forces
and production relations. But this would mean that a doubt hangs over the
legitimacy of Cohen's principal question: how to explain the 'development of
histoy'? (to which he provides the answer: in terms of the 'growth of human
productive power').
Let us look briefly at the instrument employed by Cohen to give content to
this answer. The instrument is functional explanation. Functional explanation
is, in his view, a special type of causal explanation, one which explains a
phenomenon or event (etc.) in terms of its function - roughly, in terms of its
effects. We should immediately mention his insistence that functional

508

Grakatrte Lock

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

explanations are not (necessarily)functionalist, for functionalism propounds

the dubious 'functional interconnexion thesis', to the effect that 'all elements
of social life support or reinforce one another, and the whole society'. Marxists
have, according to Cohen, often mistakenly rejected functional explanation on
account of their aversion to functionalism. As he correctly points out,
however, anti-functionalists like Althusser have made use of functional
explanations. This Althusser does for instance in his well-known article on
'Ideology and Ideological State Apparatuses' (Althusser 1971). Let us just
quote a few words from Althusser in this connexion. Talking about what he
calls the 'educational State apparatus', he describes the manner in which it
simultaneously fills children with certain kinds of 'know-how' and prepares
them for the discipline of the world of work, ejecting various masses of
children onto the labour market at the appropriate age. Thus:
Each mass ejected en route is provided with the ideology which suits the
role it has to fulfil in class society: the role of the exploited. . . ;the role of
the agent of exploitation. . . ,of the agent of repression. . ., or of the
professional ideologist?
Each child is thus provided, says Althusser, with the ideology which 'suits the
role' it has to play. But the teachers, although they are responsible for
providing the child with this - functional - ideology, are in general quite
unconscious of the task which they, in their turn, are performing: 'So little do
they suspect it that their own devotion contributes to the maintenance and
nourishment of [an] ideology of the School' according to which the School is
itself a 'natural environment purged of ideology', and so on. But nor is there
any other 'visible hand' planning the whole scheme. On the contrary: to
imagine such a thing would be fall back into the eighteenth-century idea that
Priests or Despots - in this case Ministers of Education or whatever - are
responsible for forging the appropriate 'ideological lies'. But the ruling class is
itself 'in ideology', not outside of it and controlling it.
Thus Althusser gestures towards a functional explanation of the operation
of ideology. But he says little or nothing about the mechanisms according to
which this operation takes place. Is this an (insuperable) objection? The
general point, as to whether any functional explanation must at least indicate
where such a mechanism is to be found, was in fact raised as a point of
criticism against Cohen by Jon Elster (1980). When does a reference to
function provide a sufficient basis for a functional explanation? According to
Elster only (if at all, since in that case it is dubious whether we would still want
to talk about a functional explanation) when the particular mechanism that
may justify it in a given case is indeed provided. But, says Elster, 'Cohen does
not even attempt to provide such a mechanism' in the case of historical
materialism (Elster, op. cit., p. 127). Nor of course does Althusser.
What Cohen does however do is provide a justification for the proferring of
functional explanations without specification of mechanism. I shall not
attempt here to treat the question at any depth. It happens that I agree with

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Cohen that Elster's argument is wanting, since an 'alternative route' from a

consequence-statement to a functional explanation may be available in the
form of the provision of supporting evidence which would establish a
confirming general pattern, even when little or no hint of the relevant
mechanism (or elaboration) is to be had. In Cohen's words, 'we may be
confident that a caused b in a given context because of appropriately parallel
cases in other contexts, even if we do not know how a caused b (Cohen
1980: 132). The question however remains as to whether detailed and
convincing Marxist explanations of this kind can in practice be given.
Now Cohen suggests that historical materialism may in this respect be 'in its
Lamarckian (pre-Darwinian) stage'. By this he means that Lamarck, who
provided a functional-explanatory account of biological evolution, was scientifically - justified in his belief in that account, even though his
elaboration of the relevant mechanism was - that is, turned out to be - false.
(Darwin of course provided the true elaboration in the form of the theory of
random variation and natural selection.) T o make this claim, that Marxism
may be in its Lamarckian stage, is thus to suggest that Marxists know certain
truths about society, even if they cannot yet provide accounts of the relevant
explanatory mechanisms.
T o take one example: for Cohen, Marxists are often unjustly accused of
advancing a demonstrably false conspiracy theory of history when they argue
for instance that ' "it is no accident thatn left-wing commentators receive little
space in major. . . newspapers' and so on - unjustly because such an account
may form part of a general functional-explanatory story in which no such
conspiratorial mechanism is supposed. There may of course actually be such a
conspiracy: Cohen thinks there often is. But to demonstrate the existence of a
conspiracy would be to provide the elaboration for what is so far only a
functional-explanatory claim. Cohen's claim is that the latter might be valid
without more ado. Further, it may be true that 'ruling classes are well placed to
propagate ideologies congenial to themselves'. Yet an ideology, before it can
be propagated, must be constituted.
And on that point there are traces in Marx of a Darwinian mechanism, a
notion that thought-systems are produced in comparative independence
from social constraint, but persist and gain social life following a filtration
process which selects those well adapted for ideological service. . . . There
is a kind of 'ideological pool' which yields elements in different
configurations as social requirements change. (Cohen 1978: 289-91)
Now if Marxism really is in a 'Lamarckian' stage, what kinds of elaborations
of mechanisms might it seek in order to round out its functional explanations?
Cohen notes four, among the various possibilities: (1) purposive mechanisms
(e.g. a conscious decision by the government to cow workers into submission
in regard to the acceptance of real wage-cuts by the deliberate creation of
unemployment); (2) Darwinian mechanisms (e.g. competition between large
and small companies in an industrial field where only the large can and do

5 10

Graharnr Lock

survive); (3) Lamarckian mechanisms: evolution in virtue of new characteristics acquired - but not necessarily purposively within the life history of an
'organism' (e.g. where the subjection ofworkers to the discipline ofthe factory
lays the foundations for the creation of trade-union discipline, which becomes
a tradition transmitted from one generation to another); (4) mechanisms of
self-deception, which operate 'through the minds' of agents but without their
'full acknowledgement' (pp. 287-9; the examples are mine). This last type of
elaboration seems to be of particular interest; but its statement by Cohen is
controversial, in that what is supposed or suggested by him in his formulation
of the 'psychic mechanism' involved is something like a set of unconscious
ideas which the agents in question however, so to speak, deliberately (but also
unconsciously?) reject. This, I assume, is why he uses the term 'selfdeception'. Better in my opinion would be to insist on the principled
impossibility of any such presumed self-knowledge, rather than, as in his
account, supposing that it exists but is hidden by the agents from themselves.
All these questions also have political undertones, to which Althusser's
position could not leave him indifferent. First, Cohen's brief sketch could be
utilized in support of a Kautskyan principle of scientific vanguardism, the
legitimate leaders of the working class being those who, undeceived by either
'bourgeois propaganda' or themselves (and these two states would probably be
linked in most accounts of this sort) may therefore be considered fit to 'direct'
the class - and to declare when revolution has become possible; i.e., when the
productive forces are suficiently developed to allow it to occur and to
succeed, without regression to an earlier form of society (Cohen 1978: 206).
Althusser is of course well-known for his opposition to this version of
Marxism, which is not the same thing as Lenin's doctrine.
Secondly, there is a whole tradition in the Marxist labour movement which
in fact made specific reference not just to functional explanation as a useful
instrument, but explicitly to Lamarck and Darwin. This tradition Cohen does
not treat. But it has been examined by French authors, some in a more or less
Althusserian spirit." It is a fascinating topic in its own right; and a study of it
might throw some light, not only on the line of demarcation between
legitimate and illegitimate elaborations as applied within Marxism, but also on
the possible consequences of extending the functional account to the general
relation between productive forces and production relations that is to say, on
the possible consequences of Cohen's own 'primacy thesis', which is the point
at which, as I suggested, the opposition between his position and that of
Althusser is at its sharpest.
The notion that one might be able to write an autonomous history of the
development of the productive forces, which is what Cohen's project requires,
is problematic, and is so considered by Althusser. We know for instance
from his Reply to John Lewis - that Althusser even considered Stalinism (or
more exactly, what he called the 'Stalin deviation') to be, if not a result, then an
expression of a perverse Marxism in which the idea of the 'primacy of the
productive forces' was given practical application. The theory of class

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Louis Al~husserand G.A . Cohen: a conj?onrarion

5 11

struggle, so he claimed, was expelled (officially in the 1930s) from the centre
of Stalinian theory, to be replaced by a core principle of the primacy of the
productive forces, whose development was treated as a criterion for the
successful advance of socialism. That is one reason why 'Stalinism' is claimed
to be a kind of 'posthumous revenge of the Second International', i.e. of
Kautskyism and suchlike. The expulsion of the category of class struggle except as a legitimizing ideology of policy applied after the event - meant that
Stalinian Marxism had become a humanism, albeit a cruel one. Here we find a
specific employment by Althusser of the idea of a connexion between the
ideologies of economism and humanism: the slogan of the Stalin period, he
notes, was 'Man, the most precious capital'.
The opposition between Cohen and Althusser on the point of the 'primacy
thesis' also of course turns around a difference in their respective conceptions
of the productive forces. Cohen's view is straightforwardly that they are 'not
relations'. That of Althusser/Balibar is that they are a kind of relation of
production: 'a connexion . . . of "real appropriation" between the . . .
elements: means of production, direct producers,. . . [and] non-wageearners' (Althusser and Balibar 1970: 235). This position is further linked to
their (implicit) distinction between 'real' and 'formal' appropriation (of
nature, of the means of production etc.); and this distinction is in turn
connected with their treatment of the 'real' and 'formal' subjection of the
worker, which we already discussed.
According to Althusser/Balibar, then, the productive forces are relations of
real appropriation, or as they sometimes put it, 'technical' social relations of
production. What they mean to suggest by this terminology is in any case that
the productive forces are not (even if Marx sometimes speaks this way) some
set of things like machines, scientific and technological knowledge, more or
less skilled labour power and so on. Balibar writes indeed that such a
conception of the productive forces would lead to all kinds of false ideas:
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. . . . This
leads to an interpretation of the 'level' or 'degree of development' [of the
productive forces] 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."
Here then the biological metaphor is repudiated. More importantly for our
present purpose, this account is seen to be diametrically opposed to that of
Cohen. The AlthussedBalibar definition is an historical one in a different
sense from Cohen's. Their historical definition is provided by the theoretical
insertion of the productive forces into a particular mode of production, and
therefore - in class societies - of exploitation. For if the productive forces are
indeed defined in such terms, then their definition will change with a change in
the (non-technical) social relations of production (as will the definition of

'productive power'). That the means ofproduction (especially the instruments of

production: cf. Cohen 1978: 32, 37-50) can still be identified as separate
objects, i.e. separate from the relation constituting the productive force into
which they are integrated, which is obvious, does not entail that one can after
all write a separate history of their autonomous development, in terms say of
the 'growth in knowledge of how to control and transform nature' or
something similar. T o believe in the possibility of such a history is, in my
earlier stated view," to repeat an error contained in Mam and developed by
Engels, to the effect that human history displays (as the eighteenth-century
philosophes already believed) the progressive extension of the rule of man -in
his generality - over nature. And this conception tends - in any consideration
of the future or fate of capitalism - to reduce the explanatory role of class
struggle, in all its complexities, in the determination of events; and to interpret
the so-called 'necessary transition' to socialism as just one more expression of
an underlying historical and metaphysical dialectic.
In order to illustrate my point I shall move for a moment to a more concrete
level. Cohen is faithful to Mam (that must be said) in counting instruments of
production as productive forces: at least, he is faithful to certain of Marx's
texts. His development thesis requires us to be able to provide an autonomous
history of the development of such instruments. Thus for example, he says, it
may be that in a society whose culture strongly supports one-man navigation, a
canoe is nevertheless invented which, being long, is best operated by two men.
Such an innovation does not necessitate a change in working relationships, for
the new canoes might simply be operated inefficiently. But 'we should expect
. . . a transition to double manning' because 'it is the rational thing . . . and
men are somewhat rational' (pp. 168-9). This story is of course logically
alright. T o invent a technologically superior canoe is, as he says, not
necessarily to invent or to introduce that use which renders it superior in
practice. Everyone may be aware of its technological superiority, and yet
refuse it or otherwise fail to exploit it. (Though, he suggests, they will tend not
to.) Yet it seems to me empirically true of the overwhelming majority of
technological innovations that they were and are introduced with their
application in mind; and that this fact is itself 'not accidental', but has
something to do with the nature of the division of labour under capitalism,
between tasks of planning and execution. Secondly, it appears to me just as
empirically true that such innovations-in-use are not to be explained by
reference to any abstract human rationality, but only in terms of the specific
'rationality' of the capitalist system, which includes the need to retain control
over the working class. An innovation may indeed be introduced for just this
last-named reason, rather than (merely) because it is more productively
efficient. It is true of course that such an explanation addresses itself to a
different problem from that posed centrally in Cohen's book. Still, it is
relevant to the way in which such a problem can be legitimately formulated.
Now the kind of argument which I have outlined in its Althusserian
formulation might be put in other ways, which are not specifically Marxist, let

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Louis Al~krrsserarid G. A. Cohen: a ronfion~atior~ 5 13

alone Althusserian. In one way or another it has even become a commonplace
of industrial sociology. The Tavistock Institute of Human Relations, for
instance, has concerned itself with the problems of designing technology to fit
systems of work organization and patterns of work behaviour. In this
connexion its researchers introduced the notion of the socio-technical system. It
was found, generally speaking, that not only the extent but also the type of
technological innovation possible in any enterprise depended in large part on
'behavioural' and 'psychological' (or what one might call socio-political and
ideological) factors (cf. Rose 1978: part IV). In Marxist terminology: the state
of the class struggle plays a part in determining not just what &gree of
technological rationalization is possible (is there a powerful 'Luddite'
movement? etc.), but also the qualitative course of technological change. In
consequence, it would be impossible simply to measure the extent of
improvement of the instruments of production or of the productive forces and
their eficient or inefficient deployment in quantitative terms, or to establish
thereby that in some unambiguous sense capitalism no longer deploys the
productive forces in an optimally rational way, in Cohen's sense. Moreover,
even if - as he claims capitalism is unnecessarily output-oriented, it does not
follow that a post-capitalist regime could solve that problem by just planning
(cf. infra) a better balance of work and leisure. For that might leave intact the
existing set of 'socio-technical systems', or in other words, what may be called
'capitalist technology'.
What capitalism requires is both (i) maximization of profit or something of
the sort and (ii) reproduction of the conditions necessary to profit-making, i.e.
to the reproduction of the system itself. Now (i) may often (but will not always)
be served by the introduction of 'more efficient' instruments of production; in
any case the individual capitalist firm will often rationally innovate with this
end in mind. But (ii) may nevertheless require that the more efficient
instruments ofproduction still not be introduced. The growing interventionist
tendencies of the State (which continue even in this epoch of neo-liberalism)
may be interpreted as, in part, a means of securing (ii) when the logic of profit
for the individual enterprise would tend to undermine it.
Is this account compatible with Cohen's explanatory approach, allowing for
the argument that the rational man - as long as he is not a capitalist or a
capitalist lackey - will opt for socialism, which, in contrast to capitalism, does
allow the 'most efficient alternative' to be chosen? No, and for at least two
reasons:
(1) In order for the working class to be able to overcome the capitalist
division oflabour, it may be necessary for less eficient options to be adopted, for
a long and indefinite period, such that by no means all workers can reasonably
hope to benefit from the change in 'purely' material terms (which does not
imply that only a few will cherish such hopes, or that a widely-shared hope of
such a change may not play an important role in triggering revolutionary
change). And
(2) No transition to socialism can in any case be explained in terms ofwhat is

5 14 Grakane Lock
'rationally desirable' by, or in accordance with the objective or rational
interests of the working class. Such transitions are, from the point of view of
Althusser, not matters of choice at all, and afortiori not matters of rational
choice. One reason for this has been hinted at above: that a revolutionary
upheaval is likely, if it happens at all, to be the result of a rejection by the
workers (etc.) of their 'formal' subjection to capital - that is, of a kind of
partially blind rebellion, channelled of course (this is the orthodox Leninist
position) by a vanguard party. There is moreover another consideration: the
division of the working population. If it is true that advanced capitalism is
exacerbating the technical and social division of labour, establishing new and
aggravated forms of hierarchy and division within its industrial and technological processes, then it is hard to see how it is possible to locate the
politico-ideological principle of revolution and the transition to socialism in
such things as the (working people's) simple desire, as Cohen puts it, to attain
a 'sufficiency [of goods] produced with a minimum of unpleasant exertion'
Downloaded By: [University of Oxford] At: 14:44 17 February 2010

(p. 307).

One of the great points of Marx's insistence (in his Critique of the Gotha
Programme) on the distinction between the 'first' and 'higher' stages of
communist society is to emphasize the fact that in the first stage 'the enslaving
subordination of the individual to the division of labour' and 'the antithesiq
between mental and physical labour' persist. And should we not expect these
inequalities to lead to conflict, including or even especially in economic
matters, such that 'decisions' concerning economic or political policy would
not be (mere) matters of 'collective choice', but rather the outcomes of
complex struggles, indeed of class struggles? Of course, we might still claim
that it is characteristic of socialist societies that in them planning should be
easier, for some of the anarchy of the capitalist market has been abolished. But
the market is not entirely abolished (for example, the market in labour power
still partially functions). And what 'planning' under socialism can mean is the
legal registration and regulation of value relations which, in (western)
capitalism, find their own level, rather than the abolition of such capitalist
value relations. (One might compare, of course mutatis mutandis, the
difference between free-floating and fmed currency exchange rates.) So the
advent or maintenance of socialism cannot be reduced to a matter of rational
choice; nor can it be said, on this view, that socialism, in contrast to capitalism,
allows the most efficient technological options to be seized on - at least,
nothing of that kind could be said in advance or in principle. 1conclude that
the Althusserian account, sketched above, of the double 'logic of capitalism'
(requiring profit maximization, but under the constraint of the need to
guarantee the reproduction of the capitalist system as a whole) is incompatible
with Cohen's picture of the role of rational choice in the transition to
socialism. And the connected remarks on conflict and class struggle under
socialism clash with his picture of the place of rational choice in socialist
society.
Cohen's book, which stands subjectively almost wholly outside of the crisis

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Lortis Alrhtrsser and G. A. Cohert: a confionrarion

5 15

of Marxism referred to above, is actually located four-square within it. Unlike

the work of Althusser, it seems to stand outside of theoretical as well as
political history. Its author abstracts from the conditions of production of
Marx's ideas and their sources in contemporary scientific ideology, as he
ignores the conditions of production of his own work and its relations to its
ideological 'surroundings'. Yet just because it makes no reference to Kautsky
or Bemstein, Lenin, Stalin or Mao; to the British, US, Canadian, French,
German, Czech, Polish, Czech or any other labour movement; just because it
makes no mention of the Chinese Cultural Revolution and its enormous
impact on western European Marxism - and among others, on Althusser; just
for these reasons it is an exemplary expression of the crisis, which 'dare not
speak its name'; or at least until recently dared not. In the face of this crisis,
Cohen simply returned to first principles.
I have tried to indicate in this article some of the points of conflict between
Cohen and Althusser. The idea of a return to orthodoxy is in this connexion an
interesting one, for both authors have, each in his own way, striven after such a
return, as the titles of their books indicate. Yet their work has rarely been
compared and contrasted. An exception in this respect is a review of Cohen's
book by Althusser's friend and disciple, Dominique Lecourt (1983). Lecourt
writes in particular that Cohen's work is 'one of those books on Marx, which
are so rare, whose grandeur consists in their forcing the reader to rethink the
whole structure of Marxism, from the philosophical foundations up, even if he
goes on to draw conclusions diametrically opposed to those which the author
proposes'. Lecourt goes so far as to acknowledge that 'once one has read
Cohen's book, it seems difficult to deny that such a conception of history as a
whole [viz. the one he attributes to Marx] did indeed dominate Marx's thought
right through'. The question is only whether there is not some other line of
thought in Mam too - for example, the one which Althusser claimed to have
found, and on that basis elaborated.
In this paper too I have suggested that Cohen's answers to certain key
questions of Marxism, and even his formulation of the questions, are
problematic. But the generally critical impression which I have thus given
should not be allowed to belie my admiration for and, sometimes,
astonishment at the delicate but luminous intricacies of the textual developments in his book, and for the care and thoroughness of the exegesis, which
indeed put many a 'continental' Marxist to shame. But I do not think that this
last remark applies to Althusser, for reasons which ought to have become
obvious from the above. In fact his writings contain many equally ingenious
ideas, which if worked out would certainly put Cohen's positions under great
pressure. I have tried to sketch out some of these ideas, though Althusser's
work contains of course many more.
Karl Mum's Theory of History established Cohen as the leading Marxist
philosopher of the Anglophone world. Althusser's publications, in my
opinion, established him in a similar position in France. Althusser then
announced the crisis of Marxism, and that is one of his merits. Cohen for his

part tacitly contributed to the course of this crisis. For, as Isaiah Berlin has
said, clarification may expose the shortcomings of a theory. And that is what
Cohen achieved for one version, at least of Marxism. T h e question now is
whether the crisis will be resolved by recovery or death. But that is another
story - or as Althusser might say, another 'histoire terminke, histoire
interminable'.

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Notes
1 See Althusser (1978a). Althusser was of course not the first to use this phrase, but
he did give it a new content.
2 The present paper is the product of a revision of part of an article published in a
Dutch journal (Acta Politica, Meppel, no. 3, 1981).
3 See in particular Cohen (1983).
4 See for example Althusser (1976: 94-9); Althusser (1978b: 94-6).
5 Cf. Althusser, 'Reply to John Lewis', in Althusser (1976).
6 Cohen (1978: 73-7). Thompson (1978: 298-9), comments however on what he
considers to be the merely apparent similarity between his own and Althusser's
conception of class.
7 Balibar, in Althusser and Balibar (1970: 236-7). Cf. Marx, Capital, vol. I, Part IV,
ch. XVI. His German term is 'Subsumtion'.
8 See Althusser (1976); Balibar, 'Sur la dialectique historique (quelques remarques
B propos de "Lire le Capitaln)', in Balibar (1974).
9 Althusser, 'Ideology and Ideological State Apparatuses (Notes towards an
Investigation)', in Althusser (1971: 147).
10 See for example Pacquot (1980: ch. 3).
11 Balibar, in Althusser and Balibar (1970: 234). This passage should be read in
context in order to appreciate the full sense of the argument.
12 In my (1981).

References
Althusser, Louis and Balibar, Etienne
(1970), Reading Capital, New Left
Books.
Althusser, Louis (197 I), Lenin and
Philosophj~and Other Essals, New Left
Books.
Althusser, Louis (1976), Essa)ls in Seljl
Criticisnr, New Left Books.
Althusser, Louis (1978a), 'The Crisis
of Marxism', Marxism Today,July.
Althusser, Louis (1978b), Ce qrri nepetrt
plrrs drrrer duns le part i contnrrrniste,
Maspero.
Balibar, Etienne (1974), Cinq P t r r h de
~~ratirtalis~re
historiqrre, Maspero.
Balibar, Etienne (1977), On the
Dictatorship of the Proletariat, New Left
Books.

Cohen, G. A (1978), KarlManr's Theoy

ofHistor~~:
u Dejence, Oxford University
Press.
Cohen, G. A. (1980), 'Functional
Explanation: Reply to Elster', Political
StrriPs, no. 1, March.
Cohen, G. A. (1983), 'Reconsidering
Historical Materialism', Nornos, XXVI.
Elster, Jon (1980), 'Cohen on Marx's
Theory of History', PaliticulStrriPs,
no. 1, hlarch.
Lecourt, Dominique (1983),
'Comment "defendre le matbrialisme
historique"?', Rmrrephilosopkiqrre,no. 2.
Levine, Andrew and Wright, Erik
Olin (1980), 'Rationality and Class
Struggle', Mro Left Rmiem, no. 123,
September-October.

Loitis Althusser and G. A . Colzen: a cotfrontnrio~r 517

Downloaded By: [University of Oxford] At: 14:44 17 February 2010

Lock, Grahame (1981). Tlie State and I,

Nijhoff/Brill.
Pacquot, Thierry (1980),Lrs Faiseicrs
de nrrages, Le Sycornore.

Thompson, E. P. (1978), Tile Poreq]~4

Theory, Merlin Press.