A decent introduction to Post Marxism and discourse, proving once again that good theory is no guarantee of good politics.
Ernesto Laclau and Chantal Mouffe: Post-Marxism Without Apologies
Why should we rethink the socialist project today? In Hegemony and Socialist Strategy we pointed out some of the reasons. As participating actors in the history of our time, if we are actually to assume an interventionist role and not to do so blindly, we must attempt to wrest as much light as possible from the struggles in which we participate and from the changes which are taking place before our eyes. Thus, it is again necessary to temper ‘the arms of critique’. The historical reality whereof the socialist project is reformulated today is very different from the one of only a few decades ago, and we will carry out our obligations as socialists and intellectuals only if we are fully conscious of the changes and persist in the effort of extracting all their consequences at the level of theory. The ‘obstinate rigour’ that Leonardo proposed as a rule for intellectual work should be the only guideline in this task; and it leaves no space for complacent sleights of hand that seek only to safeguard an obsolete orthodoxy.
Since we have referred in our book to the most important of these historical transformations, we need do no more here than enumerate them: structural transformations of capitalism that have led to the decline of the classical working class in the post-industrial countries; the increasingly profound penetration of capitalist relations of production in areas of social life, whose dislocatory effects—concurrent with those deriving from the forms of bureaucratization which have characterized the Welfare State—have generated new forms of social protest; the emergence of mass mobilizations in Third World countries which do not follow the classical pattern of class struggle; the crisis and discrediting of the model of society put into effect in the countries of so-called ‘actually existing socialism’, including the exposure of new forms of domination established in the name of the dictatorship of the proletariat.
There is no room here for disappointment. The fact that any reformulation of socialism has to start today from a more diversified, complex and contradictory horizon of experiences than that of fifty years ago—not to mention 1914, 1871 or 1848—is a challenge to the imagination and to political creativity. Hopelessness in this matter is only proper to those who, to borrow a phrase from J. B. Priestley, have lived for years in a fools’ paradise and then abruptly move on to invent a fools’ hell for themselves. We are living, on the contrary, one of the most exhilarating moments of the twentieth century: a moment in which new generations, without the prejudices of the past, without theories presenting themselves as ‘absolute truths’ of History, are constructing new emancipatory discourses, more human, diversified and democratic. The eschatological and epistemological ambitions are more modest, but the liberating aspirations are wider and deeper.
In our opinion, to rethink socialism in these new conditions compels us to undertake two steps. The first is to accept, in all their radical novelty, the transformations of the world in which we live—that is to say, neither to ignore them nor to distort them in order to make them compatible with outdated schemas so that we may continue inhabiting forms of thought which repeat the old formulae. The second is to start from this full insertion in the present—in its struggles, its challenges, its dangers—to interrogate the past: to search within it for the genealogy of the present situation; to recognize within it the presence—at first marginal and blurred—of problems that are ours; and, consequently, to establish with that past a dialogue which is organized around continuities and discontinuities, identifications and ruptures. It is in this way, by making the past a transient and contingent reality rather than an absolute origin, that a tradition is given form.
In our book we attempted to make a contribution to this task, which today starts from different traditions and in different latitudes. In almost all cases we have received an important intellectual stimulus from our reviewers. Slavoj Zizek, for example, has enriched our theory of social antagonisms, pointing out its relevance for various aspects of Lacanian theory.  Andrew Ross has indicated the specificity of our line of argument in relation to several attempts in the United States to address similar problems, and has located it within the general framework of the debate about post-modernity.  Alistair Davidson has characterized the new Marxist intellectual climate of which our book is part.  Stanley Aronowitz has made some interesting and friendly criticisms from the standpoint of the intellectual tradition of the American Left.  Phillip Derbyshire has very correctly underlined the theoretical place of our text in the dissolution of essentialism, both political and philosophical.  David Forgacs has posed a set of important questions about the political implications of our book, which we hope to answer in future works. 
However, there have also been attacks coming—as was to be expected—from the fading epigones of Marxist orthodoxy. In this article we will answer the criticisms of one member of this tradition: Norman Geras.  The reason for our choice is that Geras—in an extremely unusual gesture for this type of literature—has done his homework: he has gone through our text thoroughly and has presented an exhaustive argument in reply. His merits, however, end there. Geras’s essay is well rooted in the literary genre to which it belongs: the pamphlet of denunciation. His opinion about our book is unambiguous: it is ‘profligate’, ‘dissolute’, ‘fatuous’, ‘without regard for normal considerations of logic, of evidence, or of due proportion’; it is ‘shame-faced idealism’, an ‘intellectual vacuum’, ‘obscurantism’, ‘lacking all sense of reasonable constraint’, ‘lacking a proper sense of either measure or modesty’; it indulges in ‘elaborate theoretical sophistries’, in ‘manipulating concepts’ and in ‘tendentious quotations’. After all this, he devotes forty pages (one third of the May–June 1987 issue of New Left Review) to a detailed analysis of such a worthless work. Furthermore, despite the fact that Geras does not know us personally, he is absolutely definite about the psychological motivations that led us to write the book—‘the pressure . . . of age and professional status’; ‘the pressures of the political time . . . not very congenial, in the West at least, to the sustenance of revolutionary ideas’; ‘the lure of intellectual fashion’; ‘so-called realism, resignation or merely candid self-interest’, etc.—conceding, however, that such perverse motivations are perhaps not ‘consciously calculated for advantage’. (Thank you, Geras.) It is, of course, up to the reader to decide what to think about an author who opens an intellectual discussion by using such language and such an avalanche of ad hominem arguments. For our part, we will only say that we are not prepared to enter into a game of invective and counter-invective; we will therefore declare from the start that we do not know the psychological motivations behind Geras’s inspiration to write what he does and that, not being his psychiatrists, we are quite uninterested in them. However, Geras also makes a series of substantive—though not substantial—criticisms of our book, and it is to these aspects of his piece that we shall refer. We shall first consider his critique of our theoretical approach and then move on to his points concerning the history of Marxism and the political issues that our book addresses. Let us start with the central category of our analysis: the concept of discourse.
The number of absurdities and incoherences that Geras has accumulated concerning this point is such that it is simply impossible to use his critical account as the framework for our reply. We will therefore briefly outline our conception of the social space as discursive, and then confront this statement with Geras’s criticisms.
Let us suppose that I am building a wall with another bricklayer. At a certain moment I ask my workmate to pass me a brick and then I add it to the wall. The first act—asking for the brick—is linguistic; the second—adding the brick to the wall—is extralinguistic.  Do I exhaust the reality of both acts by drawing the distinction between them in terms of the linguistic/extralinguistic opposition? Evidently not, because, despite their differentiation in those terms, the two actions share something that allows them to be compared, namely the fact that they are both part of a total operation which is the building of the wall. So, then, how could we characterize this totality of which asking for a brick and positioning it are, both, partial moments? Obviously, if this totality includes both linguistic and non-linguistic elements, it cannot itself be either linguistic or extralinguistic; it has to be prior to this distinction. This totality which includes within itself the linguistic and the non-linguistic, is what we call discourse. In a moment we will justify this denomination; but what must be clear from the start is that by discourse we do not mean a combination of speech and writing, but rather that speech and writing are themselves but internal components of discursive totalities.
Now, turning to the term discourse itself, we use it to emphasize the fact that every social configuration is meaningful. If I kick a spherical object in the street or if I kick a ball in a football match, the physical fact is the same, but its meaning is different. The object is a football only to the extent that it establishes a system of relations with other objects, and these relations are not given by the mere referential materiality of the objects, but are, rather, socially constructed. This systematic set of relations is what we call discourse. The reader will no doubt see that, as we showed in our book, the discursive character of an object does not, by any means, imply putting its existence into question. The fact that a football is only a football as long as it is integrated within a system of socially constructed rules does not mean that it thereby ceases to be a physical object. A stone exists independently of any system of social relations, but it is, for instance, either a projectile or an object of aesthetic contemplation only within a specific discursive configuration. A diamond in the market or at the bottom of a mine is the same physical object; but, again, it is only a commodity within a determinate system of social relations. For that same reason it is the discourse which constitutes the subject position of the social agent, and not, therefore, the social agent which is the origin of discourse—the same system of rules that makes that spherical object into a football, makes me a player.
The existence of objects is independent of their discursive articulation to such a point, that we could make of that mere existence—that is, existence extraneous to any meaning—the point of departure of social analysis. That is precisely what behaviourism, which is the opposite of our approach, does. Anyway, it is up to the reader to decide how we can better describe the building of a wall: whether by starting from the discursive totality of which each of the partial operations is a moment invested with a meaning, or by using such descriptions as: X emitted a series of sounds; Y gave a cubic object to X; X added this cubic object to a set of similar cubic objects; etc.
This, however, leaves two problems unsolved. The first is this: is it not necessary to establish here a distinction between meaning and action? Even if we accept that the meaning of an action depends on a discursive configuration, is not the action itself something different from that meaning? Let us consider the problem from two angles. Firstly, from the angle of meaning. Here the classical distinction is between semantics—dealing with the meaning of words; syntactics—dealing with word order and its consequences for meaning; and pragmatics—dealing with the way a word is actually used in certain speech contexts. The key point is to what extent a rigid separation can be established between semantics and pragmatics—that is, between meaning and use. From Wittgenstein onwards it is precisely this separation which has grown ever more blurred. It has become increasingly accepted that the meaning of a word is entirely context-dependent. As Hanna Fenichel Pitkin points out: ‘Wittgenstein argues that meaning and use are intimately, inextricably related, because use helps to determine meaning. Meaning is learned from, and shaped in, instances of use; so both its learning and its configuration depend on pragmatics . . . Semantic meaning is compounded out of cases of a word’s use, including all the many and varied language games that are played with it; so meaning is very much the product of pragmatics.’  The use of a term is an act—in that sense it forms part of pragmatics; on the other hand, the meaning is only constituted in the contexts of actual use of the term: in that sense its semantics is entirely dependent upon its pragmatics, from which it can be separated—if at all—only analytically. That is to say, in our terminology, every identity or discursive object is constituted in the context of an action. But, if we focus on the problem from the other angle, every non-linguistic action also has a meaning and, therefore, we find within it the same entanglement of pragmatics and semantics that we find in the use of words. This leads us again to the conclusion that the distinction between linguistic and non-linguistic elements does not overlap with the distinction between ‘meaningful’ and ‘not meaningful’, since the former is a secondary distinction that takes place within meaningful totalities.
The other problem to be considered is the following: even if we assume that there is a strict equation between the social and the discursive, what can we say about the natural world, about the facts of physics, biology or astronomy that are not apparently integrated in meaningful totalities constructed by men? The answer is that natural facts are also discursive facts. And they are so for the simple reason that the idea of nature is not something that is already there, to be read from the appearances of things, but is itself the result of a slow and complex historical and social construction. To call something a natural object is a way of conceiving it that depends upon a classificatory system. Again, this does not put into question the fact that this entity which we call stone exists, in the sense of being present here and now, independently of my will; nevertheless the fact of its being a stone depends on a way of classifying objects that is historical and contingent. If there were no human beings on earth, those objects that we call stones would be there nonetheless; but they would not be ‘stones’, because there would be neither mineralogy nor a language capable of classifying them and distinguishing them from other objects. We need not stop for long on this point. The entire development of contemporary epistemology has established that there is no fact that allows its meaning to be read transparently. For instance Popper’s critique of verificationism showed that no fact can prove a theory, since there are no guarantees that the fact cannot be explained in a better way—therefore, determined in its meaning—by a later and more comprehensive theory. (This line of thought has gone far beyond the limits of Popperism; we could mention the advance represented by Kuhn’s paradigms and by Feyerabend’s epistemological anarchism.) And what is said of scientific theories can be applied to everyday languages that classify and organize objects.
Geras’s Four Theses
We can now go to Geras’s criticisms. They are structured around four basic theses: (1) that the distinction between the discursive and the extra-discursive coincides with the distinction between the fields of the spoken, written and thought, on the one hand, and the field of an external reality on the other; (2) that affirming the discursive character of an object means to deny the existence of the entity designated by that discursive object; (3) that denying the existence of extra-discursive points of reference is to fall in the bottomless abyss of relativism; (4) that affirming the discursive character of every object is to incur one of the most typical forms of idealism. Let us see.
We can treat the first two claims together. Geras writes: ‘Every object is constituted as an object of discourse means all objects are given their being by, or are what they are by virtue of, discourse; which is to say (is it not?) that there is no pre-discursive objectivity or reality, that objects not spoken, written or thought about do not exist.’  To the question posed between brackets ‘(is it not?)’, the answer is simply ‘no, it is not’. The reader who has followed our text to this point will have no difficulty in understanding why. For—returning to our previous example—whether this stone is a projectile, or a hammer, or an object of aesthetic contemplation depends on its relations with me—it depends, therefore, on precise forms of discursive articulation—but the mere existence of the entity stone, the mere material and existential substratum does not. That is, Geras is making an elementary confusion between the being (esse) of an object, which is historical and changing, and the entity (ens) of that object which is not. Now, in our interchange with the world, objects are never given to us as mere existential entities; they are always given to us within discursive articulations. Wood will be raw material or part of a manufactured product, or an object for contemplation in a forest, or an obstacle that prevents us from advancing; the mountain will be protection from enemy attack, or a place for a touring trip, or the source for the extraction of minerals, etc. The mountain would not be any of these things if I were not here; but this does not mean that the mountain does not exist. It is because it exists that it can be all these things; but none of them follows necessarily from its mere existence. And as a member of a certain community, I will never encounter the object in its naked existence—such a notion is a mere abstraction; rather, that existence will always be given as articulated within discursive totalities. The second mistake Geras makes is that he reduces the discursive to a question of either speech, writing or thought, while our text explicitly affirms that, as long as every nonlinguistic action is meaningful, it is also discursive. Thus, the criticism is totally absurd; it involves changing our concept of discourse midstream in the argument, and establishing an arbitrary identification between the being of an object and its existence. With these misrepresentations it is very easy, evidently, to attribute imaginary inconsistencies to our text.
The third criticism—relativism—does not fare any better. Firstly, ‘relativism’ is, to a great extent, an invention of the fundamentalists. As Richard Rorty has pointed our: ‘“Relativism” is the view that every belief on a certain topic, or perhaps about any topic, is as good as every other. No one holds this view . . . The philosophers who get called “relativists” are those who say that the grounds for choosing between such opinions are less algorithmic than had been thought . . . So the real issue is not between people who think one view as good as another and people who do not. It is between those who think our culture, or purpose, or intuitions cannot be supported except conversationally, and people who still hope for other sorts of support.’  Relativism is, actually, a false problem. A ‘relativist’ position would be one which affirmed that it is the same to think ‘A is B’ or ‘A is not B’; that is to say, that it is a discussion linked to the being of the objects. As we have seen, however, outside of any discursive context objects do not have being; they have only existence. The accusation of the ‘anti-relativist’ is, therefore, meaningless, since it presupposes that there is a being of things as such, which the relativist is either indifferent to or proclaims to be inaccessible. But, as we have argued, things only have being within a certain discursive configuration, or ‘language game’, as Wittgenstein would call it. It would be absurd, of course, to ask oneself today if ‘being a projectile’ is part of the true being of the stone (although the question would have some legitimacy within Platonic metaphysics); the answer, obviously, would be: it depends on the way we use stones. For the same reason it would be absurd to ask oneself if, outside all scientific theory, atomic structure is the ‘true being’ of matter—the answer will be that atomic theory is a way we have of classifying certain objects, but that these are open to different forms of conceptualization that may emerge in the future. In other words, the ‘truth’, factual or otherwise, about the being of objects is constituted within a theoretical and discursive context, and the idea of a truth outside all context is simply nonsensical.
Let us conclude this point by identifying the status of the concept of discourse. If the being—as distinct from existence—of any object is constituted within a discourse, it is not possible to differentiate the discursive, in terms of being, from any other area of reality. The discursive is not, therefore, an object among other objects (although, of course, concrete discourses are) but rather a theoretical horizon. Certain questions concerning the notion of discourse are, therefore, meaningless because they can be made only about objects within a horizon, not about the horizon itself. The following remark of Geras’s must be included within this category: ‘One could note again, for instance, how absolutely everything—subjects, experience, identities, struggles, movements—has discursive “conditions of possibility”, while the question as to what may be the conditions of possibility of discourse itself, does not trouble the authors so much as to pause for thought.’  This is absurd. If the discursive is coterminous with the being of objects—the horizon, therefore, of the constitution of the being of every object—the question about the conditions of possibility of the being of discourse is meaningless. It is equivalent to asking a materialist for the conditions of possibility of matter, or a theist for the conditions of possibility of God.
Idealism and Materialism
Geras’s fourth criticism concerns the problem of idealism and we have to consider it in a more detailed way. The first condition for having a rational discussion, of course, is that the meaning of the terms one is using should be clear. Conceptual elucidation of the idealism/materialism opposition is particularly important in view not only of the widely differing contexts in which it has been used, but also of the fact that these contexts have often overlapped and so led to innumerable confusions. The idealism/materialism opposition has been used in attempts to refer to, roughly speaking, three different types of problem.
(1) The problem of the existence or non-existence of a world of objects external to thought. This is a very popular mistake which Geras incurs throughout his discussion. For the distinction here is not between idealism and materialism, but between idealism and realism. A philosophy such as Aristotle’s, for example, which certainly is not materialist in any possible sense of the term, is clearly realist. The same can be said of the philosophy of Plato, since for him the Ideas exist in a heavenly place, where the mind contemplates them as something external to itself. In this sense, the whole of ancient philosophy was realist, since it did not put into question the existence of a world external to thought—it took it for granted. We have to reach the modern age, with a philosophy such as Berkeley’s, to find a total subordination of external reality to thought. However, it is important to realize that in this sense Hegel’s absolute idealism, far from denying the reality of an external world, is its unequivocal affirmation. As Charles Taylor has asserted: ‘This (absolute idealism) is paradoxically very different from all other forms of idealism, which tend to the denial of external reality, or material reality. In the extreme form of Berkeley’s philosophy, we have a denial of matter in favour of a radical dependence on the mind—of course God’s, not ours. Hegel’s idealism, far from being a denial of external material reality, is the strongest affirmation of it; it not only exists but necessarily exists.’  If this is the question at issue our position is, therefore, unequivocally realist, but this has little to do with the question of materialism.
(2) What actually distinguishes idealism from materialism is its affirmation of the ultimately conceptual character of the real; for example, in Hegel, the assertion that everything that is real is rational. Idealism, in its sense of opposition to materialism and not to realism, is the affirmation not that there do not exist objects external to the mind, but rather that the innermost nature of these objects is identical to that of mind—that is to say, that it is ultimately thought. (Not thought of individual minds, of course; not even of a transcendent God, but objective thought.) Now, even if idealism in this second sense is only given in a fully coherent and developed form in Hegel, philosophers of antiquity are also predominantly idealist. Both Plato and Aristotle identified the ultimate reality of an object with its form—that is, with something ‘universal’, and hence conceptual. If I say that this object which is in front of me is rectangular, brown, a table, an object, etc. each of these determinants could also be applied to other objects—they are then ‘universals’, that is form. But what about the individual ‘it’ that receives all these determinations? Obviously, it is irrational and unknowable, since to know it would be to subsume it under a universal category. This last individual residue, which is irreducible to thought, is what the ancient philosophers called matter. And it was precisely this last residue which was eliminated by a consistent idealist philosophy such as Hegel’s: it asserted the ultimate rationality of the real and thus became absolute idealism.
Thus, form is, at the same time, both the organizing principle of the mind and the ultimate reality of an object. As it has been pointed out, form ‘cut(s) across the categories of epistemology and ontology for the being of the particular is itself exhaustively defined according to the requirements of knowledge . . . Thought, word and thing are defined in relation to thinkable form, and thinkable form is itself in a relation of reciprocal definition with the concept of entity.’  The true line of divide between idealism and materialism is, therefore, the affirmation or negation of the ultimate irreducibility of the real to the concept. (For example, a philosophy such as that of the early Wittgenstein, which presented a picture theory of language in which language shared the same ‘logical form’ as the thing, is entirely within the idealist field.)
It is important to note that, from this point of view, what has been traditionally called ‘materialism’ is also to a great extent idealist. Hegel knew this so well that in his Greater Logic materialism is presented as one of the first and crudest forms of idealism, since it assumes identity between knowledge and being. (See Greater Logic, First Section, Chapter Two, final ‘remark’.) Commenting on this passage, W.T. Stace points out: ‘Atomism alleges that this thing, the atom, is the ultimate reality. Let it be so. But what is this thing? It is nothing but a congeries of universals, such perhaps as “indestructible”, “indivisible”, “small”, “round”, etc. All these are universals, or thoughts. “Atom” itself is a concept. Hence even out of this materialism proceeds idealism.’  Where, in all this, does Marx fit in? The answer cannot be unambiguous. In a sense, Marx clearly remains within the idealist field—that is to say, within the ultimate affirmation of the rationality of the real. The well-known inversion of dialectics cannot but reproduce the latter’s structure. To affirm that the ultimate law of motion of History is given not by the change of ideas in the minds of human beings but rather by the contradiction, in each stage, between the development of productive forces and the existing relations of production, does not modify things at all. For what is idealist is not the affirmation that the law of motion of History is the one rather than the other, but the very idea that there is an ultimate law of motion that can be conceptually grasped. To affirm the transparency of the real to the concept is equivalent to affirming that the real is ‘form’. For this reason the most determinist tendencies within Marxism are also the most idealist, since they have to base their analyses and predictions on inexorable laws which are not immediately legible in the surface of historical life; they must base themselves on the internal logic of a closed conceptual model and transform that model into the (conceptual) essence of the real.
(3) This is not, however, the whole story. In a sense which we have to define more precisely, there is in Marx a definite movement away from idealism. But before we discuss this, we must characterize the structure and implications of any move away from idealism. As we have said, the essence of idealism is the reduction of the real to the concept (the affirmation of the rationality of the real or, in the terms of ancient philosophy, the affirmation that the reality of an object—as distinct from its existence—is form). This idealism can adopt the structure which we find in Plato and Aristotle—the reduction of the real to a hierarchical universe of static essences; or one can introduce movement into it, as Hegel does—on condition, of course, that it is movement of the concept and thus remains entirely within the realm of form. However, this clearly indicates that any move away from idealism cannot but systematically weaken the claims of form to exhaust the reality of the object (i.e. the claims of what Heidegger and Derrida have called the ‘metaphysics of presence’). But, this weakening cannot merely involve an affirmation of the thing’s existence outside thought, since this ‘realism’ is perfectly compatible with idealism in our second sense. As has been pointed out, ‘what is significant from a deconstructive viewpoint is that the sensible thing, even in a “realist” like Aristotle, is itself unthinkable except in relation to intelligible form. Hence the crucial boundary for Aristotle, and for philosophy generally, does not pass between thought and thing but within each of these, between form and formlessness or indefiniteness.’ 
The Instability of Objects
Thus, it is not possible to abandon idealism by a simple appeal to the external object, since (1) this is compatible with the affirmation that the object is form and thus remains within the field of idealism and the most traditional metaphysics; and (2) if we take refuge in the object’s mere ‘existence’, in the ‘it’ beyond all predication, we cannot say anything about it. But here another possibility opens up at once. We have seen that the ‘being’ of objects is different from their mere existence, and that objects are never given as mere ‘existences’ but are always articulated within discursive totalities. But in that case it is enough to show that no discursive totality is absolutely self-contained—that there will always be an outside which distorts it and prevents it from fully constituting itself—to see that the form and essence of objects are penetrated by a basic instability and precariousness, and that this is their most essential possibility. This is exactly the point at which the movement away from idealism starts.
Let us consider the problem more closely. Both Wittgenstein and Saussure broke with what can be called a referential theory of meaning—i.e., the idea that language is a nomenclature which is in a one-to-one relation to objects. They showed that the word ‘father’, for instance, only means what it does because the words ‘mother’, ‘son’, etc. also exist. The totality of language is, therefore, a system of differences in which the identity of the elements is purely relational. Hence, every individual act of signification involves the totality of language (in Derridean terms, the presence of something always has the traces of something else which is absent). This purely relational or differential character is not, of course, exclusive to linguistic identities but holds for all signifying structures—that is to say, for all social structures. This does not mean that everything is language in the restricted sense of speech or writing, but rather that the relational or differential structure of language is the same for all signifying structures. So, if all identity is differential, it is enough that the system of differences is not closed, that it suffers the action of external discursive structures, for any identity (i.e., the being, not the existence of things) to be unstable. This is what shows the impossibility of attributing to the being of things the character of a fixed essence, and what makes possible the weakening of form, which constituted the cornerstone of traditional metaphysics. Human beings socially construct their world, and it is through this construction—always precarious and incomplete—that they give to a thing its being.  There is, then, a third meaning of the idealism/materialism opposition which is related neither to the problem of the external existence of objects, nor to a rigid counterposition of form and matter in which the latter is conceived as the ‘individual existent’. In this third opposition, a world of fixed forms constituting the ultimate reality of the object (idealism) is challenged by the relational, historical and precarious character of the world of forms (materialism). For the latter, therefore, there is no possibility of eliminating the gap between ‘reality’ and ‘existence’. Here, strictly speaking, there are two possible conceptual strategies: either to take ‘idealism’ and ‘materialism’ as two variants of ‘essentialism’; or to consider that all essentialism, by subordinating the real to the concept, is idealism, and to see materialism as a variety of attempts to break with this subordination. Both strategies are, of course, perfectly legitimate.
Let us return at this point to Marx. There is in his work the beginning, but only the beginning, of a movement in the direction of materialism. His ‘materialism’ is linked to a radical relationalism: ideas do not constitute a closed and self-generated world, but are rooted in the ensemble of material conditions of society. However, his movement towards relationalism is weak and does not actually transcend the limits of Hegelianism (an inverted Hegelianism continues to be Hegelian). Let us look at these two moments:
(1) One possible way of understanding this embeddedness of ideas in the material conditions of society would be in terms of signifying totalities. The ‘State’ or the ‘ideas’ would not be self-constituted identities but rather ‘differences’ in the Saussurean sense, whose only identity is established relationally with other differences such as ‘productive forces’, ‘relations of production’, etc. The ‘materialist’ advance of Marx would be to have shown that the area of social differences which constitutes the signifying totalities is much wider and deeper than it had been supposed hitherto; that the material reproduction of society is part of the discursive totalities which determine the meaning of the most ‘sublime’ forms of political and intellectual life. This allows us to overcome the apparently insoluble problems concerning the base/superstructure relation: if State, ideas, relations of production, etc. have purely differential identities, the presence of each would involve the presence of the others—as the presence of ‘father’ involves the presence of ‘son’, ‘mother’, etc. In this sense, no causal theory about the efficacy of one element over another is necessary. This is the intuition that lies behind the Gramscian category of ‘historical bloc’: historical movement is explained not by laws of motion of History but by the organic link between base and superstructure.
(2) However, this radical relationalism of Marx is immediately translated into idealistic terms. ‘It is not the consciousness of men that determines their existence, but their social existence that determines their consciousness.’  This could be read, of course, as a reintegration of consciousness with existence, but the expression could not be more unfortunate, since if social existence determines consciousness, then consciousness cannot be part of social existence.  And when we are told that the anatomy of civil society is political economy, this can only mean that there is a specific logic—the logic of the development of productive forces—which constitutes the essence of historical development. In other words, historical development can be rationally grasped and is therefore form. It is not surprising that the ‘Preface’ to the Critique of Political Economy depicts the outcome of the historical process exclusively in terms of the contradiction between productive forces and relations of production; nor is it surprising that class struggle is entirely absent from this account. All this is perfectly compatible with the basic premises of Hegelianism and metaphysical thought.
Let us now sum up our argument in this section. (1) The idealism/realism opposition is different from the idealism/materialism opposition. (2) Classical idealism and materialism are variants of an essentialism grounded on the reduction or the real to form. Hegel is, therefore, perfectly justified in regarding materialism as an imperfect and crude form of idealism. (3) A move away from idealism cannot be founded on the existence of the object, because nothing follows from this existence. (4) Such a move must, rather, he founded on a systematic weakening of form, which consists in showing the historical, contingent and constructed character of the being of objects; and in showing that this depends on the reinsertion of that being in the ensemble of relational conditions which constitute the life of a society as a whole. (5) In this process, Marx constitutes a transitional point: on the one hand, he showed that the meaning of any human reality is derived from a world of social relations much vaster than had previously been perceived; but on the other hand, he conceived the relational logic that links the various spheres in clearly essentialist or idealistic terms.
A first sense of our post-Marxism thus becomes clear. It consists in a deepening of that relational moment which Marx, thinking within a Hegelian and, in any case, nineteenth-century matrix, could only take so far. In an age when psychoanalysis has shown that the action of the unconscious makes all signification ambiguous; when the development of structural linguistics has enabled us to understand better the functioning of purely differential identities; when the transformation of thought—from Nietzsche to Heidegger, from pragmatism to Wittgenstein—has decisively undermined philosophical essentialism, we can reformulate the materialist programme in a much more radical way than was possible for Marx.
At this point we should consider Geras’s general methodological reproach that we have based our main theoretical conclusions on a false and rigid ‘either/or’ opposition; that is to say, that we have counterposed two polar and exclusive alternatives, without considering the possibility of intermediate solutions that avoid both extremes. Geras discusses this supposed theoretical mistake in relation to three points: our analysis of the concept of ‘relative autonomy’; our treatment of Rosa Luxemburg’s text on the mass strike; and our critique of the concept of ‘objective’ interest. As we will show, in all three cases Geras’s criticism is based on a misrepresentation of our argument.
Firstly, ‘relative autonomy’. Geras quotes a passage of our book where we sustain, according to him, that ‘either the basic determinants explain the nature, as well as the limits, of that which is supposed to be relatively autonomous, so that it is not really autonomous at all; or it is, flatly, not determined by them and they cannot be basic determinants . . . Laclau and Mouffe here deny to Marxism the option of a concept like relative autonomy. No wonder that it can only be for them the crudest sort of economism.’  Geras proposes, instead, the elimination of this ‘inflexible alternative’. If, for example, his ankle is secured to a stout post by a chain he may not be able to attend a political meeting or play tennis, but he can still read and sing. Between total determination and partial limitation there is a whole range of intermediate possibilities. Now, it is not very difficult to realize that the example of the chain is perfectly irrelevant to what Geras intends to demonstrate, since it involves no more than a sleight of hand whereby a relation of determination is transformed into a relation of limitation. Our text does not assert that the State in capitalist society is not relatively autonomous, but rather, that we cannot conceptualize ‘relative autonomy’ by starting from a category such as ‘determination in the last instance by the economy’. Geras’s example is irrelevant because it is not an example of a relation of determination: the chain tied to his ankle does not determine that Geras reads or sings; it only limits his possible movements—and, presumably, this limitation has been imposed against Geras’s will. Now, the base/superstructure model affirms that the base not only limits but determines the superstructure, in the same way that the movements of a hand determine the movements of its shadow on a wall. When the Marxist tradition affirms that a State is ‘capitalist’, or that an ideology is ‘bourgeois’, what is being asserted is not simply that they are in chains or prisoners of a type of economy or a class position, but rather that they express or represent the latter at a different level. Lenin, who, unlike Geras, knew what a relation of determination is, had an instrumentalist theory of the State. His vision is, no doubt, a simplistic one, but it has a considerably higher degree of realism than the chain of Geras, the latter seeming to suggest that the capitalist state is a prisoner limited by the mode of production in what otherwise would have been its spontaneous movements.
What our book asserts is not that the autonomy of the State is absolute, or that the economy does not have any limiting effect vis-`-vis the State’s action, but rather that the concepts of ‘determination in the last instance’ and ‘relative autonomy’ are logically incompatible. And, when we are dealing with logical matters, alternatives are of the either/or type. This is what we have to show. In order to do so let us put ourselves in a situation most favourable to Geras: we will take as an example not a ‘vulgar’ Marxism but a ‘distinguished’ Marxism, one that avoids crude economists and introduces all imaginable sophistication in thinking the base/superstructure relation. What conceptual instruments does such Marxism have to construct the concept of ‘relative autonomy’ starting from the concept of ‘determination in the last instance’? We can only think of two types of attempt:
(1) It might be argued that the base determines the superstructure not in a direct way but through a complex system of mediations. Does this allow us to think the concept of ‘relative autonomy’? By no means. ‘Mediation’ is a dialectical category; even more: it is the category out of which dialectics is constituted, and belongs, therefore, to the internal movement of the concept. Two entities that are related (and constituted) via mediations are not, strictly speaking, separate entities: each is an internal moment in the self-unfolding of the other. We can extend the field of mediations as much as we want: in this way we would give a less simplistic vision of social relations, but we would not advance a single step in the construction of the concept of relative autonomy. This is because autonomy—relative or not—means self-determination; but if the identity of the supposedly autonomous entity is constituted by its location within a totality, and this totality has an ultimate determination, the entity in question cannot be autonomous. According to Lukács, for instance, facts only acquire meaning as moments or determinations of a totality; it is within this totality—which could be as rich in mediations as we want—that the meaning of any identity is established. The exteriority that a relation of autonomy would require is therefore absent.
(2) So, let us abandon this attempt to use the concept of mediation and try instead a second line of defence of the logical compatibility of the two concepts. Could we, perhaps, assert that the superstructural entity is effectively autonomous—that is to say, that no system of mediations links it to the base—and that determination in the last instance by the economy is reduced to the fact that the latter always fixes the limits of autonomy (i.e., that the possibility of Geras’s hair growing as Samson’s to the point that he would be able to break the chain, is excluded)? Have we made any advance with this new solution? No; we are exactly at the same point as before. The essence of something is the ensemble of necessary characteristics which constitute its identity. Thus, if it is an apriori truth that the limits of autonomy are always fixed by the economy, then such limitation is not external to that entity but is part of its essence. The autonomous entity is an internal moment of the same totality in which the determination in the last instance is constituted—and hence there is no autonomy. (All this reasoning is, actually, unnecessary. To affirm at the same time that the intelligibility of the social whole proceeds from an ultimate determination, and that there are internal entities to that totality which escape that determination, was inconsistent from the beginning.)
Autonomy and Determination
What happens if, instead, we abandon the concept of ‘determination in the last instance by the economy’? It does not follow either that the autonomy is absolute, or that the ‘economy’ in a capitalist society does not impose fundamental structural limits on what can be done in other spheres. What does follow is (a) that the limitation and interaction between spheres cannot be thought in terms of the category of ‘determination’; and (b) that there is no last instance on the basis of which society can be reconstructed as a rational and intelligible structure, but rather that the relative efficacy of each sphere depends on an unstable relation of antagonistic forces which entirely penetrates the social. For example, the structure of capitalist relations of production in a certain moment will impose limits on income distribution and access to consumer goods; but conversely, factors such as working-class struggles or the degree of union organization will also have a limiting effect on the rate of profit that can be obtained in a political and economic conjuncture. In our book we made reference to something that has been shown by numerous recent studies: namely, that the transition from absolute to relative surplus value, far from being the simple outcome of the internal logic of capital accumulation, is, to a large extent, the result of the efficacy of working-class struggles. That is to say, the economic space itself is structured as a political space, and the ‘war of position’ is not the superstructural consequence of laws of motion constituted outside it. Rather, such laws penetrate the very field of what was traditionally called the ‘base’ or ‘infrastructure’. If determination was a last instance, it would be incompatible with autonomy, because it would be a relation of omnipotence. But, on the other hand, an absolutely autonomous entity would be one which did not establish an antagonistic relation with anything external to it, since for an antagonism to be possible, a partial efficacy of the two opposing forces is a prerequisite. The autonomy which both of them enjoy will therefore always be relative.
Our book states this clearly in the same paragraph which Geras quotes: ‘If . . . we renounce the hypothesis of a final closure of the social, it is necessary to start from a plurality of political and social spaces which do not refer to any ultimate unitarian basis. Plurality is not the phenomenon to be explained, but the starting point of the analysis. But if, as we have seen, the identity of these spaces is always precarious, it is not possible simply to affirm the equation between autonomy and dispersion. Neither total autonomy nor total subordination is, consequently, a plausible solution.’  The suggestion that we have set up a rigid alternative between total autonomy and absolute subordination is, therefore, simply an invention by Geras. All our analyses try, on the contrary, to overcome that ‘either/or’ alternative—see, for instance, our critique of the symmetrical essentialisms of the totality and the elements (pp. 103–105), or our discussion of the concept of representation (pp. 119–22). In order to overcome the alternative, however, it is necessary to construct a new terrain that goes beyond its two terms, and this implies a break with metaphysical categories such as the ‘last instance’ of the social. Geras also tries, apparently, to overcome this alternative, but he only proceeds by the trick of affirming determination in the last instance theoretically whilst eliminating it in the concrete example that he gives (the one of the chain). His overcoming of the alternative is, therefore, wishful thinking, and his discourse is lodged in permanent incoherence.
Geras’s other two examples of our ‘either/or’ reductionism can be discussed briefly, since they repeat the same argumentative strategy—and the same mistakes. Firstly, the case of Rosa Luxemburg. Geras quotes a fragment of our book where, according to him, we affirm that Marxism rests upon a well-known alternative: ‘either capitalism leads through its necessary laws to proletarianization and crisis; or else these necessary laws do not function as expected, in which case . . . the fragmentation between different subject positions ceases to be an “artificial product” of the capitalist state and becomes a permanent reality.’ On which Geras comments: ‘It is another stark antithesis. Either pure economic necessity bears the full weight of unifying the working class; or we simply have fragmentation.’  This time, however, Geras has omitted a ‘small’ detail in his quotation; and his misquotation is so flagrant that he puts us—this time for sure—before the ‘either/or’ alternative of having to conclude that he is intellectually either irresponsible or dishonest. The ‘detail’ is that our text poses this alternative, not in respect of Marxism in general, but in respect of what would be, by reductio ad absurdum, their extreme reductionist or essentialist versions. The quotation comes from a passage where, after having pointed out the presence of a double historical logic in the text of Rosa Luxemburg—the logic of structural determinism and the logic of spontaneism—we proceeded to what we called an ‘experiment of frontiers’. That is to say, we tried to see what logical consequences would follow from an imaginary extension of the operative area of either determinism or spontaneism. Thus we pointed out that it is only if Marxist discourse becomes exclusively determinist (that is, only in the imaginary case of our experiment) that the iron alternative to which Geras refers is posed. Our book presented the history of Marxism, on the contrary, as a sustained effort to escape the ‘either/or’ logic of determinism. It is exactly in these terms that we refer to the increasing centrality and area of operativity of the concept of ‘hegemony’. In fact, the second step of our experiment—the moving of frontiers in a direction that expands the logic of spontaneism—is conducive to the political alternatives which our text suggests, and which are very different from those possible within a determinist model.
Misquotations apart, it is interesting to see how Geras himself attempts to escape the ‘either/or’ alternative. As in the case of relative autonomy, his solution is a mixture of journalistic impressionism and theoretical inconsistency. (It is significant that, despite his insulting and aggressive tone, Geras is suspiciously defensive and moderate when it comes to presenting his own political and theoretical proposals.) ‘Why,’ he asks ‘may we not think that between this devil and that blue sea there is something else: notwithstanding the wide diversity, a common structural situation, of exploitation, and some common features, like lack of autonomy and interest at work, not to speak of sheer unpleasantness and drudgery, and some pervasive economic tendencies, proletarianizing ones among them, and such also as create widespread insecurity of employment; all this providing a solid, objective basis—no more, but equally no less—for a unifying socialist politics? Why may we not?’  Why may we not indeed? All these things happen under capitalism, in addition to some more things that Geras omits to mention: imperialist exploitation, increasing marginalization of vast sectors of the population in the Third World and in the decaying inner cities of the post-industrial metropolis, ecological struggles against pollution of the environment, struggles against different forms of racial and sexual discrimination, etc. If it is a matter of enumerating the unpleasant features of the societies in which we live, which are the basis for the emergence of numerous antagonisms and contesting collective identities, the enumeration has to be complete. But if it is a matter, on the contrary, of answering such fragmentation with a theory of the necessary class nature of anti-capitalist agents, no mere descriptive enumeration will do the trick. Geras’s ‘classist’ alternative is constituted only by means of interrupting at a certain point his enumeration of the collective antagonisms generated by late capitalism. The vacuity of this exercise is obvious. If Geras wants to found ‘classism’ on something other than the determinism of ‘necessary laws of history’, he has to propose a theoretical alternative of which there is not the slightest sign in his article.
Finally the question of ‘objective interests’. Ours is a criticism not of the notion of ‘interests’ but of their supposedly objective character: that is to say, of the idea that social agents have interests of which they are not conscious. To construct an ‘interest’ is a slow historical process, which takes place through complex ideological, discursive and institutional practices. Only to the extent that social agents participate in collective totalities are their identities constructed in a way that makes them capable of calculating and negotiating with other forces. ‘Interests’, then, are a social product and do not exist independently of the consciousness of the agents who are their bearers. The idea of an ‘objective interest’ presupposes, instead, that social agents, far from being part of a process in which interests are constructed, merely recognize them—that is to say, that those interests are inscribed in their nature as a gift from Heaven. How it is possible to make this vision compatible with a non-essentialist conception of the social, only God and Geras know. Again, we are not dealing with an ‘either/or’ alternative. There are interests, but these are precarious historical products which are always subjected to processes of dissolution and redefinition. What there are not, however, are objective interests, in the sense in which they are postulated in the ‘false consciousness’ approach.
The History of Marxism
Let us move now to Geras’s criticisms of our analysis of the history of Marxism. The centrality we give to the category of ‘discourse’ derives from our attempt to emphasize the purely historical and contingent character of the being of objects. This is not a fortuitous discovery which could have been made at any point in time; it is, rather, deeply rooted in the history of modern capitalism. In societies which have a low technological level of development, where the reproduction of material life is carried out by means of fundamentally repetitive practices, the ‘language games’ or discursive sequences which organize social life are predominantly stable. This situation gives rise to the illusion that the being of objects, which is a purely social construction, belongs to things themselves. The idea of a world organized through a stable ensemble of essential forms, is the central presupposition in the philosophies of Plato and Aristotle. The basic illusion of metaphysical thought resides precisely in this unawareness of the historicity of being. It is only in the contemporary world, when technological change and the dislocating rhythm of capitalist transformation constantly alter the discursive sequences which construct the reality of objects, that the merely historical character of being becomes fully visible. In this sense, contemporary thought as a whole is, to a large extent, an attempt to cope with this increasing realization, and the consequent moving away from essentialism. In Anglo-American thought we could refer to the pragmatist turn and the anti-essentialist critique of post-analytic philosophy, starting from the work of the later Wittgenstein; in continental philosophy, to Heidegger’s radicalization of phenomenology and to the critique of the theory of the sign in post-structuralism. The crisis of normative epistemologies, and the growing awareness of the nonalgorithmic character of the transition from one scientific paradigm to another, point in the same direction.
What our book seeks to show is that this history of contemporary thought is also a history internal to Marxism; that Marxist thought has also been a persistent effort to adapt to the reality of the contemporary world and progressively to distance itself from essentialism; that, therefore, our present theoretical and political efforts have a genealogy which is internal to Marxism itself. In this sense we thought that we were contributing to the revitalization of an intellectual tradition. But the difficulties here are of a particular type which is worth discussing. The article by Geras is a good example. We learn from it, with amazement, that Bernstein and Sorel ‘abandoned’ Marxism—and in Geras this has the unmistakable connotation of betrayal. What can we think about this ridiculous story of ‘betrayal’ and ‘abandonment’? What would one make of a history of philosophy which claimed that Aristotle betrayed Plato, that Kant betrayed Leibnitz, that Marx betrayed Hegel? Obviously, we would think that for the writer who reconstructs history in that way, the betrayed doctrine is an object of worship. And if we are dealing with a religious object, any dissidence or attempt to transform or to contribute to the evolution of that theory would be considered as apostasy. Most supporters of Marxism affirm its ‘scientific’ character. Science appears as separated by an absolute abyss from what mortal men think and do—it coincides with the distinction between the sacred and the profane. At a time when the philosophy of science is tending to narrow the epistemological gap between scientific and everyday languages, it seems deplorable that certain sectors of Marxism remain anchored to an image of science which is more appropriate to popular manuals from the age of positivism.
But this line of argument does not end here. Within this perspective the work of Marx becomes an origin—that is to say, something which contains within itself the seed of all future development. Thus, any attempt to go beyond it must be conceptualized as ‘abandonment’. We know the story very well: Bernstein betrayed Marx; European socialdemocracy betrayed the working class; the Soviet bureaucracy betrayed the revolution; the Western European Communist parties betrayed their revolutionary vocation; thus, the only trustees of ‘Revolution’ and ‘Science’ are the small sects belonging to imaginary Internationals which, as they suffer from what Freud called the ‘narcissism of small differences’, are permanently splitting. The bearers of Truth thus become fewer and fewer.
The history of Marxism that our book outlines is very different and is based on the following points. (1) Classical Marxism—that of the Second International—grounded its political strategy on the increasing centrality of the working class, this being the result of the simplification of social structure under capitalism. (2) From the beginning this prediction was shown to be false, and within the bosom of the Second International three attempts were made to respond to that situation: the Orthodox Marxists affirmed that the tendencies of capitalism which were at odds with the originary Marxist predictions were transitory, and that the postulated general line of capitalist development would eventually assert itself; the Revisionists argued that, on the contrary, those tendencies were permanent and that Social Democrats should therefore cease to organize as a revolutionary party and become a party of social reforms; finally revolutionary syndicalism, though sharing the reformist interpretation of the evolution of capitalism, attempted to reaffirm the radical perspective on the basis of a revolutionary reconstruction of class around the myth of the general strike. (3) The dislocations proper to uneven and combined development obliged the agents of socialist change—fundamentally the working class—to assume democratic tasks which had not been foreseen in the classical strategy, and it was precisely this taking up of new tasks which was denominated ‘hegemony’. (4) From the Leninist concept of class alliances to the Gramscian concept of ‘intellectual and moral’ leadership, there is an increasing extension of hegemonic tasks, to the extent that for Gramsci social agents are not classes but ‘collective wills’. (5) There is, then, an internal movement of Marxist thought from extreme essentialist forms—those of Plekhanov, for example—to Gramsci’s conception of social practices as hegemonic and articulatory, which virtually places us in the field, explored in contemporary thought, of ‘language games’ and the ‘logic of the signifier’.
As we can see, the axis of our argument is that, at the same time that essentialism disintegrated within the field of classical Marxism, new political logics and arguments started to replace it. If this process could not go further, it was largely due to the political conditions in which it took place: under the empire of Communist parties which regarded themselves as rigid champions of orthodoxy and repressed all intellectual creativity. If today we have to carry out the transition to post-Marxism by having recourse to a series of intellectual currents which are outside the Marxist tradition, it is to a large extent as a result of this process.
An Atemporal Critique
We will reply point by point to Geras’s main criticisms of our analysis of the history of Marxism. First, he suggests that we have designed a very simple game, choosing at random a group of Marxist thinkers and separating the categories they inherited from classical Marxism from those other aspects of their work in which, confronted with a complex social reality, they were forced to move away from economic determinism. We are then alleged to have given medals to those who went furthest in this direction. This is, obviously, a caricature. In the first place, our main focus was not on economic determinism but on essentialism (it is possible to be absolutely ‘superstructuralist’ and nevertheless essentialist). In the second place, we did not consider ‘any Marxist’ at random but narrated an intellectual history: one of progressive disintegration within Marxism of the originary essentialism. Geras says nothing of this history. However, the image he describes fits his own vision well: for him there is no internal history of Marxism; Marxist categories have a validity which is atemporal and it is only a question of complementing them here and there with a bit of empiricism and good sense.
Secondly, we are supposed to have contradicted ourselves by saying that Marxism is monist and dualist at the same time. But there is no contradiction here: what we asserted was that Marxism becomes dualist as a result of the failure of monism. A theory that starts by being pluralist would run no risk of becoming dualist.
Thirdly, Geras alleges that we have presented ourselves as the latest step in the long history of Marxism, and so fallen into the error, criticized by Althusser, of seeing in the past only a pre-announcement of oneself. Here, at least, Geras has posed a relevant intellectual question. Our answer is this: any history that deserves its name and is not a mere chronicle must proceed in the way we have proceeded—in Foucault’s terms, history is always history of the present. If today I have the category ‘income distribution’, for instance, I can inquire about the distribution of income in ancient times or in the Middle Ages, even if that category did not exist then. It is by questioning the past from the perspective of the present that history is constructed. Historical reconstruction is impossible without interrogating the past. This means that there is not an in-itself of history, but rather a multiple refraction of it, depending on the traditions from which it is interrogated. It also means that our interpretations themselves are transitory, since future questions will result in very different images of the past. For this very reason, Althusser’s critique of teleological conceptions of the past is not applicable in our case; we do not assert that we are the culmination of a process that was pre-announced, as in the transition from the ‘in itself’ to the ‘for itself’. Although the present organizes the past, it can have no claim to have disclosed its ‘essence’.
Finally, at several points Geras questions our treatment of texts by Trotsky and Rosa Luxemburg. In the case of Trotsky, we are said to have made use of ‘tendentious quotations’. What we actually said was that: (1) Pokrovsky posed a theoretical question to Trotsky: namely, whether it is compatible with Marxism to attribute to the State such a degree of autonomy from classes as Trotsky does in the case of Russia; and (2) Trotsky, instead of answering theoretically, gave an account of Russian development and attempted to deal with the specific theoretical aspect of Pokrovsky’s question only in terms of the contrast between the greenness of life and the greyness of theory (‘Comrade Pokrovsky’s thought is gripped in a vice of rigid social categories which he puts in place of live historical forces’, etc.).  Thus the type of question that Pokrovsky’s intervention implied—one referring to the degree of autonomy of the superstructure and its compatibility with Marxism—is not tackled by Trotsky at any point. The reader can check all the passages of Trotsky to which Geras refers and in none of them will s/he find a theoretical discussion concerning the relationship between base and superstructure. As for the idea that we demanded from Trotsky a theory of relative autonomy when we had affirmed its impossibility in another part of our book, we have already seen that this last point is a pure invention by Geras.
In the case of Rosa Luxemburg it is a question not of misquotations but of simplifications—that is, we are supposed to have reduced everything to the ‘symbol’. Geras starts by enumerating five points, with which it would be difficult to disagree because they are simply a summary of Rosa Luxemburg’s work on the mass strike. Our level of analysis is different, however, and does not contradict any of the five points in Geras’s summary. The fifth point, for instance, reads: ‘economic and political dimensions of the overall conflict interact, intersect, run together.’  A further nine-point enumeration then explains what this interaction is, and we would not disagree with it either since it merely gives examples of such interaction. What our text asserts—and what Geras apparently denies without presenting the slightest argument—is that through all these examples a specific social logic manifests itself, which is the logic of the symbol. A meaning is symbolic when it is a second meaning, added to the primary one (‘rose’, for example, can symbolize ‘love’). In the Russian Revolution, ‘peace’, ‘bread’ and ‘land’ symbolized a variety of other social demands. For example, a strike for wage demands by any group of workers will, in an extremely repressive political context, also symbolize opposition to the system as a whole and encourage protest movements by very different groups; in this way an increasing relation of overdetermination and equivalence is created among multiple isolated demands. Our argument was that: (1) this is the mechanism described by Rosa Luxemburg in The Mass Strike; (2) it is, for her, the central element in the constitution of the unity between economic struggle and political class struggle; (3) her text is conceived as an intervention in the dispute between syndicalist and party theoreticians about the relative weight of economic and political struggle. Since Geras does not present any argument against these three theses, it makes little sense to prolong this discussion. 
As is usual in sectarian literature, when it comes to talking about politics Geras has remarkably little to say. But we do need to deal with his assertion that it is an axiom that socialism should be democratic.  The fact is that for any person who does not live on Mars, the relation between socialism and democracy is axiomatic only in Geras’s mind. Has Geras ever heard of Stalinism, of the one-party system, of press censorship, of the Chinese Cultural Revolution, of the Polish coup d’état, of the entry of Soviet tanks into Prague and Budapest? And if the answer is that nothing of the kind is true socialism, we have to be clear what game we are playing. There are three possibilities. The first is that Geras is constructing an ideal model of society in the way that the utopian socialists did. Nothing, of course, prevents him from doing so and from declaring that in Gerasland collective ownership of the means of production and democracy go together; but in that case we should not claim to be speaking about the real world. The second possibility is to affirm that the authoritarian States of the Soviet bloc represent a transitory and necessary phase in the passage towards communism. This is the miserable excuse that ‘progressive’ intellectuals gave to support the worst excesses of Stalinism, from the Moscow trials onwards. The third possibility is to assert that these states are ‘degenerate forms’ of socialism. However, the very fact that such ‘degeneration’ is possible clearly indicates that the relation between socialism and democracy is far from being axiomatic.
For us the articulation between socialism and democracy, far from being an axiom, is a political project; that is, it is the result of a long and complex hegemonic construction, which is permanently under threat and thus needs to be continuously redefined. The first problem to be discussed, therefore, is the ‘foundations’ of a progressive politics. For Geras this presents the following difficulty: has not our critique of essentialism eliminated any possible basis for preferring one type of politics to another? Everything depends on what we understand by ‘foundation’. If it is a question of a foundation that enables us to decide with apodictic certainty that one type of society is better than another, the answer is no, there cannot be such a foundation. However, it does not follow that there is no possibility of reasoning politically and of preferring, for a variety of reasons, certain political positions to others. (It is comical that a stern critic of ‘either/or’ solutions such as Geras confronts us with exactly this type of alternative.) Even if we cannot decide algorithmically about many things, this does not mean that we are confined to total nihilism, since we can reason about the verisimilitude of the available alternatives. In that sense, Aristotle distinguishes between phronesis (prudence) and theory (purely speculative knowledge). An argument founded on the apodicticity of the conclusion is an argument which admits neither discussion nor any plurality of viewpoints; on the other hand, an argument which tries to found itself on the verisimilitude of its conclusions, is essentially pluralist, because it needs to make reference to other arguments and, since the process is essentially open, these can always be contested and refuted. The logic of verisimilitude is, in this sense, essentially public and democratic. Thus, the first condition of a radically democratic society is to accept the contingent and radically open character of all its values—and in that sense, to abandon the aspiration to a single foundation.
At this point we can refute a myth, the one which has it that our position is incompatible with humanism. What we have rejected is the idea that humanist values have the metaphysical status of an essence and that they are, therefore, prior to any concrete history and society. However, this is not to deny their validity; it only means that their validity is constructed by means of particular discursive and argumentative practices. The history of the production of ‘Man’ (in the sense of human beings who are bearers of rights in their exclusive human capacity) is a recent history—of the last three hundred years. Before then, all men were equal only in the face of God. This history of the production of ‘Man’ can be followed step by step and it has been one of the great achievements of our culture; to outline this history would be to reconstruct the various discursive surfaces where it has taken place—the juridical, educational, economic and other institutions, in which differences based on status, social class or wealth were progressively eliminated. The ‘human being’, without qualification, is the overdetermined effect of this process of multiple construction. It is within this discursive plurality that ‘humanist values’ are constructed and expanded. And we know well that they are always threatened: racism, sexism, class discrimination, always limit the emergence and full validity of humanism. To deny to the ‘human’ the status of an essence is to draw attention to the historical conditions that have led to its emergence and to make possible, therefore, a wider degree of realism in the fight for the full realization of those values.
The Transformation of Political Consciousness
Now, the ‘humanization’ of increasingly wider areas of social relations is linked to the fundamental process of transformation of political consciousness in Western societies during the last two hundred years, which is what, following Tocqueville, we have called the ‘democratic revolution’. Our central argument is that socialism is an integral part of the ‘democratic revolution’ and has no meaning outside of it (which, as we will see, is very different from saying that socialism is axiomatically democratic). In order to explain our argument we will start from an analysis of the capitalist–worker relation. According to the classical Marxist thesis, the basic antagonism of capitalist society is constituted around the extraction of surplus-value by the capitalist from the worker. But it is important to see where the antagonism resides. A first possibility would be to affirm that the antagonism is inherent in the very form of the wage-labour–capital relation, to the extent that this form is based on the appropriation by capital of the worker’s surplus labour. However, this solution is clearly incorrect: the capitalist–worker relation considered as form—that is to say, insofar as the worker is considered not as flesh and blood but only as the economic category of ‘seller of labour power’—is not an antagonistic one. Only if the worker resists the extraction of his or her surplus-value by the capitalist does the relation become antagonistic, but such resistance cannot be logically deduced from the category ‘seller of labour power’. It is only if we add a further assumption, such as the ‘homo oeconomicus’ of classical political economy, that the relation becomes antagonistic, since it then becomes a zero-sum game between worker and capitalist. However, this idea that the worker is a profit-maximizer in the same way as the capitalist has been correctly rejected by all Marxist theorists.
Thus, there is only one solution left: that the antagonism is not intrinsic to the capitalist relation of production as such, but rather, that it is established between the relation of production and something external to it—for instance, the fact that below a certain level of wages the worker cannot live in a decent way, send his/her children to school, have access to certain forms of recreation, etc. The pattern and the intensity of the antagonism depend, therefore, to a large extent, on the way in which the social agent is constituted outside the relations of production. Now, the further we are from a mere subsistence level, the more the worker’s expectations are bound up with a certain perception of his or her place in the world. This perception depends on the participation of workers in a variety of spheres and on a certain awareness of their rights; and the more democratic-egalitarian discourses have penetrated society, the less will workers accept as natural a limitation of their access to a set of social and cultural goods. Thus, the possibility of deepening the anti-capitalist struggle itself depends on the extension of the democratic revolution. Even more: anti-capitalism is an internal moment of the democratic revolution. 
However, if this is right, if antagonism is not intrinsic to the relation of production as such but is established between the relation of production and something external to it, then two consequences follow. The first is that there are no apriori privileged places in the anti-capitalist struggle. We should remember that for the Second International—for Kautsky, particularly—the idea of the centrality of the working class was linked to: (a) a vision of the collapse of capitalism as determined by the contradiction between forces and relations of production which would lead to increasing social misery—that is to say, to the contradiction between the capitalist system as a whole and the vast masses of the population; and (b) to the idea that capitalism would lead to proletarianization of the middle classes and the peasantry, as a result of which, when the crisis of the system came about, everything would be reduced to a simple showdown between capitalists and workers. However, as the second process has not taken place, there is no reason to assume that the working class has a privileged role in the anti-capitalist struggle. There are many points of antagonism between capitalism and various sections of the population (environmental pollution, property development in certain areas, the arms race, the flow of capital from one region to another, etc.), and this means that we will have a variety of anti-capitalist struggles. The second consequence is that the potential emergence of a radical anti-capitalist politics through the deepening of the democratic revolution, will result from global political decisions taken by vast sectors of the population and will not be linked to a particular position in the social structure. In this sense there are no intrinsically anti-capitalist struggles, although a set of struggles, within certain contexts, could become anti-capitalist.
If everything then depends on the extension and deepening of the democratic revolution, we should ask what the latter itself depends on and what it ultimately consists of. Marx correctly observed that capitalism only expands through permanent transformation of the means of production and the dislocation and progressive dissolution of traditional social relations. Such dislocation effects are manifest, on the one hand, in commodification, and on the other hand, in the set of phenomena linked to uneven and combined development. In these conditions, the radical instability and threat to social identities posed by capitalist expansion necessarily leads to new forms of collective imaginary which reconstruct those threatened identities in a fundamentally new way. Our thesis is that egalitarian discourses and discourses on rights play a fundamental role in the reconstruction of collective identities. At the beginning of this process in the French Revolution, the public space of citizenship was the exclusive domain of equality, while in the private sphere no questioning took place of existing social inequalities. However, as Tocqueville clearly understood, once human beings accept the legitimacy of the principle of equality in one sphere they will attempt to extend it to every other sphere of life. Thus, once the dislocations generated by capitalist expansion became more general, more and more sectors constructed the legitimacy of their claims around the principles of equality and liberty. The development of workers’ and anti-capitalist struggles during the nineteenth century was a crucial moment in this process, but it was not the only or the last one: the struggles of the socalled ‘new social movements’ of the last few decades are a further phase in the deepening of the democratic revolution. Towards the end of the nineteenth century Bernstein clearly understood that future advances in the democratization of the State and of society would depend on autonomous initiatives starting from different points within the social fabric, since rising labour productivity and successful workers’ struggles were having the combined effect that workers ceased to be ‘proletarian’ and became ‘citizens’, that is to say, they came to participate in an increasing variety of aspects of the life of their country. This was the start of the process that we have called the ‘dispersion of subject positions’. Bernstein’s view was, without any doubt, excessively simplistic and optimistic, but his predictions were fundamentally correct. However, it is important to see that from this plurality and dislocation there does not follow an increasing integration and adaptation to the system. The dislocatory effects that were mentioned above continue to influence all these dispersed subject positions, which is to say that the latter become the points which make possible a new radicalization, and with this, the process of the radical democratization of society acquires a new depth and a new impulse. The result of the process of dispersion and fragmentation whose first phases Bernstein described, was not increasingly conformist and integrated societies: it was the great mobilizations of 1968.
There are two more points which require discussion. The first refers to liberalism. If the radical democratization of society emerges from a variety of autonomous struggles which are themselves overdetermined by forms of hegemonic articulation; if, in addition, everything depends on a proliferation of public spaces of argumentation and decision whereby social agents are increasingly capable of self-management; then it is clear that this process does not pass through a direct attack upon the State apparatuses but involves the consolidation and democratic reform of the liberal State. The ensemble of its constitutive principles—division of powers, universal suffrage, multi-party systems, civil rights, etc.—must be defended and consolidated. It is within the framework of these basic principles of the political community that it is possible to advance the full range of present-day democratic demands (from the rights of national, racial and sexual minorities to the anti-capitalist struggle itself).
The second point refers to totalitarianism. Here Geras introduces one of his usual confusions. In trying to present our critique of totalitarianism, he treats this critique as if it presupposed a fundamental identity between communism and fascism. Obviously this is not the case. Fascism and communism, as types of society, are totally different. The only possible comparison concerns he presence in both of a certain type of political logic by which they are societies with a State Truth. Hence, while the radical democratic imaginary presupposes openness and pluralism and processes of argumentation which never lead to an ultimate foundation, totalitarian societies are constituted through their claim to master the foundation. Evidently there is a strong danger of totalitarianism in the twentieth century, and the reasons are clear: insofar as dislocatory effects dominate and the old structures in which power was immanent dissolve, there is an increasing tendency to concentrate power in one point from which the attempt is made ‘rationally’ to reconstruct the ensemble of the social fabric. Radical democracy and totalitarianism are, therefore, entirely opposite in their attempts to deal with the problems deriving from dislocation and uneven development.
To conclude, we would like to indicate the three fundamental points on which we consider it necessary today to go beyond the theoretical and political horizon of Marxism. The first is a philosophical point which relates to the partial character of Marx’s ‘materialism’, to its manifold dependence on crucial aspects of the categories of traditional metaphysics. In this respect, as we have tried to show, discourse theory is not just a simple theoretical or epistemological approach; it implies, by asserting the radical historicity of being and therefore the purely human nature of truth, the commitment to show the world for what it is: an entirely social construction of human beings which is not grounded on any metaphysical ‘necessity’ external to it—neither God, nor ‘essential forms’, nor the ‘necessary laws of history’.
The second aspect refers to the social analyses of Marx. The greatest merit of Marxist theory has been to illuminate fundamental tendencies in the self-development of capitalism and the antagonisms that it generates. However, here again the analysis is incomplete and, in a certain sense, parochial—limited, to a great extent, to the European experience of the nineteenth century. Today we know that the dislocation effects which capitalism generates at the international level are much deeper than the ones foreseen by Marx. This obliges us to radicalize and to transform in a variety of directions Marx’s conception of the social agent and of social antagonisms.
The third and final aspect is political. By locating socialism in the wider field of the democratic revolution, we have indicated that the political transformations which will eventually enable us to transcend capitalist society are founded on the plurality of social agents and of their struggles. Thus the field of social conflict is extended, rather than being concentrated in a ‘privileged agent’ of socialist change. This also means that the extension and radicalization of democratic struggles does not have a final point of arrival in the achievement of a fully liberated society. There will always be antagonisms, struggles, and partial opaqueness of the social; there will always be history. The myth of the transparent and homogeneous society—which implies the end of politics—must be resolutely abandoned.
We believe that, by clearly locating ourselves in a post-Marxist terrain, we not only help to clarify the meaning of contemporary social struggles but also give to Marxism its theoretical dignity, which can only proceed from recognition of its limitations and of its historicality. Only through such recognition will Marx’s work remain present in our tradition and our political culture.
 Slavoj Zizek, ‘La société n’existe pas’, L’Ane, Paris, October–December 1985.
 Andrew Ross, in m/f 11/12, 1986.
 Alastair Davidson, in Thesis Eleven, No. 16, Melbourne, 1987.
 Stanley Aronowitz, ‘Theory and Socialist Strategy’, Social Text, Winter 1986/87.
 Philip Derbyshire, in City Limits, 26 April 1985.
 David Forgacs, ‘Dethroning the Working Class?’, Marxism Today, May 1985.
 Norman Geras, ‘Post-Marxism?’, New Left Review 163, May–June 1987.
 This example, as the reader will realize, is partly inspired by Wittgenstein.
 Hanna Fenichel Pitkin, Wittgenstein and Justice, Berkeley, 1972, p. 84. See also Stanley Cavell, Must We Mean What We Say?, New York 1969 , p. 9.
 Geras, op. cit., p. 66.
 R. Rorty, Consequences of Pragmatism, Minneapolis, 1982, pp. 166–7.
 Geras, p. 69.
 Ch. Taylor, Hegel, Cambridge 1975, p. 109.
 H. Staten, Wittgenstein and Derrida, Oxford 1985, p.6.
 W.T. Stace, The Philosophy of Hegel, New York 1955, pp. 73–4.
 Staten, op. cit., p.7.
 In the same manner as reactionary theoreticians, Geras considers that he can fix the being of things once and for all. Thus, he says that to call an earthquake an expression of the wrath of God is a ‘superstition’, whilst calling it a ‘natural phenomenon’ is to state ‘what it is’. The problem is not, of course, that it does not make perfect sense in our culture to call certain beliefs ‘superstitions’. But, to counterpose ‘superstitions’ to ‘what things are’ implies: (1) that world views can no longer change (that is to say, that our forms of thought concerning the idea of ‘the natural’ cannot be shown in the future to be contradictory, insufficient, and therefore ‘superstitious’); (2) that, in contrast to men and women in the past, we have today a direct and transparent access to things, which is not mediated by any theory. With such reassurances, it is not surprising that Geras regards himself as a functionary of truth. It is said that at some point Mallarmé believed himself to be the individual mind which embodied the Absolute Spirit, and that he felt overwhelmed. Geras makes the same assumption about himself far more naturally. It is perhaps worthwhile remarking that Geras’s naive ‘verificationism’ will today hardly find defenders among philosophers of any intellectual orientation. W.V. Quine, for instance, who is well anchored in the mainstream tradition of Anglo-American analytic philosophy, writes: ‘I do . . . believe in physical objects and not in Homer’s gods, and I consider it a scientific error to believe otherwise. But in point of epistemological footing the physical objects and the gods differ only in degree and not in kind. Both sorts of entities enter our conception only as cultural posits . . . Moreover, the abstract entities which are the substance of mathematics—ultimately classes and classes of classes and so on up—are another posit in the same spirit. Epistemologically these are myths on the same footing with physical objects and gods, neither better nor worse except for differences in the degree to which they expedite our dealings with sense experiences.’ ‘Two Dogmas of Empiricism’, in From a Logical Point of View, New York 1963, pp. 44–45.
 K. Marx, A Contribution to the Critique of Political Economy, London 1971, p. 21.
 Geras reasons in a similar way. Referring to a passage in our text where we write that ‘the main consequence of a break with the discursive/extra-discursive dichotomy, is the abandonment of the thought/reality opposition’, Geras believes that he is making a very smart materialist move by commenting: ‘A world well and truly external to thought obviously has no meaning outside the thought/reality opposition’ (p. 67). What he does not realize is that in saying so he is asserting that thought is not part of reality and thus giving credence to a purely idealist conception of mind. In addition, he considers that to deny the thought/reality dichotomy is to assert that everything is thought, while what our text denies is the dichotomy as such, with precisely the intention of reintegrating thought to reality. (A deconstruction of the traditional concept of ‘mind’ can be found in Richard Rorty, Philosophy and the Mirror of Nature, Princeton, 1979.).
 Geras, op. cit., p. 49.
 E. Laclau, Ch. Mouffe, Hegemony and Socialist Strategy, Verso, London 1985, p. 140.
 Geras, p. 50.
 Ibid., p. 50.
 L. Trotsky, 1905, London 1971, p. 333.
 Geras, p. 60.
 One further point concerning Rosa Luxemburg. Geras sustains (fn., p. 62) that we deny that Rosa Luxemburg had a theory of the mechanical collapse of the capitalist system. This is not so. The point that we make is rather that nobody has pushed the metaphor of the mechanical collapse so far as to take it literally; and that, therefore, all Marxist writers of the period of the Second International combined, in different degrees, objective laws and conscious intervention of the class in their theorizations of the end of capitalism. A second point that we make in the passage in question—and here yes, our interpretation clearly differs from Geras’s—is that it is because the logic of spontaneism was not enough to ground the class nature of the social agents, that Luxemburg had to find a different grounding and was forced to appeal to a hardening of the objective laws of capitalist development. Fully to discuss this issue would obviously require far more space than we have here.
 Geras, p. 79.
 We would like to stress that, in our view, the various anti-capitalist struggles are an integral part of the democratic revolution, but this does not imply that socialism is necessarily democratic. The latter, as a form of economic organization based upon exclusion of private ownership of the means of production, can be the result, for example, of a bureaucratic imposition, as in the countries of Eastern Europe. In this sense, socialism can be entirely external to the democratic revolution. The compatibility of socialism with democracy, far from being an axiom, is therefore the result of a hegemonic struggle for the articulation of both.