Analytical Marxism – an ex-paradigm?
The odyssey of G.A. CohenMarcus Roberts
In 1978 G.A. Cohen published Karl Marxʼs Theory of History: A Defence. That this landmark work set out to defend (something like) the orthodox historical materialism of the Second International was surprising enough; that its author situated himself within the ʻanalyticalʼ tradition – and therefore engaged, and sought to defeat, Acton, Plamenatz, Popper et al. on their own methodological terrain – was surprising indeed. It is testimony to Cohenʼs analytical acuity that, from such unpropitious materials, he fashioned not a mere curio, but arguably the most accomplished defence of ʻtechnological determinismʼ ever produced, and one of the most important works of Marxist philosophy to have emerged from the Anglo-American academy. In fact, its publication heralded the emergence of a sui generis Marxism designated by its progenitors – prominent amongst whom, alongside Cohen, were Jon Elster, John Roemer, Adam Przeworski and Erik Olin Wright – as ʻAnalyticalʼ or ʻRational Choiceʼ Marxism. The architects of this new ʻparadigmʼ insisted that a necessary condition of Marxismʼs salvation was its importation into the tradition of analytical philosophical method, ʻpositivistʼ social science, and – or, at least, so argued Elster, Roemer and Przeworski – that version of rational choice theory originating in the Marginalist revolution of the 1870s and providing neo-classical economics with its deﬁnitive axioms. As one commentator observed, ʻCohen and his co-thinkers … casually crossed the supposedly impassable border between Marxism and the academic mainstream in philosophy and social theory.ʼ 
After nearly two decades, few Marxist ʻinsightsʼ have survived the attempt to ʻreconstructʼ it. Most of the Marxist heritage – Marxism-Leninism, Trotskyism, Western Marxism and Structuralist Marxism – had been consigned to the Humean ﬂames from the outset. So, too, had Marxʼs ʻmultiply confusedʼ anatomy of the capitalist mode of production.  As for ʻMarxʼs theory of historyʼ in its technological determinist incarnation, Cohen has long since confessed to doubts as to its defensibility; few Marxists – even amongst his co-workers – now share the slightest doubts about its indefensibility;3 and, anyway, Cohen himself no longer considers it to have any purchase upon the crucial problems confronting socialists at the close of the twentieth century. He argues that the pre-history of the historical materialist programme has nothing very interesting to tell us regarding either the constituency, agency and strategy of any prospective transition to socialism, or the motivational and institutional structures of a feasible socialism. Thus, in the introduction to Self-Ownership, Freedom and Equality (1995), Cohen concedes that to the extent that Marxism [i.e. Analytical Marxism – MR] is still alive, as … one may say that it (sort of) is in the work of scholars like John Roemer and Philippe Van Parijs [note the conspicuous absence of Elster – MR], it presents itself as a set of values and a set of designs for realising those values … Its shell is cracked and crumbling, its soft underbelly is exposed. 
However, while conceding the demise of Analytical Marxism, Cohen, in an article originally published in 1990, announces the advent of another new paradigm: ʻAnalytical Semi-Marxismʼ. 
If it is true that the moment anyone started to talk to Marx about morality he would laugh, then Analytical Semi-Marxism would have had him in stitches. At the end of a century providing socialists with few occasions for merriment, Cohen declares it high time for a straight-faced engagement with, and development of, the ʻethicalʼ or ʻutopianʼ socialism decried by Marx-ismʼs founders.  Marxists could excuse themselves from serious application to normative political philosophy only so long as they had anticipated both the emergence of a working-class movement impelled into class struggle by its material interests, and the advent of an era of material abundance placing humankind, at long last, beyond the ʻcircumstances of justiceʼ and, therefore, the need for theories of justice. The latest news from the Anglophone academy is that the working class has a good deal more to lose than its chains (and, it appears, a good deal less to win than a world), and that ʻ[w]e can no longer sustain Marxʼs extravagant, pre-Green, materialist optimismʼ.  The ʻplanet earth rebelsʼ against the ﬁnal elimination of material scarcity.  The Marxist explanatory programme in ruins, and – relatedly – socialism relegated from the destiny of humankind to one social blueprint amongst others, ʻMarxists, or what were Marxists, are impelled into normative political philosophyʼ. 
In view of Marxʼs notoriously inconsistent and fragmentary remarks concerning our morals and theirs, it might be anticipated that, if Marxismʼs ʻhard factual carapaceʼ has disintegrated, then its ʻsoft underbellyʼ will prove very soft indeed. So concluded Elster and Roemer, who, bereft of carapaces, set about appropriating and developing left-liberal normative philosophy in defence of versions of market socialism. Against this current, Cohen, resisting the stampede to market socialism, has launched a last-ditch defence of a more resolutely Marxian normative position. Contra Roemer, he argues that, unlike egalitarian liberals, Marxists have traditionally been, and should continue to be, interested in, and unreconciled to, capitalist exploitation.  Unfortunately, in so far as this Cohenite brand of Semi-Marxism distinguishes itself from leftliberalism, it ﬁnds itself in bad company. Speciﬁcally, or so Cohen argues, it is forced to take the rabid neo-liberalism of Robert Nozickʼs Anarchy, State and Utopia extremely seriously.
Taking nozick seriously
Cohen argues that a Semi-Marxism insistent upon the injustice of capitalist exploitation confronts a serious challenge from a neo-liberal ʻentitlementʼ theory of justice which abandons the unemployed, sick, disabled, and old – or at least those so ʻirresponsibleʼ as to have failed to make adequate private provision – to the tender mercies of private charity; refusing to tax the Trumps, Rockefellers, Gettys, Bransons, and of course the less spectacularly endowed, in order to ﬁnance welfare provisions, on the grounds that this is morally equivalent to condemning them to forced labour. Remarkably, Cohen is forced ﬁnally to conclude not only that Marxists should be exercised by Nozick but also that ʻthe Marxist doctrine of exploitation is ﬂagrantly incongruent with even the minimal principle of the welfare stateʼ. 
Those Marxists who argue that capitalist exploitation is unjust claim that, despite formal freedom of contract, workers are forced to sell their labour power to some capitalist; and that, on entering the sphere of production, they spend a portion of the working day producing surplus value appropriated by the capitalist without return. Exploitation is unjust because capitalists ʻstealʼ a portion of the workerʼs product. What general principle underlies this critique? The very principle, or so Cohen claims, to be found at the heart of Nozickian libertarianism: the self-ownership principle. Exploitation is morally objectionable because workers are coerced into exercising their powers for the beneﬁt of non-workers (i.e. capitalists). This is objectionable because it violates self-ownership: ʻI do not (fully) own myself if I am required to give others (part of) what I earn [or produce – MR] by applying my powers.ʼ  The problem is that taxing workers in order to ﬁnance welfare programmes for non-workers violates self-ownership for precisely the same reasons as capitalist exploitation does: it takes from workers and gives to non-workers (i.e. welfare claimants).  Marxists appear to be compelled, on pain of inconsistency, either to drop their egalitarian commitments or to abandon the moralized critique of capitalist exploitation; that is, unless there is some subtle way of reconciling self-ownership with egalitarianism (and, therefore, Marxian anti-capitalism with socialism).
Some obvious moves are either unsatisfactory or unavailable. For example, it is unsatisfactory to argue from the unjust origins of capitalism to its moral indefensibility. Of course, capitalismʼs origins leave something to be desired; the problem is that a consistent, and historically literate, libertarian (admittedly a rare enough beast) must concede the injustice of historical capitalism, in view of its origins in, and development via, multiple violations of those inviolable rights eulogized by the libertarian Right. However, it does not follow from capitalismʼs bloody history that the form of exploitation characteristic of the capitalist mode of production is, in and of itself, unjust. A ʻcleanʼ capitalism – that is, one with unimpeachable origins – is, at least, conceivable. For example, we might imagine everyone beginning at the beginning with an equal share of worldly resources. If some people are hard-working and frugal then they may work themselves into a position to exploit the ʻlazy rascalsʼ who, abandoning themselves to whatever ʻriotous livingʼ is afforded in a state of nature, have allowed their plots to go to seed. Moreover, the ʻexploitedʼ in such circumstances might be better provisioned than they would have been had nobody been in a position to exploit them – that is, if everyone had squandered their initial endowments. The Semi-Marxist is committed to arguing that even such an immaculately conceived and beneﬁcent capitalism as this, because it is exploitative, is unjust. 
A familiar left-liberal response to this kind of defence of inequality – and one to which Cohen is sympathetic15 – is to argue that if it is luck in the natural and social lotteries that has endowed some individuals with virtuous dispositions and scarce talents, then these fortunate individuals do nothing to deserve them, and therefore have no indefeasible moral entitlement to the material – and other – beneﬁts rewarding their exercise. However, this move is unavailable to Semi-Marxists so long as their critique of exploitation is reliant upon a self-ownership principle, according to which we are the rightful owners of our powers, not because we have done anything to deserve them, but simply because we belong to ourselves. For example, Jon and John can agree that it is sheer good fortune that has endowed Jon with, say, musical talents, but agree also that it would be unjust for John to coerce Jon into exercising them. There is a further reason why appeals to desert will not help to salvage the Semi-Marxist critique of exploitation: such arguments challenge the inegalitarian distribution of assets enabling capitalists to exploit workers, while providing no (independent) reasons for condemning exploitation as such. 
The self-ownership principle, then, asserts that what we do with ourselves is our business, so long as we mind our business. It does not ask whether we deserve our dispositions, powers and talents, nor does it require us to exercise them, at however small a cost to ourselves, to meliorate human suffering and deprivation. An obvious conclusion is that, to the extent that a moralized Marxism is unexpectedly premissed in a moral principle with such mean and unappealing implications as this – Cohen concludes that it is ʻthe principle of the bourgeois revolutionʼ  – it should be abandoned without further ado. Eventually, Cohen does abandon it, but there is further ado: he lavishes considerable intellectual ingenuity on Nozick in an effort to reconcile the Marxist critique of capitalist exploitation with socialist egalitarianism.
Libertarianism is attacked at its weakest point (at least, its weakest point once self-ownership is conceded). Self-ownership alone will not generate an unequal distribution of worldly resources. If we go back far enough, we eventually arrive at a point in (pre-)history at which there is no private ownership of the external world. The problem for the inheritors of the classical liberal tradition is to explain how it is that pre-social individuals ever get from owning only themselves to owning also bits of the world. In the Second Treatise of Government, John Locke answered that full liberal property rights are legitimately acquired when some agent mixes her labour with a previously unowned thing, is not wasteful, and leaves ʻenough and as goodʼ for others. Nozick argues that this position – and, in particular, the labour-mixture thesis – is, in various ways, unsatisfactory. In its stead, he proposes the following ʻprovisoʼ: ʻ[a] process normally giving rise to a permanent bequeathable property in a previously unowned thing will not do so if the position of others no longer at liberty to use that thing is thereby worsenedʼ. 
Cohen identiﬁes three problems with this proviso.
First, Nozick has nothing whatsoever to say about that process ʻnormallyʼ giving rise to full liberal property rights; pouring scorn upon the labour mixture thesis he gives no indication of what he thinks might replace it. Second, the decision-procedure for determining whether the position of others is worsened by an act of appropriation only, and arbitrarily, compares the post-appropriation position with the pre-appropriation position. It does not consider whether anybody is worse off than they would have been under some alternative set of property relations:
a defensibly strong Lockean proviso will forbid the formation of full liberal private property … [f]or there will always be some who would have been better off under an alternative dispensation that it would be arbitrary to exclude from consideration. 
Finally, Nozick assumes that worldly resources are originally unowned, that nobody owns anything. But we might equally well assume that, in the pre-social position, everybody owns everything (joint ownership); and, therefore, that pre-social individuals are not entitled to use worldly resources without (unanimous) consent. If we do this, then – or so Cohen argues, with considerable ingenuity and at some length – we have an original position in which self-ownership is respected and equality is maintained.
Unfortunately for the Semi-Marxist, under conditions of joint ownership no one can actually ʻuse their rights of self-ownership to achieve substantial control over their own lives, since anything that they might want to do would be subject to the veto of othersʼ.  There is little point in owning oneʼs powers if one is denied access to those worldly resources required if they are to be exercised. In view of their commitment to ʻgenuine freedom and autonomyʼ, Marxists must reject a joint ownership proposal that leaves pre-social individuals without ʻsubstantial control over their own livesʼ, and with little opportunity to exercise and develop their powers, capacities and talents. However, there are crumbs of comfort to be had here. Nozick and his supporters, argues Cohen, would be ill-advised to press this objection: in capitalist societies, the self-ownership of a propertyless proletariat is formal in the same sense and for the same reasons.[A] victory emerged from what had looked like the jaws of defeat. For I realised that my libertarian antagonists could not press [this] objection… They could not complain that joint ownership of the external world degrades self-ownership, since the pale self-ownership enjoyed by persons in a jointly owned world is at least as robust as that of selfowning propertyless proletarians, who, unlike joint owners, have no rights at all in external resources, and who also, therefore, lack real control over their own lives. Yet libertarians defend, as a realization of self-ownership, the capitalist world in which proletarians proliferate. 
Nozick is defeated, then, but the victory is a pyrrhic one. Semi-Marxists, given their commitments to egalitarianism and to ʻgenuine freedom and autonomyʼ, are obliged to attack the self-ownership thesis; but they do so at the cost of undermining the basis of their moralized critique of capitalist exploitation.
But is it marxism?
Perhaps it is already apparent why many Marxists – and not only Marxists – have been hostile to Analytical Marxism, and are unlikely to ﬁnd Analytical SemiMarxism any more congenial. In particular, Cohen and his cohorts have been attacked – by Marxists and Anti-Marxists alike22 – for their ahistoricism and ʻdonnish remotenessʼ from socialist politics. Regarding the former charge, Cohen unapologetically, and unfashionably, declares his allegiance to moral principles understood to possess transhistorical validity: ʻwhile historical circumstances undoubtedly affect what justice (for example) demands, they do so only because timelessly valid principles of justice have different implications at different timesʼ.  Regarding the charge of ʻdonnish remotenessʼ, Cohen would presumably argue that there is not much in the way of socialist politics to be remote from at present, and further, that this is, in part, due to the intellectual failings of socialist normative philosophy. He concedes that Semi-Marxism demands a very considerable diminution of Marxismʼs ʻscientiﬁcʼ pretensions, and substantial revisions to socialist normative philosophy, but insists that there is no longer any intellectually respectable alternative.
The question arises, however, as to whether or not the ʻSemi-Marxismʼ advocated by Cohen has anything to do with Marxism. The very notion of a Semi-Marxism is not a theoretically innocent one: it presupposes that the normative principles inscribed within Classical Marxism can be separated from its explanatory theses without disﬁguration.  On this view, Marxism has traditionally embraced both a theory of history and a set of ʻultimate normative truthsʼ. Semi-Marxism claims to be a species of Marxism (or, at least, to be descended from it), because it seeks to identify, elaborate and develop a distinctively Marxian normative position. It is only a Semi-Marxism because it seeks to detach these moral commitments from Marxist theories of history and of capitalism. Marxism, then, is not conceived as a theoretical totality, exiled from which its ʻcomponentsʼ cease to be that which they are. The analytical philosopher views it as an amalgam of (partially) distinct methodological, explanatory and normative theses to be prised apart and interrogated independently.  Cohen argues, in support of such a disseverance, that it is perfectly possible to embrace historical materialism ʻbut regret that the career of humanity is as it describes, and, more speciﬁcally that, as it predicts, class society will be superseded by a classless oneʼ.  Commitment to Marxism, then, is consistent with rabid anti-socialism.
However, despite attracting some prominent supporters, this understanding of Classical Marxism as a mere entanglement of explanatory and normative claims, to be disentangled by its self-appointed reconstructors, is a troubling one. Historical materialism is a theory of historical development or progress, and not simply of historical change. Whatever theory of history might remain were it to be excavated of all evaluative presuppositions, it is doubtful that it could be informatively described as an anti-socialist historical materialism, as opposed to some non-Marxist alternative to it. Of course, there is nothing unusual in combining commitment to socialist values with rejection of historical materialism. But such species of socialism are not ordinarily assumed to possess any Marxist content. As Cohen recognizes, commitment to egalitarianism and to ʻgenuine freedom and autonomyʼ are by no means distinctive to Marxism. In order to trace the contours of a Marxian normative position, it is necessary to move from the proclamation of allegedly ʻtimelessʼ moral principles to consider the criteria governing their application in varied historical circumstances. In doing so we ﬁnd that a distinctively Marxian normative position cannot be speciﬁed other than by invoking historical materialism. For example, Marxists would not advocate egalitarianism in pre-capitalist or early capitalist times: the transition to socialism is historically possible only once its material preconditions have been delivered by a succession of inegalitarian and exploitative class societies.
Consideration of Cohenʼs dissatisfaction with his own joint-ownership proposal underscores this point. This proposal, or so he argues, reconciles egalitarianism with self-ownership, but only at an unacceptable price. Marxists will reject it because they value ʻreal freedom and autonomyʼ. What is odd here is Cohenʼs insistence that Marxists will object to joint ownership in a pre-social original position because it frustrates the autonomous development of pre-social individuals. From the perspective of historical materialism, and certainly that version of it defended by Cohen in the 1970s, to permit pre-social or ʻprimitiveʼ individuals to exercise their powers and capacities by transforming worldly things into use-values is to initiate a process of (pre)historical development tending towards the creation of a surplus product and, consequently, the emergence of exploitative class societies. According to historical materialism, ʻgenuine freedom and autonomyʼ is possible only at the close, and not prior to the outset, of a long process of economic, social and political development.
Moreover, regardless of the problem of reconciling Cohenʼs discussion of joint ownership with an allegiance to historical materialism, there are further problems with its proposal. What are we to make of the idea of joint world ownership? To see how strange this proposal is, consider an individual in the pre-social position wishing to grow turnips on an acre of land in (what we now know as) the Norfolk Broads. Is she required – as far as possible – to seek the consent of everyone in the world? Perhaps it is a little less fanciful to imagine joint owners agreeing to a set of general rules governing the appropriation of worldly resources – for example, they might agree to rules permitting the emergence of capitalism. But, if these joint owners are permitted to choose between, say, continued joint ownership and capitalism, then what is to be said in favour of an original position within which opting for capitalism requires unanimity, while the continuance of joint ownership is secured by a single dissenting vote? Is it not ʻarbitrary to exclude from considerationʼ the continuance of joint ownership itself? After all, this might be the worst possible option for the majority (or even all) of these pre-social individuals. Or, to put the same point another way, even if there is no good reason for not starting with joint ownership (Cohen) as opposed to ʻnon-ownershipʼ (Nozick), is there any good reason for starting with it, and not with any other aboriginal property system?
More importantly, leaving aside the organizational and communications problems that would confront anybody attempting to assemble the worldʼs population to discuss such pressing matters within a state of nature, none of this makes sense: there can be no pre-social ʻvetosʼ, ʻrulesʼ or ʻproperty rightsʼ. None of these concepts has any application outside of some pre-existing set of economic, social and juridical relations. Short of resorting to the profoundly anti-Marxist notion of a natural right to property, there is no way that there can be joint ownership, or any other form of ownership, prior to the emergence of all social and juridical institutions. Now, of course, Cohen may have in mind not joint ownership of the entire world (and everything in it – except, of course, each other), but communal ownership of particular worldly resources (for example, tribal ownership of a particular tract of land). This, of course, is a more historically realistic, and formally unobjectionable, starting point for enquiry into the development of property regimes. But it places us at the beginning of an altogether different sort of story to that narrated by Cohen. To see this, notice that, given the project on which he has embarked, Cohen is prohibited from beginning here: questions will arise with respect to communal ownership of particular worldly resources similar to those he is preoccupied with in his discussion of individual acts of appropriation.
The absence of capitalism
Cohenʼs discussion of joint ownership evinces the sort of ahistoricism that has been attacked by many of Analytical Marxismʼs most incisive critics.  As Cohen recognizes, Marxists are generally hostile to these sorts of ʻthought experimentsʼ – or ʻout-of-theway examplesʼ as he calls them. While he is right to respond that such argumentative devices can serve the useful purpose of clarifying and dramatizing problems within normative philosophy, there are limitations to how far the imagination can be stretched. In particular, the designer of such thought experiments should not specify counterfactual conditions that are not simply ʻout-of-the-wayʼ, but actually inconceivable. For example, we cannot be asked to imagine joint ownership in a pre-social state of nature, because this is unimaginable. Similar problems arise with the ʻout-of-the-way examplesʼ which ﬁgure prominently in Cohenʼs discussion of self-ownership and exploitation.
In particular, we are invited to imagine, or so it appears, consenting adults indulging in capitalist acts in the absence of capitalism. Consider the parable of Inﬁrm and Able, which – in various guises – ﬁgures prominently throughout Self-Ownership, Freedom and Equality:Think of a worker who very much enjoys both his work and the wages it brings him and who works for a wholly inﬁrm neighbour who leads a miserable life but who, unlike the worker, has managed to possess himself of a stock of capital.
This inﬁrm capitalist lops off just enough of the workerʼs product so that he, the capitalist, can stay alive. We can suppose that, if something like the stated capital imbalance did not obtain, then the worker would produce for himself alone and callously let his inﬁrm neighbour die. And we can also suppose that it was because he knew that he would die without the power over the worker that capital would give him that the inﬁrm man decided to acquire and exercise that power. 
Cohen wishes to make three points about this ʻout-ofthe-way exampleʼ: (i) Inﬁrm exploits Able; (ii) only the self-ownership principle can provide credible grounds for objecting to Inﬁrmʼs ʻtheftʼ of a portion of Ableʼs product; and (iii) while their moralized critique of exploitation appears to commit Marxists to condemning Inﬁrmʼs treatment of Able, they will feel uncomfortable about this in view of their egalitarian – and other – moral commitments.
So what is to be said about this ʻout-of-the-way exampleʼ? Various questions arise. Why make Inﬁrm and Able neighbours? What is the relevance of Inﬁrmʼs ʻmiserable lifeʼ to the justice or otherwise of his exploitation of Able? Or are we to suppose that Ableʼs wages are generous? And, if so, are we to imagine that it is Inﬁrmʼs generous disposition that determines the wage rate, and not conditions in the labour market? It is this last question that directs us to what might be designated the ʻabsence of capitalismʼ problem. Cohen writes that ʻ[t]his inﬁrm capitalist lops off just enough of the workerʼs product so that he, the capitalist, can stay aliveʼ. It seems that we are supposed to imagine Inﬁrm ʻlop[ping] off just enough of the workerʼs productʼ to secure his (physical) survival as a human being. But this is unimaginable; or, at least, it needs to be added that Inﬁrm must reproduce himself qua capitalist in order to survive qua human being.
Reliant upon successful competition with other capitals for his survival, Inﬁrm may be forced to expand production, to take on other workers, to bear down upon their wage rates, to increase the length of the working day, to support high unemployment as a means to secure low wages and long hours, to empty Ableʼs work of any meaningful content, to sack her during a crisis, and so on. All this not because acquisition of the predication ʻcapitalistʼ transmogriﬁes Inﬁrm from good neighbour to demon incarnate, but because, in order to survive, he must now compete successfully in the market place. Notice also that, if Inﬁrm is to exploit Able, it will not be enough that he should possess himself of a stock of capital; he will also have to deprive Able – who, it seems, starts out as an independent producer – of access to the means of her reproduction – that is, to ensure that she will now die without Inﬁrmʼs ʻassistanceʼ. Moreover, inﬁrm will not be capable of unilaterally separating Able from her means of (re)production, and it is unlikely that anyone else would oblige him by throwing Able off her land. In order to render this extraordinary transition in the affairs of Able and Inﬁrm at all historically credible, we must reconceive it as one manifestation of a transition between modes of production (from, say, independent peasant production to early capitalism). It scarcely needs to be added here that no individual proto-capitalist (such as Inﬁrm) could, prior to such a transition, pick out some particularly callous and mighty unneighbourly proto-worker (Able) for exploitation following it.
So where does all this lead us? Away from the beliefs, dispositions, motives, actions and predicaments of particular individuals condemned to live within the constraints imposed upon them by the capitalist mode of production, and towards analysis and critique of capitalism itself: to consider the actions required of individuals qua ʻpersoniﬁcations of economic categories, the bearers of particular class relations and interestsʼ. 
Thus understood, this parable should not unduly trouble those still committed to Marxʼs critique of capitalist exploitation. The relationship between Inﬁrm (qua capitalist) and Able (qua worker) is simply one more manifestation of the exploitation of labour by capital – structurally (and normatively) indistinguishable from any other such manifestation. Its power to disconcert, such as it is, rests upon a feature of the situation which is beside the central Marxist point: that Inﬁrm is forced to acquire capital for the very same reason that Marxists have traditionally argued that proletarians are forced to sell their labour power – that is, to survive. But this should not disconcert us unduly: the moral justiﬁablity of Inﬁrmʼs actions, given the constraint he faces, is an altogether different matter to that of the justice or otherwise of the relationship he is compelled to enter into vis-à-vis Able (and vis-à-vis labour in general). Nor does the fact that Inﬁrm appears to be an awfully nice person, and Able a rather callous one, make the slightest difference: the injustice inscribed in the relation between capital and labour is not of Inﬁrmʼs doing – on becoming a capitalist he has to do what he has to do. As for Able, callous people can suffer injustice too.
Cohenʼs tendency to conceive capitalist exploitation exclusively in terms of bilateral relationships between individual capitalists and workers neglects the extent to which, for Marx, exploitation was a relationship between social classes – that is, between capital and labour. Cohenʼs claim that the traditional Marxist critique of capitalism invokes a strong version of the self-ownership principle appeals exclusively to those passages in which Marx writes of the ʻrobberyʼ of particular workers by those capitalists employing them at some particular time. On this reading, those outside the sphere of production – the unemployed, the incapacitated, the families and dependents of workers, and so on – are not (currently) members of the exploited class at all. They appear in Self-Ownership, Freedom and Equality in the guise of the ʻpoorʼ and ʻneedyʼ (who are ever with us), and, of course, as non-exploited non-workers pumping welfare beneﬁts out of the direct producers.
Marxʼs position is a rather different one. In Capital, and elsewhere, he is concerned to analyse an exploitative relationship between capital and labour irreducible to a sum of bilateral relationships between capitalists and labourers.  If Marx does offer a moralized critique of capitalism, then many of its central claims pose a serious challenge to Cohenʼs reading of Marx; in particular, to Cohenʼs claim that it covertly relies upon invocation of the self-ownership principle. For example, Marx argues at length that labour, in producing surplus value, is condemned to (re)produce and augment ʻthe alien powerʼ that dominates and exploits it: that is, capital. Labourʼs own activity also turns against it by producing an Industrial Reserve Army: ʻThe working population … produces both the accumulation of capital and the means by which it is itself made relatively superﬂuous; and it does this to an extent which is always increasing.ʼ 
It seems reasonable to conclude from this that the revelation that capital is nothing more or other than past labour – that is, that workers are subjected to the dominion of their own product – reveals a particularly viscous, and unjust, aspect of capitalist exploitation. However, this analysis of the innermost secrets of capitalist exploitation would have no moral signiﬁcance if the commitment undergirding the critique of exploitation was, as Cohen claims, to self-ownership. It is not this workerʼs product that confronts her as an alien power, but labourʼs total product that so confronts her. On Cohenʼs reading of Marx, however, there could be nothing unjust about a situation in which each individual worker confronts the product of all other workers in the form of capital. Their product is their product; it can stand in no morally signiﬁcant relationship to her. For similar reasons, on Cohenʼs reading, no special kind of injustice could arise because labourʼs work activity tends to produce labourʼs future inactivity in the form of unemployment. This might be unjust only if it could be demonstrated that some particular individualʼs own labour activity was instrumental in producing her unemployment.
Consider, in this context, Cohenʼs illustration of his own general thesis regarding the incompatibility of self-ownership and egalitarianism, with speciﬁc reference to Clause IV of the British Labour Partyʼs 1918 constitution (RIP).  This, it may be recalled, had promised ʻto secure for workers by hand or by brain the full fruits of their labourʼ. To it Cohen responds that, ʻif that promise were fulﬁlled, then full-time single parents, inﬁrm people and other unemployables [sic] would get nothingʼ. What he fails to note is that self-ownership is not only incompatible with welfarism, but also with common ownership. If Marxism does invoke a strong version of the self-ownership thesis, it is obliged to render Clause IV as follows: ʻto secure for each worker by hand or by brain the full fruits of his or her labourʼ. No generation of workers could claim that ʻlabourʼ (i.e. those currently involved in the production process) possessed any entitlement to the ʻfull fruits of labourʼ (i.e. the means of production accumulated over generations). Justice will require common ownership only if the injustice to be rectiﬁed is the ʻrobberyʼ by capital of labour, and not the violation by capitalists of the self-ownership of each labourer.
Reasons to be cheerful?
If Cohen has resisted unconditional surrender to leftliberalism à la Roemer, Van Parijs and Elster, then he remains also less than enthusiastic about market socialism: ʻmarket socialism is at best second best, even if it is the best (or more than the best) at which it is now reasonable to aimʼ.  To what, then, does Cohen consider it to be ʻsecond bestʼ? Presumably to something a little closer to the ʻhigher phaseʼ of communism envisaged by Marx. However, for Marx, such a society was possible only on the basis of that material abundance against which the planet earth rebels. Indeed, Cohen argues that it has been the anticipation of abundance (the Marxist ʻtechnological ﬁxʼ) which has provided a convenient – if entirely unsatisfactory – resolution to the problem of reconciling self-ownership and egalitarianism. The defeat of material scarcity leads to the emergence of a wholly non-coercive egalitarian society:
An overﬂowing abundance renders it unnecessary to press the talent of the naturally better endowed into the service of the poorly endowed for the sake of establishing equality of condition, and it is therefore unnecessary to trench against or modify self-ownership, in order to achieve equality. 
If we can no longer share Marxʼs ʻextravagant, pre-Green, materialist optimismʼ, must we abandon all hope of such ʻvoluntary equalityʼ? Not according to Cohen:
it is not Utopian to speak of a level of material plenty which, while too low to abolish conﬂicts of interest as such, is high enough to allow their resolution without coercion and in favour of equality … [people] … are or might become sufﬁciently just to support an egalitarian distribution … in conditions of modest abundance. 
Here we ﬁnd one of the few remaining remnants of historical materialism to have survived the wreckage of the Analytical Marxist programme – and, in particular, of Cohenʼs defence of a reconstructed ʻtechnological determinismʼ: socialism is materially possible as human industry develops to, and beyond, some ʻhighʼ level (that is, as it delivers ʻconditions of modest abundanceʼ). This claim survives in Cohenʼs version of Analytical Semi-Marxism exiled from its theoretical context: it has survived the virtual abandonment of Marxʼs theory of history, of capitalism, and of class struggle. Can the connection between the development of human industry and the possibility of a successful transition to socialism be maintained when so much else is lost?To begin with, it is not clear precisely what Cohen has in mind here. Is he claiming that individuals will be more amenable to ʻvoluntary equalityʼ as they become more prosperous? Surely not. Prosperous individuals are not renowned for their egalitarian sentiments. Is the idea, then, that such sentiments will become more pronounced as GNP – or total global product? – increases to some ʻhighʼ level? If so, it is difﬁcult to imagine what causal mechanism might take us there. Perhaps Cohenʼs claim is not a predictive but a prescriptive one? Under conditions of ʻmodest abundanceʼ people ought to be moved by egalitarian considerations. But, if egalitarian principles of justice possess timeless and universal validity, then the advent of ʻmodest abundanceʼ does not obviously strengthen the arguments in support of – and therefore the reasons for acting in accordance with – egalitarian principles.
In fact Cohen does not envisage any world-historical transformation of human motivations. Human beings have always possessed both self-interested and ʻgenerousʼ dispositions. As the common wealth increases it is no longer necessary to be ʻzealously just and altruisticʼ – to make any ʻheroicʼ sacriﬁce of self-regarding interestsʼ – in order to be moved by egalitarian considerations.  However, the question arises again of whether this point is predictive or prescriptive. If predictive, then ʻmodest abundanceʼ has had no detectable impact of this kind in the advanced capitalist countries – the per-capita incomes of which are, Cohen assures us, ʻalready beyond the required levelʼ.  If evaluative, then it is unclear why we should be any more forgiving towards a feudal lord unmoved by ʻegalitarianʼ considerations, than towards a corporate manager or an afﬂuent worker.
The weaknesses in Cohenʼs arguments – here and elsewhere – are often signalled by an uncharacteristic absence of conceptual clariﬁcation: where he is least sure of his ground, his self-confessed ʻpedantryʼ is wont to desert him. So it is with the troubling notion of a ʻmodest abundanceʼ. Consider how one might go about answering the question, ʻIs there more or less material scarcity now than there was in 1896?ʼ It is a mistake to identify the progressive conquest of scarcity with ever increasing levels of productivity. Scarcity is a function not only of productivity but also of human need: that is, as Marx insisted, as productivity increases human beings acquire historically determined needs. The modern worker has needs that would not even have occurred to the medieval peasant.
Two points follow from this. First, to decide whether material scarcity is yielding to (modest) abundance we require not only a knowledge of productivity levels but also a philosophical anthropology. It will be necessary to distinguish mere desires and preferences (ʻfalse needsʼ) from genuine human needs (ʻtrue needsʼ) – and to do so while recognizing that historical agents acquire (genuine) historically determined needs which are not shared by all human beings regardless of historical and cultural location. Second, and following on from this, we cannot straightforwardly conclude that, because the planet Earth rebels against the indeﬁnite expansion of material productivity, it is therefore in rebellion against the conquest of material scarcity. The concepts of ʻscarcityʼ and ʻabundanceʼ are not purely descriptive ones; they possess an irreducibly evaluative and contestable aspect – that is, unless we regard naked consumer preference as sacrosanct, the way we apply these concepts will depend upon a theory of human need.
Cohenʼs own position clearly does invoke such a philosophical anthropology. In claiming that per-capita incomes within the advanced capitalist countries are indicative of a condition of ʻmodest abundanceʼ he is, presumably, claiming that their productive powers are sufﬁcient, if redeployed within a socialist society, to satisfy not mere – and often manipulatively created – consumer preferences, but genuine human needs. This recalls Cohenʼs efforts, in Karl Marxʼs Theory of History: A Defence, to identify a distinctive contradiction in advanced capitalism. He had argued that – because of capitalismʼs ʻoutput expanding biasʼ – any increase in its productive powers would be (mis)used to produce (ever more frivolous) consumer goods, rather than to alleviate toil: ʻelectric carving knives are ﬁne but nothing beats freedomʼ.  Socialismʼs mission, then, is to bring about a ʻbenign realignment of labour, leisure and educationʼ.  Some conception of human nature and the conditions of its ﬂourishing obviously premiss these arguments. The problem is that he has tried to argue that the development of human industry to a level at which the transition to a feasible socialism is, at long last, on the historical agenda will be fortuitously coincident with a more or less spontaneous disillusionment with consumerism or an outbreak of ʻvoluntaryʼ egalitarianism amongst capitalismʼs ʻtalentedʼ. Neither of these claims is particularly plausible, so why is Cohen so concerned to defend them?
The answer, perhaps, is that he is nostalgic for a Marxism in which he no longer believes, according to which the evolution of capitalism tends to produce and radicalize the agents of its own dissolution. Unfortunately, the development of Analytical Marxism has rendered capitalism inert: the stimulus for social transformation, such as it is, is no longer to be sought within history, but outside of it – in the realm of ʻtimelessʼ principles. Having concluded that the prospects for working-class radicalization in the advanced capitalist countries are between dim and nil, and having abandoned the special theory of capitalism wholesale, Cohen is, it seems, here making a last-ditch attempt to forge some kind of tenuous connection between the development of human industry within capitalism, and the emergence of the agents of – or, at least, a constituency more amenable to – the transition to some form of socialism. In place of class struggle, socialist must hope that (post-) industrial capitalism will so overprovision its ʻtalentedʼ that they become increasingly disposed to indulge their ʻgenerousʼ dispositions and to embrace a ʻvoluntary egalitarianismʼ. It scarcely needs to be added that even if, under conditions of ʻmodest abundanceʼ, ʻgenerousʼ dispositions (or the tendency to act upon them) do become ever more pronounced, there is no reason to assume, and very good reasons for doubting, that generous people will be any more likely to embrace socialist (or any other species of) egalitarianism. The question would remain: why socialism rather than, say, some form of liberalism?
A staple diet
Cohen does not claim to possess a well-worked-out model of a feasible socialism. However, in view of his rejection of large-scale economic planning – ʻVon Mises and Hayek were rightʼ  – it is unclear what alternative he might offer to some type of market socialism other than something along the lines of Nordic social democracy. Of course, a further possibility, to which Cohen confesses his vulnerability, is despair:
Vanity of Vanities, or rather, the form of it that has tempted me, says: genuine socialism is impossible, or virtually impossible to achieve. It is overwhelmingly likely that the best we shall ever get is some kind of capitalism, and it is for others to ﬁnd the strength to ﬁght for a better capitalism. 
If Analytical Semi-Marxism represents Marxismʼs best effort to come to terms with the inhospitable economic, social, political and ideological conditions of the late twentieth century, then despair is indeed indicated. For what is Marxism without either historical materialism (aside, that is, from a few isolated and disﬁgured remnants) or the special theory of capitalism (the labour theory of value, the law of the tendency of the rate of proﬁt to fall, and the corresponding theories of crisis, class formation and class struggle)?
Alan Ryan, writing in 1987, expressed a bewilderment, shared by many others, regarding Analytical Marxistsʼ claim to be ʻreconstructingʼ Marxism. Of Elster, given his hostility to the Marxist tradition, he asked, why ʻbother with Marx at allʼ? As for Roemerʼs ethical egalitarianism, it is ʻnot Marx or Marxistʼ; while in Cohenʼs work he detected a ﬁght to the death between ʻan essentially sentimental and old-fashioned communismʼ and (emerging victorious) a powerful, analytical, intelligence. In conclusion, Ryan anticipated the impending recruitment of the Analytical Marxist to ʻthe endless and repetitive debates over Rawls, Nozick and Dworkin which constitute the staple diet of academic political theoryʼ:
the project of showing that Marx can be made to talk in ways that Oxford philosophers of thirty years ago would have thought sufﬁciently sanitary seems a mistake. It satisﬁes neither the desire to get inside the mind of the historical Marx nor the need for an intellectual apparatus adequate to the complexities and absurdities of the late 20th century. 
The recent works of Cohenʼs erstwhile co-workers strikingly conﬁrm Ryanʼs prognosis. Roemerʼs defence of market socialism appeals directly to left-liberal moral principles; Elster – who, along with Przeworski, has departed the ranks of the ʻNon-Bullshit Marxism Groupʼ – openly declares his allegiance to the basic principles informing ʻthe writings of John Rawls and Ronald Dworkinʼ;43 and Van Parijs is ʻcommitted … to some broadly “Rawlsian” conception of justiceʼ.  As for Cohen, all the indications are that he can no longer postpone the embrace of a socialism of avowedly leftliberal, rather than Marxist, provenance.
This conclusion is unlikely to perturb the (ex) Analytical Marxists who have been impatient with ʻsterileʼ debates ﬁxated upon the question of whether they are ʻreallyʼ (Semi-)Marxists. Perhaps such debates are sterile. However, the complaint against them is not only that their work is ʻnot Marx or Marxistʼ, but, more strongly, that it is entirely alien to the spirit – as well as the letter – of ʻhistoricalʼ Marxism, and wholly inadequate to ʻthe complexities and absurdities of the late 20th centuryʼ. Both these failures are apparent in its total neglect of the Marxist critique of liberalism, and of Marxʼs reasons for disdaining ʻutopianʼ socialism.
In particular, if Marx laughed at the mention of morality, this, perhaps, was not symptomatic of monstrous self-delusion and wanton inconsistency, but rather an expression of his conviction that normative principles were all but impotent in the struggle to change the world – the comedy that pitches hopeless inadequacy against terrifying immensity (with tragic, as well as hilarious, consequences). About this, there are still good reasons to share Marxʼs scepticism and despair. A socialism appealing not to the material interests of any social class, divorced from any identiﬁable mass movement, but rather to the reason, and reasonableness, of disembodied and rational individuals – and which, in place of a systemic analysis of the developmental tendencies of capitalism, offers a selection of blueprints for an ideal society (or, in Cohenʼs case, omits to provide such blueprints) – is unlikely to prevail in the face of a global neo-liberal capitalism possessing formidable economic as well as coercive and ideological resources.
In a piece originally published in Radical Philosophy in the summer of 1986, Joseph McCarney – responding to Gregor McLennanʼs favourable review of Elsterʼs Making Sense of Marx – enquired as to whether Analytical Marxism constituted ʻA New Marxist Paradigm?ʼ  At the outset, he conceded to McLennan that Elsterʼs work might well be ʻthe ﬂagship for a new armada of Marxist analytical scholarship and reassessmentʼ, adding that the other ʻcaptains in [this] armadaʼ included Cohen and Roemer. However, as McCarney proceeded to argue, if Elsterʼs ship was the standard-bearer for a new ʻMarxismʼ, then it was a ship that sought to ease its passage into the paciﬁc waters of the philosophical and social scientiﬁc mainstream by throwing a ʻdufferʼ like Marx – ʻa minor post-Ricardian whose heart [was] in the right placeʼ – overboard at the outset of its all too brief voyage from residual allegiance to a few ʻisolated insightsʼ to be trawled from the work of Marx, to unapologetic embrace of left(ish) liberalism.
Such taunts cannot be levelled at Cohen with equal justiﬁcation. He has persisted, despite the general trajectory of the analytical reconstruction of Marxism, in holding Marx in some esteem, and there is no question of the profundity of the ties that continue to bind him to the socialist project.  However, Cohenʼs best efforts to keep his own ship aﬂoat have led him progressively to jettison its original Marxist cargo. If anything remains of the Marxism he had sought to defend some twenty-ﬁve years ago, then it is little more than a slightly nuanced left-liberal moral position, alongside a commitment to the idea that the development of capitalism has some tendency to produce more favourable conditions for the transition to socialism. Regardless of whether or not Analytical Marxism constituted ʻa new Marxist paradigmʼ some ten years ago, it is now, without doubt, an ex-Marxist one. Cohenʼs recent work has made its contribution to those ʻendless … debates over Rawls, Nozick and Dworkinʼ; Karl Marxʼs Theory of History: A Defence will remain a classic Marxist text, despite the failure of its central argument. However, there is warrant for concluding that, when the intellectual history of this period comes ﬁnally to be written, Cohenʼs efforts over the last quarter of a century to reconstruct a version of Classical Marxism will be viewed as something of an anachronism: the tenacious pursuit of his own peculiar brand of ʻold-fashionedʼ Marxism all the way to its dead end.
1. ^ A. Callinicos, ʻIntroductionʼ to Marxist Theory, Oxford University Press, Oxford, 1989, p. 3.
2. ^ J. Elster, Making Sense of Marx, Cambridge University Press, Cambridge, 1985, p. 554.
3. ^ See, in particular, R. Brenner, ʻThe Social Basis of Economic Developmentʼ, in J. Roemer, ed., Analytical Marxism, Cambridge University Press, Cambridge, 1986; and E.O. Wright, A. Levine and E. Sober, Reconstructing Marxism, Verso, London, 1992.
4. ^ G.A. Cohen, Self-Ownership, Freedom and Equality, Cambridge University Press, Cambridge, 1995, p. 6.
5. ^ See G.A. Cohen, ʻMarxism and Contemporary Political Philosophy, or: Why Nozick Exercises Some Marxists More than He Does Any Egalitarian Liberalsʼ, in Cohen, Self-Ownership, Freedom and Equality (originally published in the Canadian Journal of Philosophy, supp. vol. 16, 1990).
6. ^ See the introduction to Self-Ownership, Freedom and Equality, esp. pp 1–7.
7. ^ Ibid., p. 10.
8. ^ Ibid., p. 7.
9. ^ Ibid., p. 8.
10. ^ Ibid., pp. 204–8.
11. ^ Ibid., pp. 159–60.
12. ^ Ibid., p. 216.
13. ^ Ibid., p. 151.
14. ^ See ibid., pp. 121ff.
15. ^ Although he doubts that such arguments will be conclusive against a ʻmoderately sophisticatedʼ defence of self-ownership (ibid., pp. 229ff).
16. ^ To see this, notice that someone reliant exclusively upon this sort of argument will have no grounds for objection to a possible world within which exploitative relations did emerge as a consequence of, initially equal, historical agents exercising dispositions and talents which they did deserve.
17. ^ Ibid., p. 229.
18. ^ Quoted in ibid., p. 75. See pp. 74–91 for Cohenʼs devastating assault upon the Nozickian proviso.
19. ^ Ibid., p. 87.
20. ^ Ibid., p. 14.
21. ^ Ibid.
22. ^ See, in particular, John Gray, ʻThe Academic Romance of Marxismʼ, and ʻAgainst Cohen on Proletarian Unfreedomʼ, in his Post-Liberalism, Routledge, New York and London, 1993; and for a partial and characteristically uncompromising response, Cohen, Self-Ownership, Freedom and Equality, pp. 62–5.
23. ^ Ibid., p. 2.
24. ^ It would be misleading to trace Cohenʼs intellectual biography from historical materialism to ethical socialism. He has long been concerned with normative philosophy. However, his belief that there are ʻtimelessʼ ethical principles did not, in and of itself, raise any doubts in his mind regarding his historical materialist afﬁliations.
25. ^ For a general argument concerned with the tendency of analytical philosophy to ʻanalyseʼ complex totalities into their (allegedly) separate components, and the problems this poses for Cohenʼs anti-dialectical interpretation of Marxʼs theory of history, see S. Sayers, ʻMarxism and Dialectic Method: A Critique of G.A.
Cohenʼ, in S. Sayers and P. Osborne, eds, Socialism, Feminism and Philosophy, Routledge & Kegan Paul,
26. ^ Self-Ownership, Freedom and Equality, p. 1.
27. ^ See, in particular, E. Meiksins Wood, ʻRational Choice Marxism: Is the Game Worth the Candle?ʼ, New Left Review 177, 1989; M. Lebowitz, ʻIs “Analytical Marxism” Marxism?ʼ, Science and Society 52, 1988; and A.
Callinicos, ʻIntroductionʼ in A. Callinicos, ed., Marxist Theory, Oxford University Press, Oxford, 1989. See also M. Roberts, Analytical Marxism – A Critique, Verso, London and New York, 1996.
28. ^ Self-Ownership, Freedom and Equality, pp. 149–50.
29. ^ K. Marx, ʻPreface to the First Editionʼ, Capital 1, Penguin, Harmondsworth, 1976, p. 92.
30. ^ See, for example, Capital 1, chs 24–25.
31. ^ Ibid., p. 783.
32. ^ Self-Ownership, Freedom and Equality, p. 155 n. 18.
33. ^ Ibid., p. 256.
34. ^ Ibid., p. 122.
35. ^ Ibid., p. 128.
36. ^ Ibid., p. 135.
38. ^ G.A. Cohen, Karl Marxʼs Theory of History: A Defence, Oxford University Press, Oxford, 1978, p. 318.
39. ^ G.A. Cohen, History, Labour, Freedom: Themes from Marx, Oxford University Press, Oxford, 1989, p. 121.
40. ^ Ibid., p. 260.
41. ^ Ibid., p. 254.
42. ^ A. Ryan, ʻCan Marxism be Rescued?ʼ, London Review of Books, 17 September 1987, pp. 9–10.
43. ^ J. Elster, Solomonic Judgements, Cambridge University Press, Cambridge, 1989, p. 205.
44. ^ P. Van Parijs, ʻRawlsians, Christians and Patriotsʼ, European Journal of Philosophy, vol. 1, no. 3, 1993, p. 309.
45. ^ J. McCarney, ʻA New Marxist Paradigm?ʼ, Radical Philosophy 43, Summer 1986.
46. ^ See, in particular ʻThe Future of a Disillusionʼ, in SelfOwnership, Freedom and Equality.