Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, POLITICS ( (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 25 March 2017

Analytical Marxism

Summary and Keywords

Marxists believe that an understanding of human society presupposes an understanding of the nature of the production of its material surplus and the nature of control over that surplus. This belief forms part of the “hard core” of the Marxist scientific research program. This hard core is complemented by a set of auxiliary hypotheses and heuristics, constituting what Imre Lakatos has called a scientific research program’s “protective belt.” The protective belt is a set of hypotheses protecting a research program’s hard core. Over the past century and a half, Marxists have populated the protective belt with an economic theory, a theory of history, a theory of exploitation, and a philosophical anthropology, among other things. Analytical Marxism is located in Marxism’s protective belt. It can be seen as a painstaking exercise in intellectual housekeeping. The exercise consists in replacing the tradition’s antiquated, superfluous, or degenerate furnishings with concepts, methods, and auxiliary hypotheses from analytic philosophy and up-to-date social science.

The three most influential strands in analytical Marxism are, roughly: its reconstruction of Marx’s theory of history, historical materialism; its philosophical anthropology, including the theory of freedom; and its theory of exploitation, including the theory of class.

Keywords: freedom, egalitarianism, exploitation, domination, critical theory, labor theory of value, Karl Marx


Marxists believe that an understanding of human society presupposes an understanding of the nature of the production of its material surplus and the nature of control over that surplus. This belief forms part of the “hard core” of the Marxist scientific research program. This hard core is complemented by a set of auxiliary hypotheses and heuristics, constituting what Lakatos (1978) has called a scientific research program’s “protective belt.” A research program is a set of theories within a scientific domain. Ptolemaic astronomy, neoclassical economics, and Weberian functionalism are all scientific research programs. A protective belt is a set of hypotheses that protects the hard core of the research program from empirical refutation. In Ptolemaic astronomy, for example, claims about geocentrism belong to the hard core and claims about the epicyclical motion of the planets belong to the protective belt: the latter protect the former from empirical refutation. Lakatos offers a set of normative standards for evaluating a scientific research program, depending on whether its protective belt consistently generates novel confirmable predictions.1

Over the past century and a half, Marxism’s protective belt has been populated with an economic theory, a theory of history, a theory of exploitation, and a philosophical anthropology, among other things. Much like Althusserian Marxism, or the Marxism of the Frankfurt school, analytical Marxism is largely located in Marxism’s protective belt. It can be seen as a painstaking exercise in intellectual housekeeping. The exercise consists in replacing the tradition’s antiquated, superfluous, or degenerate furnishings with concepts, methods, and auxiliary hypotheses from analytic philosophy and up-to-date social science.

The thematic repertoire of analytical Marxism is extraordinarily rich: it comprises pioneering work in economics (Bardhan, 2003; Bowles & Gintis, 1990; Roemer, 1982, 1996; Van Parijs, 1993), history (Bardhan, 1983; Brenner, 1985, 2006), philosophy (Cohen, 1978, 1988, 1995; Steiner, 1994), political science (Elster, 1985; Przeworski, 1985; van der Veen, 2003), and sociology (Wright, 1985, 2010). It is no exaggeration to say that analytical Marxism was, for many years, the avant-garde of the economics and philosophy of the Left.2

There are now several comprehensive introductions to analytical Marxism, covering the full gamut from method to economics to history (see, for example, Bertram, 2008; Levine, 2003; Mayer, 1994; Van Parijs & Vrousalis, 2014; Veneziani, 2012). Instead of offering a comprehensive introduction, this article discusses the three most influential strands in analytical Marxism: its reconstruction of Marx’s theory of history, historical materialism; its philosophical anthropology, including the theory of freedom; and its theory of exploitation, including the theory of class. The focus is on the most important contributions in the area, with emphasis on seminal work by G. A. Cohen, John Roemer, and Erik Olin Wright. Of those three, Cohen’s work stands out in setting the research agenda. The narrative is therefore structured around his contributions. The article concludes with a critique of liberal egalitarianism inspired by analytical Marxism.

Before moving on to the substantive debates, a brief note on method. According to Cohen (1978), analytical Marxism “respects two constraints: on the one hand, what Marx wrote, and, on the other, those standards of clarity and rigour which distinguish twentieth-century analytical philosophy” (Cohen, 1978, p. ix). Cohen (2000) contrasts two senses of “analytical”: in the broad sense, “analysis” is contrasted with “dialectic.” In the narrow sense, “analysis” is contrasted with “holism.” Analytical Marxism is analytical in the broad sense, through its commitment to standards of rigor and clarity, as opposed to “dialectical” forms of thought (exemplified in the work of authors like David Harvey and Bertell Ollman). It is not, however, necessarily analytical in the narrow sense. What has come to be called rational choice Marxism (exemplified in the work of authors like Jon Elster and John Roemer) is distinct from analytical Marxism, in that the former, unlike the latter, is committed to methodological individualism. Methodological individualism is the view that all social phenomena are in principle explicable by appeal to properties of individuals and their relations.3 Since not all analytical Marxists affirm methodological individualism, it follows that not all analytical Marxists are rational choice Marxists.

Historical Materialism

Analytical Marxism’s founding text is Cohen’s 1978 book, Karl Marx’s Theory of History: A Defence. This book defends an “orthodox” interpretation of Marx’s theory of history, that is, an interpretation taking its cue from the Marxism of the Second International.4 Cohen’s historical materialism consists in an extensive defense of two claims: the first primacy thesis, according to which the productive forces (roughly, technology) have explanatory primacy over the economic structure (roughly, economic relations of power), and the second primacy thesis, according to which the economic structure has explanatory primacy over the superstructure (roughly, laws and ideology). What makes Cohen’s materialism materialist is that it provides a framework for understanding epochal transitions—from, say, antiquity to feudalism, or from feudalism to capitalism—through an explanatory hierarchy attaching priority to how societies reproduce themselves, as opposed to their conceptions of themselves or of their reproduction. Since most of the controversy and debate surrounding Cohen’s Marxism are concerned with the first primacy thesis, it receives the most attention in this article.5

The argument for the first primacy thesis can be roughly summarized as follows: the productive forces have an autonomous tendency to develop on the basis of asocial facts about human nature, such as rationality and self-interest (Cohen calls this the development thesis). These forces are only compatible with certain economic structures (Cohen calls this the compatibility thesis). Therefore, as the productive forces grow, there will be a tendency for the whole set of production relations to change in a direction facilitating that growth. Concomitant with growth in the productive forces, and the consequent change in economic structure, different classes will shrink and others grow, depending on the extent to which they are capable of presiding over the development of the productive forces. Ruling classes will therefore rise and fall depending on whether they can carry productive power to its optimal levels. It follows that the development of the productive forces explains the nature of the economic structure and not vice versa. As Marx (1992, p. 119) famously put it: “The hand-mill gives you society with the feudal lord, the steam-mill gives you society with the industrial capitalist.”

One novelty of Cohen’s materialism was its enfranchisement of functional explanations into Marxism. Such explanations are commonplace in fields like evolutionary biology, but not in social science. Take the common-stock example of hollow bones in birds: the possession of hollow bones normally facilitates flight, and flight is normally a consequence of the possession of hollow bones. Hollow bones, in other words, are the proximate cause of a healthy bird’s flying disposition. But we know that the main explanation why birds possess hollow bones consists in the fact that such bones tend to facilitate flight. If we take a sufficiently long time-horizon, the possession of hollow bones is explained by that consequence, namely the flying facility it furnishes. More precisely, let property P have consequence C. A functional explanation provides an explanation of P in terms of P’s tendency to facilitate C, given environmental conditions. Functional explanations in social theory introduce a similar explanatory structure. They therefore contrast with garden-variety nonfunctional explanations, which typically explain C exclusively in terms of P.

The defense of functional explanation is central to Cohen’s account of historical materialism because it offers a way out of the “overdetermination” problem plaguing Marxist thought since Plekhanov (1947). The problem, roughly stated, is that the economic structure unquestionably impacts on the productive forces, and the superstructure on the economic structure. How, then, can Cohen’s first and second primacy theses be true? A similar question is at the center of Althusser’s (1969) discussion of historical materialism. It is also the reason why he procures an overbearing conceptual machinery of “instances,” “structures in dominance,” “determination in the last instance,” and so on. According to Cohen, the overdetermination quandary can be easily solved by enfranchizing functional explanation into Marxism. Thus, (i) a particular set of production relations obtains, as it does, when and because it tends to develop the corresponding productive forces, and (ii) a particular superstructure obtains, as it does, when and because it tends to foster the corresponding set of production relations. Condition (i) is compatible with the relations impinging on the forces and (ii) is compatible with superstructure impinging on the relations. A cloud of Althusserian scholasticism is condensed into a drop of explanatory functionalism.

Cohen’s account was roundly criticized on a number of counts. Joshua Cohen (1982) criticized the asocial characterization of rationality in the elaboration of the development thesis. Elster (1980, 1982) criticized the use of functional explanation, and Levine and Wright (1980) the failure to locate the class struggle more prominently in the discussion of both primacy theses. The consensus view among Marxists was that Cohen’s development thesis is insufficiently motivated. It followed that the economic structure must have a more central role in the explanation of epochal change than Cohen’s interpretation allows. In effect, the practical upshot of Cohen’s book was a proliferation of nontechnological interpretations of historical materialism. One such interpretation is due to Richard Miller.

Miller (1984) defends an interpretation of historical materialism based on a “mode of production reading of Marx.” The mode of production is a broad term for how an item gets produced. Narrowly construed, the mode of production refers only to the technology and know-how used to produce an item. Broadly construed, the mode of production includes technology and the division of labor. Unlike Cohen, Miller construes the mode of production broadly. He also drops Cohen’s sharp content–form distinction, that is, the sharp distinction between forces and relations of production. Consider: any distinction between material and nonmaterial properties—a distinction that coincides, in Cohen’s view, with the distinction between the productive forces and everything else—must assign some pride of place to the division of labor. But it is not obvious that the division of labor is a material property, since material properties are defined, by Cohen, as properties that make no reference to power relations. And since the division of labor does entail power relations, it falls out of his definition of a material property. Cohen’s account of the productive forces is therefore excessively narrow (see Vrousalis (2015) for discussion). Miller realizes this, and drops Cohen’s distinction between the forces and the relations of production. He thus drops the explanatory primacy of the former over the latter.

Part of Miller’s explicit motivation for this revision is that it makes room for the symmetrical explanatory relationship between forces and relations of production that Cohen (1978, p. 138) disparagingly calls the “zig-zag dialectic.” An upshot of the “zig-zag dialectic” is that the relations of production, and more specifically (the nature of) control over economic surplus, take center stage once again. This is not only more faithful to the writings of Marx, says Miller, but also makes actual history directly relevant to historical materialism. The main explanation for the rise of the medieval manor, or the decline of manufacturing, for example, need not be traced back to the productive forces, but can ultimately be explained by contradictions within the mode of production itself:

In feudalism, the lords of the manor dominate, not because undominated peasants would have been less productive but because the lords’ and the peasants’ locations in the mode of production give the former so much more coordination, mobility, and concentrated leadership in the use of force.

(Miller, 1984, p. 227)

Miller’s interpretation broaches an explanatory pluralism that assigns explanatory priority to none of the entities constitutive of modes of production (technology, division of labor, economic structure). It thereby drops the first primacy thesis. It is, however, materialist in that it preserves the second primacy thesis almost intact. That is, the mode of production retains explanatory priority over the political, legal, and juridical (in a word: spiritual) phenomena covered by the superstructure. Cohen himself devotes considerable attention to spelling out exactly which phenomena are covered by such priority. In doing so, he distinguishes between inclusive and restricted historical materialism. Unlike its inclusive variant,

Restricted historical materialism does not say that the principal features of spiritual existence are materially or economically explained.… [It] only asserts that the material determines the spiritual to the extent necessary to prevent the spiritual from determining the material.

(Cohen, 1988, pp. 159–160)

In this view, shared by Miller, materialism would be conclusively refuted if one could offer examples where religious charisma, political leadership, or military genius effect epochal transformation, that is, offer explanations that have no final recourse to claims about economic structure. This is why so much Marxist firepower has been concentrated against Weber’s (1905) explanation of the rise of capitalism in terms of Calvinist ideology, to take one prominent example (see Hobsbawm, 1999, pp. 15–16). Thus a fundamental assumption of historical materialism that puts it at odds with much liberal thought is that moral and political leadership are a consequence, never the principal cause, of social change. “Spiritual leadership” is an oxymoron.

The pluralism of Miller’s mode of production interpretation flattens the explanatory hierarchy of Cohen’s materialism. Its obvious downside is that it lacks the analytical elegance of that hierarchy. It also leaves us in the dark as to which structure in the mode of production is actually doing the explanatory work. One might respond that Miller’s interpretation only provides us with a materialist armory designed to catch spiritual prey. What weapon is used in each offensive will depend on the nature of the weapon and of the prey. But the mooted objection runs deeper. For if we stick to the positivist model of scientific explanation favored by Cohen (1978), then the mode of production interpretation is a non-starter. For the latter interpretation offers many, too many, ways to explain away recalcitrant data, meaning ad hoc ways to reinterpret the theory such that it accommodates observations that appear to disconfirm it. This is not true of technological materialism, which seems more straightforwardly confirmable. Miller (1984, p. 277 and passim) grants this, but argues that Cohen’s positivism presupposes an unduly restrictive, and therefore untenable, account of what counts as a scientific theory. To what extent this recalcitrance of the mode of production interpretation turns that interpretation into a degenerating research program (in Lakatos’ sense) has not, apparently, been studied.

To sum up: the founding document of analytical Marxism is Cohen’s (1978) defense of historical materialism. Historical materialism is a substantive thesis about what explains what in history. Cohen’s historical materialism consists in an extensive defense of two claims: the first primacy thesis, according to which the productive forces—roughly, human technology—have explanatory primacy over the economic structure, and the second primacy thesis, according to which the economic structure has explanatory primacy over the superstructure—roughly, laws and the state. Both these were roundly criticized, with emphasis on Cohen’s use of functional explanation, on his commitment to the development thesis, and on his excessively narrow specification of the productive forces.


The publication of a series of rebuttals to Cohen’s technological materialism led to a lively debate in economics and analytic philosophy (Levine & Wright, 1980; Lukes, 1985; Roemer, 1982). Increasing interest in Marx’s philosophical anthropology, coupled with mounting pessimism about the explanatory credentials of historical materialism, eventually led the analytical Marxists away from the theory of history. By the mid-1980s, most analytical Marxists had turned their energies toward the theory of exploitation and the theory of freedom (Cohen, 1988; Miller, 1984; Van Parijs, 1993). Freedom is discussed in this section, exploitation in the next.

The ascendancy of the New Right in the 1980s, represented by Thatcherism in Britain and Reaganism in the United States, created a new set of problems for the Left. One important ideological endeavor of the New Right was a redefinition of the concept of freedom in ways friendly to inequality and private property. In making “promiscuous … use of the rhetoric of freedom” (Cohen, 1995, p. 67), the New Right was attempting to usurp one of the Left’s most cherished values, freedom. Cohen played a significant part in providing intellectual ammunition against this offensive. In the midst of the offensive, Cohen wrought a weapon indispensable to the critique of capitalism, a critical theory of money. A critical theory offers an understanding of a phenomenon that simultaneously constitutes a critique of it. Cohen writes:

To lack money is to be liable to interference, and the assimilation of money to physical, or even mental, resources is a piece of unthinking fetishism, in the good old Marxist sense that it misrepresents social relations of constraint as people lacking things. In a word: money is no object.

(Cohen, 2011, p. 177)

The relevant notion of “freedom” here is negative freedom, that is, absence of interference with what one might want to do (see Berlin (1969) for the original account and Carter (2012) for an overview). The fetishist “misrepresentation” of money, according to Cohen, issues from the fact that money appears to be, but is not, a mere thing. Rather, money is:

as Karl Marx said, “social power in the form of a thing,” but it is not, like a screwdriver, or a cigarette lighter, itself a thing (meaning, here, by a “thing,” a physical object), for social power is not a thing. If you swap your ten one-pound coins for a ten-pound note, you’ve got a different thing from what you had before, but the very same money. You’ve got the same license to travel, to acquire goods and services and so on, the same social entitlement, the same prospects of non-interference that you had before …

(Cohen, 2011, p. 185)

What, precisely, is the role of money? To answer this question, Cohen draws an analogy with a moneyless economy in which “the state owns everything, and in which courses of action available to people, courses they are free to follow without interference, are laid down by the law” (Cohen, 2011, p. 181). The state gives people tickets, which permit each person to perform a conjunction of actions (action X and action Y) or a disjunction of conjunctions (actions U and V, or actions W and X, or actions Y and Z, etc.). The state thus determines who can get what when. This is similar to what money does: “A sum of money is tantamount to (≠ is) a license to perform a disjunction of conjunctions of actions, actions like, for example, visiting one’s sister in Glasgow, or taking home, and wearing, the sweater on the counter at Selfridge’s” (Cohen, 2011, p. 182). But there are disanalogies:

Whereas it is the government that restricts a person’s freedom in the moneyless society, it is not, standardly, the government, but the owner of the good to which a person desires access, who, in the first instance, restricts her freedom in the money case. What the government in a money economy does is to enforce the asset-holder’s will, inter alia when that will is a will to deny access except in return for money. And the strategic role of the asset holder’s will means both that money does not absolutely ensure access … and that lack of money does not absolutely ensure lack of access.

(Cohen, 2011, pp. 182–183)

In other words, unlike the government in the state-ticket scenario, money is neither necessary nor sufficient for the performance of a particular action (or a disjunction of conjunctions thereof). It is not necessary because one might simply give the good in question without recourse to monetary exchange. It is not sufficient because one has to choose to engage in some particular act of exchange for it to be deemed socially valid. Money is, rather, an INUS condition (that is, an Insufficient but Necessary part of a condition which is itself Unnecessary but Sufficient) for that act, or the conjunction of acts thereof (Cohen, 2011, p. 177). But now note that the state-ticket case can be amended to also reflect an INUS condition:

Imagine, then, that, like money, the state tickets are neither always necessary nor always sufficient to secure goods, because state-appointed asset administrators are free, to some small extent, to grant access to ticketless people and to withhold it from people with tickets: this is an officially recognized perk of office. The administrators, let us further suppose, exercise bias in favor of some citizens and against others to precisely the same extent that private asset-holders do in the money economy … tickets establish what you are free and not free to do, not, now, to be sure, as we originally supposed, tout court, but within the feasible set established by asset administrators’ spheres of discretion.

(Cohen, 2011, p. 183)

The analogy is extraordinary because it lays bare, with unprecedented clarity, the parallels between a despotism of bureaucrats and a despotism of money. To see this, imagine that the administrators in Cohen’s original example are systematically and irreversibly biased in favor of possessors of green eyes, who happen to be 1% of the population. The 1% thus ends up with the bulk of freedom tickets. The remaining 99% get some, but the average non-green-eyed person is significantly less free than the average green-eyed person. Now suppose that the administrators are biased in favor of those who possess green pieces of paper, perhaps featuring images of dead people at their center. Imagine—by no great stretch of the imagination—that 1% of the population owns about a third of the total stock of such green papers. The 1% thus ends up with the bulk of freedom tickets. Indeed, if there is a fixed amount of total tickets, and each gets an amount proportional to their green-holdings, then the 1% gets exactly one third of the total tickets. The remaining 99% gets the rest, but the average green-poor person is significantly less free than the average green-rich person. This kind of despotism bears a remarkable resemblance to contemporary market capitalism.

Cohen’s argument has polemical significance. In Capitalism and Freedom, Milton Friedman emphasizes “the role that inequality plays in providing independent foci of power to offset the centralization of political power” (Friedman, 2002, p. 168). Friedman’s argument has a plausible major premise: the desirability of centralization-attenuating foci of political power. But Friedman invalidly infers from this premise the desirability of inequality. What he should have inferred is the desirability of equality. And this is what Cohen argues for: if your goal is the widest possible “dispersal” of freedom, then you had better opt for socialism.6 In other words, capitalist relations of production tend to reproduce centralization of political power, not in the hands of the bureaucrats of a monolithic state à la Friedman, but rather in the hands of a moneyed few: “To think of capitalism as a realm of freedom is to overlook half of its nature” (Cohen, 2011, p. 152). Both state capitalism—described in the bureaucrats case two paragraphs back—and market capitalism—described in the money scenario of the previous paragraph—are sources of unfreedom.

If Cohen is right, then lack of money implies lack of freedom, in the Right’s conception of freedom. Having hoisted one “bourgeois defense of capitalism with its own petard” (Cohen, 1995, p. 162), Cohen employs a similar line of reasoning to argue that private property also restricts freedom. According to Cohen, the Right claims that private property protects freedom. But, so long as we are talking about negative freedom, that is, absence of subjunctive interference by others, that claim is false. The enforcement of private property does imply that owners are guaranteed certain robust negative freedoms. These negative freedoms are, however, thereby denied to non-owners. I cannot, for example, pitch my tent in your backyard without (being liable to) interference from you, or your bodyguards or the police in a system where private property is universally enforced. Indeed, private property is simply a distribution of freedom and unfreedom or, more precisely:

a distribution of rights of interference. If A owns P and B does not, then A may use P without interference and B will, standardly, suffer interference if he attempts to use P.

(Cohen, 2011, p. 176)

Capitalism requires vast amounts of (threats of) interference in the service of private property: next to the invisible hand of the market, the visible fist of the capitalist state. It follows that one “cannot both deny that justice restricts freedom and claim that private property is just” (Cohen, 1988, p. 252). For private property manifestly restricts freedom, and if it is just—as proponents of capitalist private property claim—then justice restricts freedom. This is one reason why those proponents often eschew talk about freedom and instead choose to talk about liberty: the latter is more amenable to moralized, rights-based, definition than the former, in a way that does not lock them into the antiproperty dilemma just rehearsed. This is precisely the strategy pursued by some intellectuals of the New Right, such as Robert Nozick.7

Now, the defender of capitalism, and therefore of private property, might respond that under a well-enforced system of private property rights, total negative freedom is maximized. But this move, Cohen points out, is unwise, for it turns the capitalism vs. socialism debate into an empirical question—one whose answer might not satisfy defenders of capitalist private property (Cohen, 2011, p. 188). One reason why pavements, for example, should be public is that public pavements normally entail greater freedom of movement than do private pavements. Under a system of private pavements, I am liable to interference at every possible juncture of my walking life: when I walk to work, to university, to the pub. One might add that a system of free public transport analogously removes interference for long-range movement: trains, buses, and other forms of mass transit are, in that sense, just pavements on wheels (compare: seat-equipped, spacious, kilometer-long conveyor belts like the ones used at airports). And all that on the Right’s definition of freedom as negative freedom.

Thus, according to Cohen, negative freedom cannot do the work the Right assigns to it and may, indeed, serve to justify anti-Right political goals. Cohen concluded from all this that we need a theory of the just distribution of freedom. In constructing one, he thought, socialists must inevitably try to make good on the old but ambiguous socialist slogan that capitalist private property is theft, that is, a species of unjust taking. Nudged by social and political exigency, and through a series of logical steps, the analytical Marxists gradually turned to normative political philosophy.

Proletarian Unfreedom

By the beginning of the 1990s, some of the analytical Marxists defended revisions of the orthodox Marxist characterization of the relations of production, invariably premised, as it was, on the idea of force. In revising the orthodox position, Cohen (1995), Roemer (1996), Van Parijs (1993), and others attacked some of the fundamental premises on which the debate on proletarian unfreedom is commonly conducted.

According to a common-stock argument about freedom, if an agent is forced to φ‎, then she is unfree to φ‎. People on the Left assert that workers are forced to work. They infer that workers lack freedom, again defined negatively, as the absence of interference. People on the Right, on the other hand, deny that workers are unfree to work. They infer that workers are not forced to work. Cohen (1988, 1995) dissects this dialectic of modus ponens and modus tollens by pointing out that both inferences presuppose a claim about forcing. And he points out that such a claim is false: if A is forced to φ‎, then all that follows is that A is not free not to φ‎. It does not follow that A is unfree to φ‎. Indeed, Cohen argues that if A is forced to φ‎, then A must be free to φ‎: “For how can someone be forced to do what he is not free to do?” Therefore, both Left and Right are mistaken, and the common-stock claim about freedom with which this paragraph began is false (Cohen, 1988, pp. 239–255). Workers are forced to work under capitalism precisely because they are free to work.

Having made this move, Cohen goes further, by arguing that the term “workers” involves a fundamental equivocation. According to Cohen, the vast majority of able-bodied workers under capitalism are not forced to work for any capitalist. This directly contradicts Marxist lore. Indeed, most workers under capitalism are individually free, in the sense that each can make her way into the petty bourgeoisie and escape her proletarian predicament without (liability to) interference from others. But, Cohen argues, it does not follow, and it is false, that workers are collectively free—that is, free as a class. To infer collective freedom from individual freedom is to commit the fallacy of composition. For suppose all the workers were to attempt self-emancipation together. They would then find themselves incapable of jointly squeezing through the petty bourgeois loophole. For successful collective self-emancipation entails the end of capitalism (Cohen, 1988, pp. 261–265). That is, if capitalist society is founded on the exploitation of workers by capitalists, and if that exploitation depends on their collective unfreedom, then it follows that capitalism is incompatible with collective self-emancipation (see Marx, 1976, p. 1079). To illustrate this set of claims, Cohen asks us to imagine that there are ten people in a cave and only one key. Once a single person exits, the door is automatically sealed forever. Prior to someone using the key, each of the room’s denizens is free to exit. And if, for whatever reason, nobody attempts to escape, it is true of each of the denizens of the cave that she is free to exit. Cave denizens are not, however, collectively free to do so, for they cannot escape collectively. They are therefore unfree as a class. By the same token, workers are free to exit the working class in sensu diviso but not in sensu composito. This is why, according to Cohen, socialism promises “a better liberation: not just from the working class, but from class society” (Cohen, 1988, p. 265).

Proletarian unfreedom thus consists in a complex structural configuration of collective unfreedoms and individual freedoms for members of the working class. Cohen’s thesis, if true, leads to a further significant point. Take contemporary ideological beliefs, such as the American Dream. Its content ostensibly is not, “Everyone can become rich,” but “I can become rich.” This has ideological value only if the stated disposition is sometimes realized for some working class “I”s. That is, part of the allure of servitude-perpetuating ideologies issues from the fact that their content is sometimes realized for working-class people. And that fact in turn vindicates the notion that some of these people are, in fact, free. This has further implications for the theory of exploitation and class.

In summary, the claims of this section are: The idea of negative freedom, Cohen argues, cannot do the work the Right assigns to it. Indeed, when properly precisified, that idea can be turned against the Right. This possibility has led some to moralize the concept of freedom, that is, turn from a definition of freedom in terms of interference to a specification in terms of rights. The definition is then made dependent on a question about justice: who is entitled to what, when. This “entitlement turn” explains why the analytical Marxists became more interested in structures of property rights and their justification in the 1990s.


Cohen’s historical materialism set the agenda for much subsequent discussion of exploitation and the justification of private property. In Cohen’s view, Marxism sees history as “a protracted process of liberation from the scarcity imposed on humanity by nature, and from the oppression imposed by some people on others” (Cohen, 1988, p. vii). Thus, exploitation assumes different forms under different modes of production, but it necessarily involves a transfer of material surplus from one class to another. With the concomitant prodigious growth in this surplus comes a gradual diminution in “socially necessary” forms of exploitation, until all exploitation is abolished under communism. Much of John Roemer’s work in political philosophy can be read as an elaboration and extension of this view. According to Roemer (1982), the taxonomy of socially necessary forms of exploitation breaks down into ancient, feudal, capitalist, and socialist. Let Ei signify the average rate of exploitation at each stage of productive development, where i = 1 (ancient), 2 (feudal), 3 (capitalist), 4 (socialist). Cohen’s interpretation of historical materialism entails that, for i = 2, 3, that is, under feudal and capitalist relations of production, a certain amount of exploitation Eimin is socially necessary for productive progress. The rest of this section concentrates on capitalist exploitation, the historically specific form that exploitation assumes under capitalism.

According to Roemer, capitalist exploitation obtains when, for any coalition of agents A and its complement B, the following is true:

  1. (i) were B to withdraw from the society, endowed with its per capita share of society’s alienable property (that is, produced and nonproduced goods), and with its own labor and skills, then B would be better off (in terms of income and leisure) than it is at the present allocation;

  2. (ii) were A to withdraw under the same conditions, then A would be worse off (in terms of income and leisure) than it is at present;

  3. (iii) were B to withdraw from society with its own endowments (not its per capita share), then A would be worse off than at present. (Roemer, 1996, p. 40)

Conditions (i) and (ii) express the idea that exploitation is a result of distributive inequality. When A exploits B, B is worse off than she would have been with an egalitarian distribution of assets. Condition (iii) expresses the idea that if A exploits B, then A benefits at B’s expense. A, in other words, is better off than she would have been had B not existed. Conditions (i)–(iii) are therefore necessary and sufficient for capitalist exploitation.

Socialist exploitation, by contrast, is what remains once capitalist private property has been abolished. It consists in inequality due to differences in skills, talents, knowledge, etc. It is identified by replacing “alienable” for “inalienable” in the definition above. A society free from both capitalist and socialist exploitation would tolerate neither inequality due to arbitrary differences in alienable resources, such as the means of production, nor due to arbitrary differences in inalienable resources, such as talents, organizational knowledge, etc. Roemer thus claims that non-exploitation obtains if and only if both inequalities in alienable and inalienable resources are removed or significantly attenuated (this distinction takes its cue from the liberal egalitarian theories put forward by Rawls (1971) and Dworkin (2000)).8

One corollary of Roemer’s view, substantively endorsed by Cohen (1988, pp. 233–234), is that an unequal exchange of labor, understood as inequality between the amount of labor B receives in her consumption bundle, and the amount she expends in production,9 is not what makes a transaction exploitative. What matters, says Roemer, is not the flow injustice of unequal exchange (“You give fewer hours than I do”), but the stock injustice of asymmetric property relations (“You own more than I do”):

A flow of value … constitutes exploitation only if the contract it fulfils arises out of an unfair bargaining situation, and regardless, moreover, of whether or not that situation precisely forces the worker to sell his labor to the capitalist‖. The crucial question for exploitation concerns the justice of the distribution of the means of production.

(Cohen, 1988, p. 234)

This contradicts the orthodox Marxist view—entailed by the labor theory of value—to the effect that an unequal exchange of labor is both necessary and sufficient for exploitation. The analytical Marxists thus construe exploitation as a form of unreciprocated, though not necessarily forced, exchange, whose justice or fairness depends upon the justice or fairness of the distribution in the means of production. If one adds the premise that capitalist ownership of the means of production is unjust, then one gets the desired conclusion, namely that all capitalist exchanges between consenting adults are exploitative. What is crucial here is that any unjust flow constituting exploitation inherits its injustice from the background injustice in distribution.10

This account of exploitation cross-fertilized numerous auxiliary hypotheses in sociology and economics (see Wright, 1985; Skillman, 1996). It gave rise, for example, to a set of new hypotheses about the nature of class and evinced wholesale reconsideration of the individuation of capitalist relations of production. The project of redefinition of capitalist relations of production is considered here first and class is addressed second.

In the first volume of Capital, Marx assumes that prices do not diverge from values.11 He infers that surplus value arises only if there exists a commodity that produces more value than it consumes in the labor process. He then argues that the said commodity is labor power.12 In effect Marx (1976) assumes wage labor to be the only vehicle of exploitation of labor by capital. Marx (1993) then relaxes the price-value equivalence assumption. It follows that surplus value can be produced in the absence of wage labor. Usury capital and merchant capital are paradigmatic instances of this form of capitalist exploitation. Hence Marx’s (1976) definition of capital, which subsumes usury and merchant capital under the so-called “general formula for capital,” is inconsistent with his (ostensibly prior) value-theoretic assumptions.13 It further follows that the commodification of labor power is a sufficient but not a necessary condition for capital-positing labor: capitalist relations of production do not presuppose labor markets. Roemer proves this result by appeal to an isomorphism theorem. According to that theorem, capitalist exploitation is possible either through labor or through a credit market.14 Roemer’s conclusion has received historical corroboration by Banaji (2010), whose studies of usury and merchant capital in India and the Mediterranean contradict the idea that capitalist social relations presuppose the “formal” or “real” subsumption of labor by capital.15

Roemer’s isomorphism result, if true, has important implications. It implies that, insofar as capital exploits labor, “labor”—in the relevant sense—is broader than “wage labor.” Isomorphism entails, for example, that self-employed workers subsisting through dependence on moneylenders are capitalistically exploited by the latter. Not unlike usury-dependent landed peasants—who lack nonmarket access to the means of production—self-employed professionals can be capitalistically exploited by merchants, banks, and other credit institutions. It follows that the so-called “petty bourgeoisie” can be capitalistically exploited. This possibility has major implications for the Marxist theory of class. Analytical Marxism thus inaugurates a “revolution” in class theory (Van Parijs, 1993) through the notion of contradictory class locations.

The idea of a contradictory class location is due to Wright (1985). It has a basically Roemerian pedigree: Wright enlists a variant of the property-relations account of exploitation, but substantially enhances its explanatory ambitions. Wright follows Roemer in distinguishing between different kinds of alienable assets (raw materials, machinery) and inalienable assets (organizational resources, skills). Those who control such assets optimize by exploiting others. Those who do not, optimize by being exploited. Crucially, some social groups control enough of these assets to exploit those below them, but not enough to avoid being simultaneously exploited by those above them. The standard case is managers. They occupy “contradictory locations within class relations” in the following sense: managers possess enough skill or organizational knowledge that they optimize by appropriating the surplus value produced by workers who lack such assets. But they normally do not possess enough alienable assets to optimize by setting up production directly. That role lies with the bourgeoisie proper, which owns and controls the alienable means of production. Under certain conditions, this allows the bourgeoisie to exploit both managers and workers. Managers therefore occupy contradictory class locations: they exploit the propertyless, but they are exploited by the propertied. Wright extends the discussion of skills, organizational assets, and alienable assets to all types of employer, manager, and worker. This gives rise to novel theoretical predictions about class consciousness, some of which have received empirical discussion (see, e.g., Wright, 1997).

In summary, the argument of this section is: Analytical Marxism is largely responsible for the rejuvenation of exploitation theory in the 1980s and 1990s. John Roemer’s seminal contribution to this debate (Roemer, 1982, 1996) focused that debate on property relations, as opposed to the traditional Marxian focus on the unequal exchange of labor. Roemer’s contributions also generated an important literature attempting a redefinition of capitalist relations of production, as well as a novel account of the nature of class.

The Analytical Marxist Critique of Liberal Egalitarianism

Most contemporary political philosophers, including some self-described analytical Marxists, are liberal egalitarians. That is, they tend to subscribe to some version or other of Rawls’ (1971) recasting of the liberal paradigm. In this section the tools of analytical Marxism are turned against liberal egalitarianism. If this approach is correct, there is at least one sense in which liberal egalitarianism is ideological. For all present purposes, a doctrine is ideological if it serves the interests of the powerful at the expense of the vulnerable. Analytical Marxists are uniquely placed to show why this is the case.

Some Left critics of Rawls dismiss his system as “pure normativism.”16 Pace those critics, Rawls’ system has explanatory content. As Cohen puts it:

Among what contributes to the greatness of A Theory of Justice, and of the entire Rawlsian achievement, is that, to put the matter as Hegel would have done had he agreed with me, John Rawls grasped his age, or, more precisely, one large reality of his age, in thought. In his work the politics of liberal (in the American sense) democracy and social (in the European sense) democracy rises to consciousness of itself.

(Cohen, 2008, p. 11)

Thanks to its theoretical breadth and depth, Rawlsian political theory demonstrates the material ambitions of social democracy, ideally conceived. The theory is somewhat less successful, however, in accounting for its own material preconditions. Indeed, at least a part of contemporary liberal egalitarianism is untenable for exactly this reason. This critique can therefore be construed as indirect corroboration of the second primary thesis discussed above.

In his seminal study of genealogy, Bernard Williams (2002, pp. 221–223) introduces what he calls the critical theory test. The test says that if a story S serves as a putative justification for a power relationship P, and P causally explains acceptance of S, then S constitutes no genuine justification of P and P is pro tanto unjust. Now consider the following application of the critical theory test to liberal egalitarianism. (1) Capitalists have power over workers; call that power P. (2) There is a story S, liberal egalitarianism, that serves, under certain conditions, as a putative justification of P. (3) Acceptance of liberal egalitarianism is causally explained by P. So (4) liberal egalitarianism is not a genuine justification of P and P is pro tanto unjust.

Suppose this form of critique is acceptable. Then Marxists are uniquely placed to offer it with this particular content. For “P” in (1)–(4) refers to economic power, indeed to capitalist relations of production. Liberal egalitarians, on the other hand, are uniquely misplaced to offer this particular critique, for it is directed against them. However, there are some ways in which analytical Marxism can be enlisted to defend premise (3). For ease of exposition, the discussion is confined to the paradigmatic form of liberal egalitarianism, that of John Rawls.

Rawls (1971) defends what he calls the difference principle, the view that economic inequalities are justified if and only if they are necessary to improve the position of the least well off. One Rawlsian justification for this principle is that inequality provides incentives for material contribution. The idea is that unequalizing incentives might be deemed necessary by those higher up the economic hierarchy in order to contribute to the livelihood of those lower down.17 Cohen (2008) argues that this proposed justification of incentives is morally equivalent to the putative justification offered by a kidnapper to the kidnapped child’s parents. The kidnapper says: “Children should be with their parents. Unless you pay me, I shall not return your child. So you should pay me” (Cohen, 2008, p. 39). By the same token, Cohen’s Rawls says: “Economic inequalities are justified when they make the worst off people materially better off. When the top marginal rate is 40 percent, the talented produce more than when it is 60 percent, and the worst-off are, as a result, materially better off. Therefore the top tax rate should not be raised above 40 percent to 60 percent” (Cohen, 2008, p. 34). The exact details of Cohen’s critique of incentives do not matter for our purposes. What does matter is that Rawls’ difference principle purports to elicit contribution from the economically powerful for the benefit of the powerless when the former are unwilling to contribute under conditions of equality. In other words, the acceptance of Rawls’ putative justification of inequality through the difference principle is causally explained by the power of capitalists, where “capitalist” is construed broadly enough to include all those who reap market rewards by virtue of their control over marketable, capital-positing, assets. Premise (3) is therefore true: it is capitalist power that explains how those who accept the Rawlsian incentives story come to accept it.

Analytical Marxists have studied other ways in which (3) might be true. Take, for instance, the widespread liberal egalitarian concern with “efficiency” and “feasibility.” These notions are typically construed in terms of Pareto efficiency: A distribution X is more efficient than Y if and only if X makes at least someone better off and no one worse off compared to Y. The Pareto principle mandates Pareto improvements, that is, movements from X to Y when X is more efficient than Y. Przeworski (1985) has argued that, even on the Paretian definition of efficiency—whose metric is the welfare of individual workers—capitalism is more efficient than socialism. That is, even if it is admitted that socialism is generally more productive than capitalism, there is a significant valley of transition from capitalism to socialism that is likely to make workers worse off under socialism than under capitalism for a significant period of time. It is therefore in the workers’ medium-term interest not to opt for socialism and instead to strive to improve capitalism. The mechanism suggested by Przeworski is very simple: growth under capitalism depends on investment, which depends on profit, which depends on stable and relatively low wages. The prospect of a government set on significantly raising wages and/or socializing investment is therefore sufficient to precipitate a cutback of investment—a capital strike, as it is sometimes called—thereby engendering a deep and lasting recession. Recovery on the basis of socialist production, even if forthcoming, is likely to take years or decades. It follows that some workers will be significantly worse off than they would have been under capitalism. According to Przeworski (1985, pp. 171–204), the “material conditions of class compromise” ensure capitalism’s Pareto superiority to socialism.

Now note that, unlike Rawls and some Rawlsians, Przeworski never presents the Pareto story as a putative justification for inequality under social democracy. Indeed, the mechanisms identified by Przeworski lend further support to (3): it is just in virtue of capitalist power that claims about the Pareto superiority of capitalism have any purchase. Capitalist power makes it the case that socialist equality—quite literally—does not pay: equality means high pay and full employment, which engenders low profit, which engenders, by dint of capitalist power, low investment, which implies low pay and high unemployment. Thus the critical theory test forswears the Rawlsian Pareto story as a genuine justification for inequality. Capitalist relations of production remain pro tanto unjust.

Consider an analogy. Suppose you are a liberal in 19th-century England. Like most other liberals, you support the emancipation of women. You believe that a first step in that direction is the extension of the franchise. You also believe—correctly, as it happens—that all the suffragist candidates you could support are unelectable. In light of this fact, you come to ungrudgingly support a nonsuffragist candidate, reassured in your belief that the latter is going to be better for women in the long run. If you are that liberal, then you are in the grip of an ideology. Your unwrinkled reassurance, in other words, issues from the very form of power it purports to justify: patriarchy. For it is patriarchy that ensures unelectability of suffragists. Now suppose that you are on the Left, and that you have to choose between supporting a socialist candidate and a social democrat. Suppose, further, that you opt not to support the socialist because she is “unelectable” in a possible general election. If you ungrudgingly do this—in blissful allegiance to some Pareto-type justification—then you are in the grip of an ideology. For socialist candidacies in capitalist democracies are never merely unelectable, they are forced into unelectability by capitalist power, just as suffragists were once forced into unelectability by patriarchal power. No non-ideologically blighted advocate of women’s suffrage can ungrudgingly vote against all unelectable suffragists. By the same token, no non-ideologically blighted Leftist can ungrudgingly vote against all unelectable socialists.

The Cohen–Przeworski hypotheses, in conjunction with the critical theory test, thus serve to undermine the complacency of some liberal egalitarians vis-à-vis capitalist social relations. Insofar as liberal egalitarianism serves to justify the very form of power that explains its promulgation and acceptance, that theory is ideological in the pejorative sense. This argument, if sound, illuminates one way in which the analytical Marxist armory is superior in firepower relative to that of the liberal egalitarian. Non-Marxists may, of course, proffer a similar critique on the basis of the critical theory test. Some critical theorists, after all, are not Marxists. The difference between the former and the latter is that the latter have a scientific account of the explanatory relationship between power—or, more generally, the economic structure—and ideology—or, more generally, the superstructure. Analytical Marxists are therefore privy to a scientifically respectable set of explanatory claims that are unavailable to non-Marxists.


This article discusses the three most important contributions of analytical Marxism to social and political theory: historical materialism, the theory of freedom, and the theory of exploitation. Analytical Marxism’s founding document, Cohen’s (1978) Karl Marx’s Theory of History: A Defence, was largely responsible for setting the agenda in all three areas. Its main aspirations were, first, to understand “what Marx wrote,” and, second, to do so using “those standards of clarity and rigour which distinguish twentieth-century analytic philosophy.” These two aspirations remain at the center of the analytical Marxist approach. The analytical Marxists have also bequeathed us with a number of unresolved philosophical problems: Is historical materialism a progressive or a degenerate research program? How are freedom and exploitation related? What are the connections between exploitation and competitive markets? What is the place of exploitation in Marx’s critique of political economy? These are all questions that analytical Marxists are well placed to answer. Indeed, if the criticism of liberal egalitarianism outlined here is sound, then they are uniquely well placed to answer them.


Thanks to Kasper Lippert-Rasmussen, Roberto Veneziani, and two anonymous referees for very helpful written comments on an earlier draft of this article.


Althusser, L. (1969). For Marx. Harmondsworth, U.K.: Penguin.Find this resource:

Banaji, J. (2010). Theory as history. London: Haymarket Books.Find this resource:

Bardhan, P. (1983). Agrarian relations in West Bengal. Bombay: Somaiya.Find this resource:

Bardhan, P. (2003). Poverty, agrarian structure and political economy in India. New Delhi: Oxford University Press.Find this resource:

Barry, B. (1989). Theories of justice. Berkeley: University of California Press.Find this resource:

Berlin, I. (1969). Four essays on liberty. Oxford: Oxford University Press.Find this resource:

Bertram, C. (2008). Analytical Marxism. In J. Bidet & S. Kouvelakis (Eds.), Critical companion to contemporary Marxism. Leiden, The Netherlands: Brill.Find this resource:

Bowles, S., & Gintis, H. (1990). Contested exchange. In L. Putterman & R. Kroszner (Eds.), The economic nature of the firm: A reader. Cambridge, U.K.: Cambridge University Press.Find this resource:

Brenner, R. (1985). Agrarian class structure and economic development in pre-industrial Europe. In T. H. Alston & C. H. E. Philpin (Eds.), The Brenner debate. Cambridge, U.K.: Cambridge University Press.Find this resource:

Brenner, R. (2006). The economics of global turbulence. London: Verso.Find this resource:

Burawoy, M. (1990, December). Marxism as Science. American Sociological Review, 55, 775–793.Find this resource:

Carter, I. (2012). Positive and negative liberty. In N. Zalta (Ed.), Stanford Encyclopedia of Philosophy. Stanford, CA: The Metaphysics Research Lab.Find this resource:

Cohen, G. A. (1978). Karl Marx’s theory of history: A defence. Princeton, NJ: Princeton University Press.Find this resource:

Cohen, G. A. (1988). History, labour and freedom. Oxford: Oxford University Press.Find this resource:

Cohen, G. A. (1995). Self-ownership, freedom and equality. Cambridge, U.K.: Cambridge University Press.Find this resource:

Cohen, G. A. (2000). Karl Marx’s theory of history: A defence (Rev. ed.). Oxford: Oxford University Press.Find this resource:

Cohen, G. A. (2008). Rescuing justice and equality. Cambridge, MA: Harvard University Press.Find this resource:

Cohen, G. A. (2011). On the currency of egalitarian justice (M. Otsuka, Ed.). Princeton, NJ: Princeton University Press.Find this resource:

Cohen, J. (1982). Review of Karl Marx’s theory of history. Journal of Philosophy, 79, 253–273.Find this resource:

Dworkin, R. (2000). Sovereign virtue. Cambridge, MA: Harvard University Press.Find this resource:

Elster, J. (1980). Cohen on Marx’s theory of history. Political Studies, 28, 121–128.Find this resource:

Elster, J. (1982). Marxism, functionalism, and game theory. Theory and Society, 11, 453–482.Find this resource:

Elster, J. (1985). Making sense of Marx. Cambridge, U.K.: Cambridge University Press.Find this resource:

Friedman, M. (2002). Capitalism and freedom. Chicago: University of Chicago Press.Find this resource:

Geras, N. (1985). The controversy about Marx and justice. New Left Review, 150, 47–85.Find this resource:

Gramsci, A. (1971). Selections from the prison notebooks (Q. Hoare, Ed.). London: International Publishers.Find this resource:

Heath, J. (2009). Habermas and analytical Marxism. Philosophy and Social Criticism, 35, 891–919.Find this resource:

Hobsbawm, E. (1999). Industry and empire. Harmondsworth, U.K.: Penguin.Find this resource:

Lakatos, I. (1978). The methodology of scientific research programmes: Philosophical papers, Volume I. Cambridge, U.K.: Cambridge University Press.Find this resource:

Leopold, D. (2008). Dialectical approaches. In D. Leopold & M. Stears (Eds.), Political theory: Methods and approaches. Oxford: Oxford University Press.Find this resource:

Levine, A. (2003). A future for Marxism? London: Pluto Press.Find this resource:

Levine, A., & Wright, E. O. (1980). Rationality and class struggle. New Left Review, 123, 47–68.Find this resource:

Lukes, S. (1985). Marxism and morality. Oxford: Oxford University Press.Find this resource:

Mayer, T. F. (1994). Analytical Marxism. London: SAGE.Find this resource:

Marx, K. (1976). Capital (Vol. I). Harmondsworth, U.K.: Penguin.Find this resource:

Marx, K. (1992). Poverty of philosophy. New York: International Publishers.Find this resource:

Marx, K. (1993). Capital (Vol. III). Harmondsworth, U.K.: Penguin.Find this resource:

Miller, R. W. (1984). Analyzing Marx. Princeton, NJ: Princeton University Press.Find this resource:

Plekhanov, G. (1947). The monist view of history. London: Lawrence and Wishart.Find this resource:

Przeworski, A. (1985). Capitalism and social democracy. Cambridge, U.K.: Cambridge University Press.Find this resource:

Rawls, J. (1971). A theory of justice. Cambridge, MA: Harvard University Press.Find this resource:

Roemer, J. E. (1982). A general theory of exploitation and class. Cambridge, MA: Harvard University Press.Find this resource:

Roemer, J. E. (1988). Free to lose. Cambridge, MA: Harvard University Press.Find this resource:

Roemer, J. E. (1995). Analytical Marxism. Cambridge, U.K.: Cambridge University Press.Find this resource:

Roemer, J. E. (1996). Egalitarian perspectives. Cambridge, U.K.: Cambridge University Press.Find this resource:

Steiner, H. (1994). An essay on rights. Oxford: Basil Blackwell.Find this resource:

Skillman, G. (1996). Marxian value theory and the labor-labor power distinction. Science and Society, 60, 427–451.Find this resource:

Skillman, G. (2007). Value theory vs. historical analysis in Marx’s account of capitalist exploitation. Science and Society, 71, 203–226.Find this resource:

Van der Veen, R. J. (2003). Basic income on the agenda. Amsterdam: Amsterdam University Press.Find this resource:

Van Parijs, P. (1993). Marxism recycled. Cambridge, U.K.: Cambridge University Press.Find this resource:

Van Parijs, P., & Vrousalis, N. (2014). Analytical Marxism. In J. Wright (Ed.), International encyclopedia of social and behavioral sciences (2d ed., Vol. I, pp. 665–667). Oxford: Elsevier.Find this resource:

Veneziani, R. (2012). Analytical Marxism. Journal of Economic Surveys, 26, 649–673.Find this resource:

Vrousalis, N. (2015). The political philosophy of G. A. Cohen. London: Bloomsbury.Find this resource:

Warren, P. (1997). Should Marxists be liberal egalitarians? Journal of Political Philosophy, 5, 47–68.Find this resource:

Warren, P. (2015). In defence of the Marxian theory of exploitation. Social Theory and Practice, 41, 286–308.Find this resource:

Williams, B. A. O. (2002). Truth and truthfulness. Cambridge, U.K.: Cambridge University Press.Find this resource:

Wolff, R. P. (1985). Understanding Marx. Oxford: Basil Blackwell.Find this resource:

Wright, E. O. (1985). Classes. London: Verso.Find this resource:

Wright, E. O. (1997). Class counts. Cambridge: Cambridge University Press.Find this resource:

Wright, E. O. (2010). Envisioning real utopias. London: Verso.Find this resource:

Wright, E. O., Levine, A., & Sober, E. (1992). Reconstructing Marxism. London: Verso.Find this resource:


(1.) For an attempt to apply Lakatos’ normative standards to Marxism, see Lakatos (1978, pp. 5–6). For a contrary view, see Burawoy (1990).

(2.) For a short history of the so-called “September group,” formed by Cohen, Elster, and Roemer in 1981, see Cohen (2000).

(3.) Methodological individualism rules out structural constraints on social explanation. It is therefore inconsistent with the views of those analytical Marxists who affirm the existence of such constraints. For an internal critique of rational choice Marxism along these lines, see Veneziani (2012).

(4.) The Second International was the world’s largest and most powerful federation of socialist parties before the First World War. It was founded in the 1880s with the active support of Friedrich Engels. Its chief theoreticians (Karl Kautsky, Eduard Bernstein, August Bebel, and Franz Mehring) were all associated with the Social Democratic Party of Germany.

(5.) For a more detailed discussion of Cohen’s materialism and a critique of the distinction between forces and relations of production, see Vrousalis (2015, chapter 1).

(6.) Socialism is, of course, compatible with the existence of money as a medium of exchange. But it is incompatible with the pervasiveness of money as capital, that is, as a monopoly of the means of production in the hands of one class.

(7.) This strategy led to a series of responses, collected in Cohen (1995). For a summary of the Cohen–Nozick debate, see Vrousalis (2015).

(8.) Rawls, for example, distinguishes between natural and social primary goods. A primary good is an “all-purpose means,” a good that is indispensable to the realization of a rational plan of life. Natural primary goods are goods like health and talent. Social primary goods are goods like income and wealth. Although both Roemer and Rawls are interested in distribution, the latter has no theory of exploitation.

(9.) More precisely, an unequal exchange of labor between A and B obtains if A’s consumption bundle embodies X hours of labor time, B’s consumption bundle embodies Y hours of labor time, A’s product embodies Y, B’s product embodies X, and X > Y > 0. Roemer (1982, 1996) enlists the unequal exchange theory as a coherent alternative to the labor theory of value.

(10.) Cohen changed his view on the significance of labor flow but remained ambiguous about the exact wrong-making features of exploitation (see Vrousalis, 2015, chapter 4). Warren (1997, 2015) discusses the implications of this view for the distinction between Marxism and liberal egalitarianism.

(11.) Marx’s (1976) argument for the labor theory of value has been extensively criticized by Cohen (1988, chapter 11), Roemer (1982), Wolff (1985), and Skillman (1996). The consensus among economists is that Marx’s argument, interpreted as a theory of equilibrium price, is logically incoherent.

(12.) As Roemer (1996, chapter 3) shows, this further inference is invalid.

(13.) This point has been extensively defended by Skillman (1996, 2007).

(14.) The isomorphism theorem is discussed informally in Roemer (1988, pp. 95–97) and is proven formally in Roemer (1982, pp. 89–94).

(15.) This argument is discussed extensively in a paper titled “How Capitalists Dominate,” which is available from the author upon request.

(16.) Heath (2009, p. 918) attributes this reaction to Habermas, upon reading Rawls. This kind of reaction is not uncommon among Leftist critics of Rawls.

(17.) Rawls seems to sanction the Pareto argument for inequality (in Rawls, 1971, p. 78). Barry (1989, pp. 213–215) formulates the justification more succinctly.