Why is (it that) common priors are implicit or explicit in the vast majority of the differential information literature in economics and game theory? Why has the economic community been unwilling, in practice, to accept and actually use the idea of truly personal probabilities in much the same way that it did accept the idea of personal utility functions? After all, in (Savage's expected utility theory), both the utilities and probabilities are derived separately for each decision maker. Why were the utilities accepted as personal, and the probabilities not?

It is universally agreed that involuntary unemployment is an evil for unemployed individuals, who lose both income and the non-pecuniary benefits of paid employment, and for society, which loses the productive labor that the unemployed are unable to expend. It is nearly as widely agreed that there is at least a prima-facie case for alleviating this evil – for reasons of justice and/or benevolence and/or social order. Finally, there is little doubt that the evils of involuntary unemployment cannot be adequately addressed in contemporary societies without state intervention – whether through monetary or fiscal policies, cash payments or other subsidies to the unemployed, direct provision of employment by the state, or some combination of these measures.

Roemer has a striking and novel criticism of Marx's thesis that capitalism is exploitative. Even if so, Roemer claims, that does not establish that it is morally objectionable. He argues from idealized models that involve what he calls ‘Marxian exploitation’, but do not have the consequences which, he says, Marxists think make exploitation bad – domination, alienation, and inequality – or have them only in unobjectionable ways. These models, Roemer thinks, show that the moral problem with capitalism that Marxian exploitation locates only imperfectly is unjust inequality. He develops an alternative notion of ‘property relations exploitation’, based on this charge, to replace the morally uninteresting notion of Marxian exploitation in assessing arguments for socialism.

In his book A General Theory of Exploitation and Class (1982) (hereafter GT), John Roemer employs the tools of mainstream general equilibrium and game-theoretic analysis to develop a fundamental critique and broadbased reformulation of Marxian economic theory. Perhaps Roemer's most striking departure from traditional Marxian tenets lies in his explanation of the material basis of exploitation in capitalist economies. Roemer argues that capitalist exploitation must be understood as essentially the consequence of exchange given differential ownership of relatively scarce productive assets (DORSPA). In particular, Roemer concludes that capitalist exploitation does not fundamentally depend on capitalist domination of production, or what Marx termed the subsumption of labor under capital.

The prisoner 's dilemma game (henceforth, PD) has acquired large literatures in several disciplines. It is surprising, therefore, that a good definition of the game is hard to find. Typically an author relates a story about captured criminals or military rivals, provides a particular payoff matrix and asserts that the PD is characterized, or illustrated, by that matrix. In the few cases in which characterizing conditions are given, the conditions, and the motivations for them, do not always agree with each other or with the paradigm examples elsewhere. In this paper we describe several varieties of PD's. In particular, we suggest there are two distinctions among PD's with philosophical significance, the pure/impure and the utilitarian/nonutilitarian distinctions. In the first section, we explain and characterize the two distinctions. In the second, we discuss an issue of moral philosophy that illustrates the significance of the former.

In his Comment ‘Adam Smith on the Morality of the Pursuit of Fortune’, Richard Arlen Kleer accepts much of the argument in my article ‘Signifying Voices’ (Brown, 1991) but insists that I have ‘gone too far’ (Kleer, 1993). Kleer agrees that there is a moral hierarchy in Adam Smith's Theory of Moral Sentiments (TMS) where benevolence and self-command are ranked higher than justice and prudence, but he is uneasy with the conclusion that economic activity and the pursuit of gain are ‘amoral’ activities and insists that they do have a significant moral standing. In addition, although Kleer accepts a good deal of the stylistic analysis, again he is uneasy with the results that are derived from it. This reply will take each of these aspects in turn.