* Full disclosure: I was one of the readers of Cohen's book for Harvard University Press. The dust jacket quotes the part of my report to the press characterizing the book as the deepest and most sophisticated critical work on Rawls's theory of justice. Another part of that same report said that ‘almost every step of the way, I found myself objecting to particular claims, worrying that Cohen had misconstrued Rawls, and developing counter arguments of my own. Each of these responses would require at least an article of its own, and are a measure of the book's depth and engagement.’ This critical notice is the first installment of that response. I am grateful to Peter Benson, Mohammad Fadel, Joseph Heath, Katrin Flikschuh, John Mandle, Christopher Morris, Sergio Tenenbaum, Helga Varden, and Jacob Weinrib for comments on an earlier draft, and to Paul Weithman and the members of his seminar at Notre Dame for discussing it and Cohen's book with me. Just before he died, Jerry Cohen sent me a note in which he described my argument as ‘very challenging’ and said that he would respond to it. Sadly, that was not to be. I dedicate this essay to his memory.

1 See Cohen, G.A. Self-ownership Freedom and Equality (Cambridge: Cambridge University Press 1995), 5ff.CrossRefGoogle Scholar

2 Parenthetical references are to Cohen, Rescuing Justice and Equality.

3 Rawls, John A Theory of Justice (Cambridge MA: Harvard University Press 1971), 74.Google Scholar Abbreviated as TJ.

4 TJ, 67

5 In the example through which he introduces the regress argument, he endorses the suggestion that a particular facts can be relevant to a principle simply by instantiating it, citing the example of promise-keeping. However, instantiation does not figure prominently in his deployment of the argument against Rawls.

6 Despite Cohen's characterization of the mode of argument in the first part of the book, his initial argument about the difference between bargaining and community also engages the Rawlsian project more directly, by interrogating the grounds on which citizens can make claims against each other. Cohen's point in that argument operates within Rawls's idea that a basis for justification presupposes citizens capable of taking responsibility for their choices, and uses that idea to show that those who are unwilling to work without incentives are treating their own choices as natural facts about them, for which they bear no responsibility. As such, the earlier argument takes seriously Rawls's claim to be seeking a regulative principle appropriate to the object in question, that is, society conceived as a fair system of cooperation, which is supposed to govern the claims that citizens can make against each other and their, institutions. Cohen's argument is, in its simplest form, is that a fair system of cooperation cannot coherently allow citizens to bargain or threaten with respect to the terms on which they are required to contribute to their society and its institutions. The incentives argument also focuses on the basic structure, because it objects to those who, like Nigel Lawson, Margaret Thatcher's Chancellor of the Exchequer, seek to change the tax code.

7 Pogge, Thomas ‘Cohen to the Rescue,’ Ratio (new series) 21 (2008), 454–75.Google Scholar

8 Pogge, 460

9 TJ Second edition 137. Cohen objects to Rawls's further remark that such a strategy entails making up facts too (261). But Rawls's point is that the principle appropriate to a thing depends on its nature.

10 Rawls, Political Liberalism, 4.Google Scholar In the preface to A Theory of Justice, Rawls characterizes his theory as the ‘most appropriate moral basis for a democratic society’ (TJ xviii). See also the discussion in Cohen, Joshua ‘For a Democratic Society’ in The Cambridge Companion to Rawls, Freeman, Samuel ed. (Cambridge: Cambridge University Press 1999).Google Scholar

11 It thus meets Cohen's requirements better than Cohen's own luck-egalitarianism. Pogge points out that although Cohen presents it as an example of a fact-free principle, it fails by his stated standards. Its directives ‘are not general enough to settle how they apply to animals, beings on distant planets or humans living in distant ages — let alone how they apply in very different possible universes’ (‘Cohen to the Rescue’ 469, n. 15). At various points in the book, Cohen also endorses the idea of an ‘agent centered prerogative’ and rejects the ‘moral rigoist’ position that denies people the right to pursue self-interest to some extent (61), in favor of a ‘properly prerogative-informed egalitarian principle.’ It appears to be an exception to, rather than a part of distributive justice (10, 387-9), but it is difficult to see what its moral basis is supposed to be in the absence of some defense of the factual premise that people have inclinations to pursue things other than justice. If so, it appears to violate Cohen's declared methodological scruples.

12 Though sometimes under that name. In the Critique of Pure Reason, Kant distinguishes ideas, which are pure rational concepts, from ideals, which include an empirical part but have practical power as regulative principles ‘and form the basis of perfection of certain types of actions. He gives the example of ‘The wise man (of the Stoics) is, however, an ideal, that is, a man existing in thought only, but in complete conformity with the idea of wisdom’ (Kant, Critique of Pure Reason, Guyer, Paul and Wood, Alan trans. [Cambridge: Cambridge University Press 1998]CrossRefGoogle Scholar A569/B597). The Stoic sage is an ideal of a perfectly wise human being, not a set of precepts adopted on empirical grounds about how one is most likely to succeed in achieving a type of wisdom that exceeds the human. On ideals as partial determination of principles in Kant, see Herman, Barbara ‘A Cosmopolitan Kingdom of Ends’ in her Moral Literacy (Cambridge, MA: Harvard University Press 2007).Google Scholar

13 This is the form of ‘Problems I and II’ in Kant, Critique of Practical Reason, Gregor, Mary trans., in Kant, Practical Philosophy (Cambridge: Cambridge University Press 1996), 162,Google Scholar AK 5:29.

14 For a helpful discussion of some of the differences between Kant and Rawls, see O’Neill, Onora ‘Constructivism in Rawls and Kant,’ in The Cambridge Companion to Rawls Freeman, Sam ed. (Cambridge: Cambridge University Press 2003) 247–67.Google Scholar

15 ‘Kantian Constructivism in Moral Theory’ in Collected Papers, Samuel Freeman, ed. (Cambridge, MA: Harvard University Press 1999), 304

16 Rawls's treatment of Kant's categorical imperative explicitly incorporates facts in a way that could be read as seeking to respond to Hegel's charge that Kant's ethics is merely an ‘empty formalism’ (Rawls, Lectures on the History of Moral Philosophy, Herman, Barbara ed. [Cambridge, MA: Harvard University Press 2000], 163Google Scholar).

17 A Theory of Justice 159/138, quoted at 262

18 See Rawls, Justice as Fairness: A Brief Restatement, Kelly, Erin ed. (Cambridge, MA: Harvard University Press 2001), 3;Google Scholar Lectures on the History of Political Philosophy, Freeman, Sam ed. (Cambridge, MA: Harvard University Press 2007), 10.Google Scholar Neither Kant nor Hegel figures in Rawls's Lectures on the History of Political Philosophy, but Hegel's critique of Kantian ethics is a central theme of his Lectures on the History of Moral Philosophy.

19 See the discussion of this point in Thompson, Michael Life and Action: Elementary Structures of Practical Thought (Cambridge MA: Harvard University Press 2008).CrossRefGoogle Scholar

20 In the Critique of Pure Reason, Kant also talks about species concepts, through which the operation of the parts of a living thing are thought of as subordinated to a thing as a whole (A 316/B373). Section 65 if the Critique of The Power of Judgment makes it clear that he regards natural teleology as a regulative concept for reflective judgment ‘to guide our investigation of organized objects,’ which does not provide knowledge of the objects, but only a principle to guide study of them, so that in the case of a living thing, the species concept is imposed on material nature (Critique of the Power of Judgment, Paul Guyer and Eric Matthews, trans. [Cambridge: Cambridge University Press 2001] 5:373). Kant's approach differs from Hegel's because he regards teleological judgments as rational constructs, so that their perfective aspect is not real in time.

21 TJ, 7

22 See Joshua Cohen, ‘For a Democratic Society.’

23 Rawls's rejection of the analytic/synthetic distinction counts against attributing an essentialist position to him. Cohen's embrace of that distinction makes the essentialism charge a difficult one for him to press. Cohen expresses misgivings about the role of biological metaphors in Hegel's philosophy in Chapter 4 of If You’re an Egalitarian, How come You’re so Rich? (Cambridge, MA: Harvard University Press 2000). Perhaps some of these could be raised against the Hegelian elements in Rawls.

24 See James, Aaron ‘Constructing Justice for Existing Practice: Rawls and the Status Quo,’ Philosophy & Public Affairs 33 (2005) 281–316.CrossRefGoogle Scholar

25 As Christine Korsgaard puts it, ‘Practical philosophy, as conceived by Kant and Rawls, is not a matter of finding knowledge to apply in practice. It is it is rather the use of reason to solve practical problems. The concepts of moral and political philosophy are the names of those problems, or more precisely of their solutions.’ See ‘Realism and Constructivism in Twentieth Century Moral Philosophy’ in her The Constitution of Agency: Essays on Practical Reason and Moral Psychology (Oxford: Oxford University Press 2008), 322.

26 TJ 151/131, cited by Cohen at 81. Cohen offers an elaboration and defense of the idea that people should serve one another in his Why Not Socialism? (Princeton: Princeton University Press 2009).

27 For such contestation, see Onora O’Neill, ‘Political Liberalism and Public Reason: A Critical Notice of John Rawls, Political Liberalism,’ The Philosophical Review 106 (1997) 411-28. I raise a different Kantian objection to Rawls's focus on benefits of cooperation in the first chapter of Force and Freedom (Cambridge MA: Harvard University Press 2009).

28 TJ, 242. Rawls appears to be deploying the distinction between the demands of justice and the acceptable rules of regulation in his characterization between justice and peace at ‘the legislative stage,’ suggesting that ‘citizens may affirm the law as the lesser of two evils.’ He concedes that the law is unjust, remarking that ‘Since bitter dissensions exist, there is no way to prevent some injustices, as we ordinarily think of them, from occurring. All that can be done is to limit these injustices in the least unjust way.’ In this example, the existence of bitter dissensions enters as a limit on the achievement of justice.

29 Dworkin, Ronald Justice for Hedgehogs (Cambridge, MA: Harvard University Press forthcoming 2010)Google Scholar

30 In the next paragraph, he does note, however, that someone might regard promissory fidelity as itself a fundamental principle.

31 Rawls himself points to two differences between Sidgwick's view and constructivism: Sidgwick regards justification as ‘primarily an epistemological problem’ possiabout an independent order of moral values, rather than as an issue for practical reason about working out the implications of a conception of the person, ‘Kantian Constructivism’ in Collected Papers (554), and in the role each assigns to human powers of reflection and judgment (560). These differences help to explain Sidgwick's attitude towards rules, since the possibility of exceptions both makes them poorly suited to be objects of knowledge and demands too much scope for reflection and judgment. It is not clear whether Cohen shares Sidgwick's views on these matters. Both his methodology of approaching principles separately and his claim that ‘political philosophy is concerned with what we should think rather than what we should do’ suggest some attachment to them.

32 Mill, Utilitarianism, Chapter 5 in The Collected Works of John Stuart Mill Vol. 10, Robson, J.M. ed. (Toronto: University of Toronto Press 1969), 259Google Scholar. Mill had earlier characterized the broader distinction between right and wrong in terms of liability to some sort of sanction, whether coercive, social or in conscience (246).

33 Sidgwick characterizes a plurality of potentially conflicting ultimate principles as ‘absolute proof that at least one of the formulae needs qualification’ (Methods of Ethics [Indianapolis: Hackett 1981] Book III, Chapter XI, 341).

34 Cohen's luck egalitarian conception of justice as a property of distributions and his pessimism about its achievement also echo Sidgwick, who writes ‘For in any case it does not seem possible to separate and practice that part of the man's achievement which is due strictly to his free choice from that part which is due to the original gift of nature and of favoring circumstances’ (Methods of Ethics, Book III, Chapter V, 285).

35 Although I cannot take this matter up in any detail here, it is difficult to see the basis, within his own view, for Cohen's appeal to facts about the cognitive mechanisms thorough which people deal with competing claims, or about the way principles present themselves ‘in competitive array.’ The fact that people experience conflicts of value, or have no difficulty resolving them, may speak to what they do, but not to what they ought to do or think about such conflicts. Some fact-free principle would be required to make these facts relevant.

36 Lavin, Douglas ‘Practical Reason and the Possibility of Error,’ Ethics 114 (2004) 424–57CrossRefGoogle Scholar

37 It must be conceded here that Rawls's presentation of his view sometimes invites the suspicion that he shares the idea that rules are to be selected on the basis of factors that are not themselves rule-based. In A Theory of Justice his discussion of promissory obligation distinguishes what he calls the ‘constitutive convention’ of promise-keeping from its moral justification. In so doing he appears to commit himself to Bentham's fundamental distinction between the existence of a rule and its merits or demerits, with the concomitant thought that the merits are not themselves a matter of rules. His appeal to H.L.A. Hart's ‘principle of fair play’ might seem to underscore his attitude. However, Cohen's attribution of the idea that rules are instruments for bringing about results that are not themselves rule-like does not draw on this passage.

38 I explain the role of postulates in Kant's Rechtslehre in Force and Freedom, Appendix.

39 See The Methods of Ethics Book III, Chapter V, 350, where Sidgwick concludes that the principles of justice, among which he explicitly includes what he calls the ‘socialistic principle of requiting Desert’ cannot be ultimate because it cannot be made suitably precise.

40 Cohen says that his focus is on ‘distributive justice, by which I uneccentrically mean justice (and its lack) in the distribution of benefits and burdens to individuals’ (126). I do not mean to deny that there is such a topic, only that it is a central topic in Rawls's theory.

41 TJ, 74

42 See, for example, Nagel, Thomas ‘Rawls and Liberalism’ in The Cambridge Companion to Rawls, Freeman, Samuel ed. (Cambridge: Cambridge University Press 2003);Google Scholar Kymlicka, Will Contemporary Political Philosophy: An Introduction (Oxford: Oxford University Press 1991).Google Scholar Thomas Nagel and Liam Murphy present Rawls as endorsing the luck-egalitarian principle, but as skeptical about the ability of institutions to implement it. See The Myth of Ownership (New York: Oxford University Press 2002), 55. Cohen himself credits David Lyons with first noticing that the argument for the difference principle was not one of justice in his ‘The Nature and Soundness of the Contract and Coherence Arguments,’ in Norman Daniels, ed., Reading Rawls (New York: Basic Books 1975), 152-3.

43 Samuel Freeman suggests that the luck-egalitarian reading originates in Robert Nozick's discussion of Rawls in his Anarchy, State and Utopia (New York: Basic Books 1974) 213-31. See Freeman, ‘Rawls and Luck Egalitarianism’ in his Justice and the Social Contract: Essays on Rawls (Oxford: Oxford University Press 2009), 114.Google Scholar Jon Mandle points out that Rawls explicitly says that ‘the natural distribution is neither just nor unjust; nor is it unjust that persons are born into society at some particular position. These are simply natural facts. What is just and unjust is the way that institutions deal with these facts.’ See Mandle, Rawls's Theory of Justice: an Introduction (Cambridge: Cambridge University Press 2009), 49,CrossRefGoogle Scholar quoting TJ 102/87. Mandle also points to Rawls's explicit rejection of the claim that inequalities of birth and natural talent are to be redressed (24), citing TJ 102/87.

44 Political Liberalism, 265

45 TJ, 76

46 Cohen mentions pure procedural justice in passing three times, at 126 n. 20, 220, and 332 n. 89. In each case, he rejects the idea that rules could be the subject matter of justice.

47 As early as his 1958 paper ‘Justice as Fairness,’ Rawls distinguishes the role of equality in justice from its role in a comprehensive conception. Collected papers, Freeman, Samuel ed. (Cambridge MA: Harvard University Press 1999), 48.Google Scholar

48 Scheffler, Compare ‘What is Basic About the Basic Structure?’ in The Egalitarian Conscience: Essays in Honour of G.A. Cohen, Sypnowich, Christine ed., (Oxford: Oxford University Press 2008).Google Scholar

49 Rawls, ‘Social Unity and Primary Goods,’ in Utilitarianism and Beyond, Sen, Amartya & Williams, Bernard eds., (Cambridge: Cambridge University Press 1982), 170Google Scholar

50 A parallel point could be made about publicity, a topic to which Cohen devotes an entire chapter. If justice is the organizing principle of a system of cooperation in light of which people make claims against their common social institutions, the principles will need to be broadly graspable by ordinary citizens making such claims. It does not follow from this that they must be public in the strong sense in which Cohen criticizes Andrew Williams for regarding as necessary or desirable to justice, that is, providing exact prescriptions for particular actions. Instead, they only need to be capable of being given public institutional expression through, for example, tribunals and courts, so that there is a public procedure for ascertaining the merits of particular claims. This same need for publicly articulated standards underlies Rawls's objection to intuitionism in political philosophy. Principles that present themselves ‘in competitive array’ do not provide a public basis of justification. Once again, the fundamental contrast is between Rawls's practical conception of normative theory and the epistemological conception found in Sidgwick and Cohen.