1 Cf. 1971: 19-21, 46-51, 579-81; 1974: 7; 1980: 534. References to Rawls’ writings are by year of publication and page. My discussion will draw on ‘Outline for a Decision Procedure in Ethics,’ Philosophical Review 60 (1951) 177–97CrossRefGoogle Scholar reprinted in Thomson, J. and Dworkin, G. eds., Ethics (New York: Harper and Row 1968);Google Scholar A Theory of Justice (Cambridge, MA: Harvard University Press 1971); ‘The Independence of Moral Theory,’ Proceedings and Addresses of the APA 48 (1974) 5–22Google Scholar; ‘Fairness to Goodness,’ Philosophical Review 84 (1975) 536-54; ‘A Well-Ordered Society’ in Laslett, P. and Fishkin, J. eds., Philosophy, Politics, and Society, fifth series (New Haven: Yale University Press 1979)Google Scholar; ‘Kantian Constructivism in Moral Theory’ The Journal of Philosophy 77 (1980) 515–72CrossRefGoogle Scholar; and ‘Social Unity and Primary Goods’ in Sen, A. and Williams, B. eds., Utilitarianism and Beyond (New York: Cambridge University Press 1982)CrossRefGoogle Scholar.

2 Cf. Daniels, Norman ‘Wide Reflective Equilibrium and Theory Acceptance in Ethics’ Journal of Philosophy 76 (1979) 256–82CrossRefGoogle Scholar and my Moral Realism and the Foundations of Ethics (New York: Cambridge University Press forthcoming), chapters 5 and 6.

3 We can distinguish between relativist and nonrelativist forms of constructivism. Relativist constructivism (relativism) is true just in case there are a plurality of sets of moral facts each constituted by different moral beliefs or different bodies of moral beliefs. According to moral relativism, x is a moral fact for S (x is true-for S) just in case S believes x, S would believe x upon reflection, S is part of a social group the majority of whom believe x, or some such thing. Nonrelativist constructivism (constructivism) holds that there is a single set of moral facts which are constituted by some function of our moral beliefs, e.g., by our moral beliefs in reflective equilibrium. Identifying the moral facts with those moral beliefs in reflective equilibrium will be a form of constructivism if there is a unique equilibrium point; it will be a form of relativism if there are only equilibria.

4 Another alternative would be for Rawls to make the normative claim that it is always right for a society to maintain and pursue those assumptions which can be ascribed to its political tradition, regardless of the moral content of those assumptions. I assume that it is obvious why this sort of moral and political conservativism makes this alternative not very promising.

5 Someone might try to offer a methodological reading of the first two sentences of this passage as descriptions of how the contract device looks, as it were, from the inside. We saw that a methodological construal could claim that the contractors do not view themselves as discovering moral truth but as creating it. But the availability of this reading cannot save Kantian constructivism from anti-realism. For (a) the previous passage, the last sentence of this passage, and other passages (listed below) are straightforward statements of the anti-realist commitments of Kantian constructivism, and (b) as we have seen, a methodological construal cannot support Rawls’ defense of justice as fairness or explain his contrast between Kantian constructivism and intuitionism. Compare: ‘The search for reasonable grounds for reaching agreement rooted in our conception of ourselves and in our relation to society replaces the search for moral truth interpreted as fixed by a prior and independent order of objects and relations, whether natural or divine, an order apart from how we conceive of ourselves’ (1980: 519). Cf. Rawls 1980: 516, 537-8, 551-2, 564, 569.

6 After this paper was accepted for publication, Rawls published ‘Justice as Fairness: Political not Metaphysical’ Philosophy & Public Affairs 14 (1985) 223–51Google Scholar. In this paper, Rawls claims to eschew controversial philosophical and metaphysical claims and to defend his conception of justice as no more than a reasonable basis of agreement among members of a constitutional democracy. This claim (which itself seems to make controversial philosophical assumptions) runs counter to my metaphysical interpretation of Kantian constructivism and accords more closely with what I call the methodological interpretation. However, I don't think this need affect the merits or interest of my interpretation of the Dewey Lectures. First, Rawls concedes that his new paper may not be entirely consistent with his previous writings (1985: 224). Second, the ‘political’ interpretation which he offers there, like what I call the methodological interpretation of 1980, cannot, I think, adequately support his defense of justice as fairness or explain his contrast between Kantian constructivism and intuitionism. Finally, whether or not Rawls intends the metaphysical interpretation of Kantian constructivism, his writings suggest it and others have so understood them. Indeed, as I will now explain, his writings suggest an argument for constructivist anti-realism. For these reasons, it is worth exploring the metaphysical interpretation.

7 Rawls seems to recognize this distinction; see Rawls 1974: 17, 21; 1980: 534, 571. But he typically fails to mark or observe it.

8 See Locke, John An Essay Concerning Human UnderstandingCrossRefGoogle Scholar, ed. Nidditch, P.H. (New York: Oxford University Press 1975) II, xxvii.Google Scholar

9 Cf. Daniels, Norman ‘Moral Theory and the Plasticity of Persons,’ The Monist 62 (1979), 274.CrossRefGoogle Scholar

10 Cf. Rawls 1974: 17-20; 1980: 516-7, 535-6; 1982: 169, 180-1. Also see Daniels, ‘Moral Theory and the Plasticity of Persons’ and ‘Reflective Equilibrium and Archimedian Points,’ Canadian Journal of Philosophy 10 (1980) 83–103CrossRefGoogle Scholar and Scheffler, Samuel ‘Moral Skepticism and Ideals of the Person’ The Monist 62 (1979), 297CrossRefGoogle Scholar. Someone might resist the claim that Rawls assigns a uniquely important role in moral theory to ideals of the person by appeal to the public conception of justice in the justification of Rawls’ two principles of justice. ‘The model-conception of the well-ordered society,’ as well as ‘the model-conception of the moral person,’ helps to determine the selection of the two principles of justice (1980: 517, 537-8, 555). However, even if, contrary to fact, the model-conception of the well-ordered society were independent of the model-conception of the moral person, Rawls’ texts make clear the greater importance of the model-conception of the moral person (1971: 584; 1979; 6, 20; 1980: 516-17, 518, 520, 535-6, 547-52, 554, 559-60, 571; 1982: 172f.). Buchanan, Cf. Allen ‘Revisability and Rational Choice’ Canadian journal of Philosophy 5 (1975) 395–408Google Scholar; Scanlon, T.M. ‘Rawls’ Theory of Justice’Google Scholar in Daniels, N. ed., Reading Rawls (New York: Basic Books 1975), 171, 178–9Google Scholar; Daniels, ‘Moral Theory and the Plasticity of Persons’ and ‘Reflective Equ∼ibrium and Archimedian Points’; and Scheffler, 295Google Scholar. In fact, the model-conception of the well-ordered society is not independent of the model-conception of the moral person; the former is heavily influenced by the latter. The various features of the well-ordered society depend in important ways upon the Kantian ideal of persons as free, equal, rational, and socially cooperative (1979: 6, 20; 1980: 519-22, 543-7; 1982: 172).

11 Rawls 1971: 251-65, 584; 1979: 6-7, 19-20; 1980: 516, 519-22, 534-6, 547-52, 554, 559-60, 571; 1982: 169, 172, 180-1. For useful discussion of Rawls’ ideal of the person and how it supports his two principles of justice, see Buchanan; Scanlon; Scheffler; and Daniels, ‘Moral Theory and the Plasticity of Persons’ and ‘Reflective Equilibrium and Archimedian Points.’

12 Cf. Scheffler, 295-300.

13 Indeed, since Rawls’ argumentative strategy contains an argument for underdetermination, it supports a relativistic version of constructivism. Thus, Rawls’ argument for constructivism creates a tension with his official agnosticism in the Dewey Lectures between relativism and nonrelativism. Cf. Rawls 1980: 569-70.

14 See, e.g., Quine, W.V.O. Word and Object (Cambridge, MA: MIT Press 1960)Google Scholar, chapter 2 and ‘Ontological Relativity’ in Ontological Relativity and Other Essays (New York: Columbia University Press 1969).

15 Parfit, Reasons and Persons (New York: Oxford University Press 1984)Google Scholar, chapter 15. Cf. Daniels, ‘Moral Theory and the Plasticity of Persons,’ 267-9Google Scholar.

16 See, e.g., Hare, R.M. ‘Rawls’ Theory of Justice’Google Scholar reprinted in Daniels, ed. Reading Rawls and Moral Thinking (New York: Oxford University Press 1981), 12, 40Google Scholar; Dworkin, Ronald ‘The Original Position’Google Scholar reprinted in , Daniels ed., Reading Rawls;Google Scholar Singer, Peter ‘Sidgwick and Reflectjve Equilibrium,’ The Monist 57 (1974) 490–517CrossRefGoogle Scholar; Lyons, David ‘The Nature and Soundness of Contract and Coherence Arguments’Google Scholar in , Daniels ed., Reading Rawls;Google Scholar and Brandt, Richard A Theory of the Good and the Right (New York: Oxford University Press 1979), 16–23Google Scholar. Actually, unlike the others, Dworkin embraces, rather than criticizes, the constructivist implications of the epistemological argument.

17 Here one might compare Rawls 1975 and 1980 with Dworkin, ‘The Original Position.’

18 See my Moral Realism and the Foundations of Ethics, chapters 2, 5, and 6 and appendix 2 for fuller discussion of these issues.

19 Cf.Lawrence, Bonjour ‘The Coherence Theory of Empirical Knowledge,’ Philosophical Studies 30 (1976) 281–312Google Scholar and Williams, Michael ‘Coherence, Justification, and Truth,’ Review of Metaphysics 34 (1980) 243–72Google Scholar.

20 I would like to thank Bennett, John G. Irwin, T.H. Sidelle, Alan Sturgeon, Nicholas Whiting, JenniferGoogle Scholar, audiences at Cornell University and Bates College, and a referee for the Canadian Journal of Philosophy for helpful comments on earlier versions of this paper.