1 See, for example, Rawls, John A Theory of Justice (Cambridge, MA: Harvard University Press 1971)Google Scholar; and ‘kantian Constructivism in Moral Theory,’ Journal of Philosophy 78 (1980) 515-72.

2 See, for example, Resnik, Michael ‘Logic: Normative or Descriptive? The Ethics of Belief or A Branch of Psychology?’ Philosophy of Science 52 (1985) 221-38CrossRefGoogle Scholar; and Sosa, Ernest ‘Equilibrium in Coherence?’ in Bender, J.W. ed., The Current State of the Coherence Theory (Kluwer: Academic Press 1989) 242-50CrossRefGoogle Scholar.

3 Norman Daniels has written much to help clarify Rawls’s notion of reflective equilibrium. See his ‘Reflective Equilibrium and Archimedean Points,’ Canadian Journal of Philosophy 10 (1980) 83-103, and ‘Reflective Equilibrium and Theory Acceptance in Ethics,’ Journal of Philosophy 76 (1979) 264-73. As an early expositor of Rawls on reflective equilibrium, Daniels has greatly influenced the discussion of this concept. In the following exposition, I stick primarily to Rawls’s text itself, yet my interpretation is no doubt influenced by that of Daniels.

4 A Theory of Justice, 20

5 In ‘Kantian Contructivism in Moral Theory,’ Rawls suggests that the justification of ethical principles is not so much to show that they are true, as with justification in epistemology, but to show that they are reasonable for persons living in a free and democratic society. Given this purpose it is not entirely clear that one could move radically away from the common presumptions we share and still justify principles for all who share those presumptions. I will return to this issue below.

6 Norman Daniels, in ‘Reflective Equilibrium and Archimedean Points,’ is particularly helpful in clarifying the relevance of these broader theoretical issues to reflective equilibrium. Daniels clarifies Rawls’s distinction between ‘narrow reflective equilibrium’ and ‘wide reflective equilibrium.’ See, however, Holmgren’s, Margaret ‘The Wide and Narrow of Reflective Equilibrium,’ Canadian Journal of Philosophy 19 (1989) 43-60CrossRefGoogle Scholar, for an argument that Daniels overdraws this distinction.

7 ‘We begin by describing [the original position] so that it represents generally shared and preferably weak conditions. We then see if these conditions are strong enough to yield a significant set of principles. If not, we look for further premises equally reasonable. But if so, and these principles match our considered convictions of justice, then so far well and good’ (A Theory of Justice, 20).

8 See A Theory of Justice, 21, 578.

9 Rawls himself suggests this through his use of the term ‘coherent’ to describe beliefs in reflective equilibrium. Norman Daniels seems to suggest that (wide) reflective equilibrium is a coherentist method, although he most explicitly contrasts it with a strong version of foundationalism represented by moral intuitionism (‘Reflective Equilibrium and Archimedean Points,’ 100-3; and ‘Wide Reflective Equilibrium and Theory Acceptance in Ethics,’ 264-73). More recently, Kai Nielsen presents his own modification of the concept of wide reflective equilibrium as ‘a coherentist model of justification’ (’Reflective Equilibrium and the Transformation of Philosophy,’ Metaphilosophy 20 [1989], 240). Similarly, Mark Timmons has called reflective equilibrium ‘the most popular version of moral coherentism’ (’On the Epistemic Status of Considered Moral Judgments,’ Southern Journal of Philosophy 29 [1990] Supplement).

10 DePaul, Michael ‘Two Conceptions of Coherence Methods in Ethics,’ Mind 96 (1987) 463-81CrossRefGoogle Scholar

11 This seems similar to the way Daniels has exposited the notion of ‘narrow reflective equilibrium’ in contrast to ‘wide reflective equilibrium,’ although there the focus is not merely on the relative stability of the initial considered moral judgments but on the kind of judgments which are relevant. The two issues are related. Narrow reflective equilibrium tends to be conservative since all that is brought into balance are the considered moral judgments and moral principles. Wide reflective equilibrium introduces background theories and other beliefs and thus provides many more factors which could force revision of the initial considered moral judgments. Thus it tends to be more radical.

12 Although I am not entirely convinced that Rawls has room for the radical extreme suggested by this conception of reflective equilibrium, some proponents are quite open to it.

13 Normal Daniels places significant weight on these two themes in taking wide reflective equilibrium to be a coherentist approach.

14 In ‘Reflective Equilibrium and Foundationalism,’ American Philosophical Quarterly 23 (1986) 59-69, Michael DePaul concludes that reflective equilibrium is consistent with some forms of foundationalism. DePaul argues that Rawls’s method of reflective equilibrium is compatible with foundationalism because the two are really positions on different, but related topics. Reflective equilibrium is ‘a method of theory construction,’ while foundationalism is ‘a type of account of the epistemic status of our beliefs’ (68). I find his argument well-reasoned and persuasive. His conclusion alone would undermine the thesis that reflective equilibrium provides a coherentist model of justification. The conclusion I am defending, however, is even stronger than DePaul’s. I will argue below that the method of reflective equilibrium, at least when taken as a method of seeking justification, is best construed as involving a form of modest foundationalism.

15 We have already seen that Rawls rejects the traditional ‘intuitionist’ interpretation of moral judgments. In seeing justification in the practical way he does, Rawls makes another move that takes him away from classical ‘intuitionist’ foundationalism, which would picture the task of justification as the epistemological one of having reasons for believing that certain principles are objectively true. A careful look at Rawls’s constructivist alternative would take us away from our theme in this paper. For a critical look at several constructivist proposals, including Rawls’s, see Margaret Holmgren, ‘Wide Reflective Equilibrium and Objective Moral Truth,’ Metaphilosophy 18 (1987) 108-24.

16 Although I will not develop it here, it would be possible to argue that presumptions must play the role of foundations even if we grant that the ‘common presumptions’ could radically change through a kind of communal reflective process. The argument would take a form parallel to that taken in considering the possibility of radical change in considered moral judgments. But the possibility of radical change is much less plausible here than in the case of moral judgments, since the whole point seems to be the articulation of principles for persons who see things as we now do in our free and democratic society. If we were to change our common presumptions radically, we would no longer see things in this way. Thus the principles would not be those which are reasonable for persons with the presumptions we share in our free and democratic society.

17 It is not my intention to suggest that these three directions are exhaustive or even mutually exclusive. The following explorations are meant to merely suggest some possible ways of accounting for the status of considered moral judgments, not to map all the possible approaches.

18 There are, of course, some who defend ‘epistemic conservatism’ according to which the fact that we believe a proposition does increase our justification for beliefs based on that proposition. For a critical discussion of this view, see Foley, Richard ‘Epistemic Conservatism,’ Philosophical Studies 43 (1983) 165-82CrossRefGoogle Scholar.

19 See, for example, ‘Kantian Constructivism in Moral Theory,’ 519.

20 I will leave a discussion of whether these objections can be overcome to a future occasion. For a further criticism of constructivist approaches to ethics see Margaret Holmgren, ‘Wide Reflective Equilibrium and Objective Moral Truth.’

21 For one naturalist moral realist view of the supervenience of moral on non-moral properties see Railton, Peter ‘Moral Realism,’ The Philosophical Review 95 (1986) 163-207CrossRefGoogle Scholar.

22 For arguments in the defense of rationalist approaches to ethics see Audi, Robert ‘Moral Epistemology and the Supervenience of Ethical Concepts,’ Southern Journal of Philosophy 29 Supplement (1990) 1-24CrossRefGoogle Scholar; and Simon, Caroline ‘On Defending a Moral Synthetic A Priori,’ Southern Journal of Philosophy 26 (1988) 217-34CrossRefGoogle Scholar. In general, a number of philosophers have published recent articles exploring possibilities in the modest objectivist foundationalist direction. Mark Timmons, in ‘Foundationalism and the Structure of Ethical Justification,’ argues that there are ‘a plethora of foundationalist views of justification that are at least worthy of consideration’ (Ethics 97 [1987]595-609, at 595). Margaret Holmgren has also published several articles which present helpful arguments in this direction. See, especially, ‘Wide Reflective Equilibrium and Objective Moral Truth’; and ‘The Poverty of Naturalistic Moral Realism: Comments on Timmons,’ Southern Journal of Philosophy 29 Supplement (1990) 131-35. See also Simon, Caroline J. ‘The Intuitionist Argument,’ Southern Journal of Philosophy 28 (1990) 91-114CrossRefGoogle Scholar. Both Holmgren and Simon argue that elements of the ‘intuitionist’ view should be taken more seriously than they have been in the contemporary discussion.

23 Along these lines, Copp, David in ‘Normativity and the Very Idea of Moral Epistemology’ (Southern Journal of Philosophy 29 Supplement [1990)189-210)CrossRefGoogle Scholar suggests that normative propositions are justified epistemologically in virtue of their relationship to moral standards, but that these moral standards are justified practically.

24 I am grateful to Robert Audi, Albert Casullo, and Joseph Mendola for helpful comments on an earlier version of this paper, and to Peter Vallentyne for his comments at the 1990 AP A Eastern Division Meeting. I have also benefited greatly from suggestions made by the referees of this journal.