1 Williams, Ethics and the Limits of Philosophy (Cambridge, MA: Harvard University Press 1985), 22; some emphasis added.Google Scholar

2 Republic. Trans. Grube, G.M.A. (Indianapolis, IN: Hackett 1974), 358cGoogle Scholar

3 One usually hears this sort of project being called rational justification. I use the term 'rationalist’ simply for the following reason: I shall, later in this paper, suggest that we should consider a sort of justification which is not rationalistic. I do not want to call this sort of justification anything like ‘non-rational’ justification, to contrast it with ‘rational’ justification. Such a name may (wrongly) suggest that the procedure of justification in question does not follow canons of rationality. Of course, this need not be the case; Hume was quite rational in arguing that rationality is not itself the basis for morality.

4 Thus Plato and Aristotle are not rationalists, although they do argue that we are rationally required to be moral. Their arguments rely not only on the claim that we are rational, but also upon some substantive, non-instrumental (perhaps morally loaded) account of the good.

5 Donagan, Alan The Theory of Morality (Chicago, IL: University of Chicago Press 1977) 4, 6CrossRefGoogle Scholar; emphasis added. Donagan makes his own philosophical prejudices apparent here when he attributes the first rationalist, and hence the first reasonably clear, conception of morality to the Stoics.

6 Whether or not Rawls's name should be on this list is a matter of some controversy. However, there is a reading, which I think is quite plausible, of Rawls's, A Theory of Justice (Cambridge, MA: Harvard University Press 1971)Google Scholar (not his more recent work) which makes this addition to the list quite appropriate.

7 Hence the title of this paper. The Foole is, in an important respect, the paradigm amoralist for rationalism. Rationalists must believe that amoralists are fools— i.e., that they can maintain their amoralism only by persisting in a mistake, acting contrary to reason.

8 And here we should remember those lines from Ogden Nash (made famous for some of us by Gauthier, David: ‘Oh Duty! / Why has thou not the visage of a sweetie or a cutie?’ (‘Kind of an Ode to Duty’ in I Wouldn't Have Missed It: Selected Poems of Ogden Nash [Boston, MA: Little, Brown 1975), 141])Google Scholar; see Morals by Agreement (New York: Oxford University Press 1986), 1.

9 One which is not based on false beliefs.

10 Gauthier, ‘Why Contractarianism?’ in Contractarianism and Rational Choice, Vallentyne, Peter ed. (New York: Cambridge University Press 1991), 17–18Google Scholar; emphasis added.

11 Gauthier, Morals by Agreement, 1Google Scholar. Cf. his ‘Why Contractarianism?’ in which Gauthier says morality ‘perishes’ if we cannot answer the amoralist (e.g., 16, 19). Also, cf. Nagel, The Possibility of Altruism (Princeton, NJ: Princeton University Press 1970), e.g., 4, 6.Google Scholar

12 For the sake of argument, I shall not question the assumption that rationality is essential to agents in this paper. However, it should be noted that one could undermine rationalism if one could argue that rationality is not essential to us.

13 The Possibility of Altruism, see esp. chs. 2, 3, 5, and 6.

14 Note that rationalists deny that their project presupposes anything about the content of morality. In his general characterization of meta-ethics, David Copp explains:

15 Here I simply describe what rationalist views are like — as I do throughout this discussion characterizing rationalism. I do not intend to prescribe what rationalism should be like, nor do I intend to endorse any of the assumptions I ascribe to rationalism.

16 Ethics and the Limits of Philosophy, 22-3

17 See Republic, 338a-b, d. Williams suggests that Plato provides us with a wonderful depiction of an amoralist with the character of Callicles in his Gorgias; see Ethics and till’ Limits of Philosophy, 22. Plato also brings the a moralist to life with the character of Thrasymachus.

18 See Republic, 340e-341b.

19 Sometimes in the literature the names ‘internalism’ and ‘externalism’ are given to views other than the ones I discuss here. Sometimes they refer to views about the relation between reasons and morality, sometimes to views about the relation between morality and motivating force. Sometimes (as when Bernard Williams uses them) they refer to non-instrumentalist and instrumentalist views of practical reason. When I use the terms in this paper, however, I mean to refer only to views about the relation between reasons and motivation (suitably characterized).

20 See Nagel, The Possibility of Altruism, 65-6Google Scholar. The list is his, and I don't believe it is meant to be exhaustive.

21 In the sense in which anything is of some color; I do not want to get bogged down by the difficulties associated with secondary qualities here.

22 One might object (as Shelly Kagan did) that I am begging the question against views like Kant's here. After all, if we have respect for the moral law (or for moral rights), then we will be motivated by Gewirth's demonstration that all agents have moral rights. However, I don't see any questions being begged. I merely say that sound inference does not necessarily preserve motivational efficacy, and surely that is true, whether or not we have respect for morality. If we do have respect for morality, then it is only the combination of sound inference to a moral truth and moral respect that gives us motivational efficacy. Sound inference to a non-moral truth and moral respect will not, and neither will sound inference to a moral truth without moral respect. So sound inference does not necessarily preserve motivational efficacy. And insofar as the success of this sort of justification relies on our having respect for morality (even if our having such respect isn't a part of the basis — i.e., the premises — of the justification) then, if the Kantian is to claim success for his justification, he must demonstrate that we all do essentially have such moral respect. I suggest that this tactic will not get the Kantian very far; after all, phenomenological evidence seems to suggest that we are not all motivated by this sort of respect. Now one might object (again, with Shelly Kagan) that phenomenological evidence cannot tell against our having respect, because respect is a capacity we all have, and its motivational efficacy comes into play only after we've sufficiently developed this capacity. However, I do not believe this will help the Kantian either. Now in order to demonstrate the success of his justification, the Kantian must demonstrate two things: (1) that we all essentially have the capacity of respect, and (2) that we are required to develop this capacity. If respect is not an essential capacity, then those who lack it can opt out of the justification. If we have no decisive motivating reason to develop the capacity, then respect needn't come into play motivationally. Therefore we will not necessarily be motivated by the justification, even if we do all essentially have moral respect (in the capacity sense).

23 Parfit, Reasons and Persons (Oxford: Oxford University Press 1984), 117–20Google Scholar

24 Gauthier, Morals by Agreement, ch. 6. See Parfit, Reasons and Persons, 17–23Google Scholar, for an interesting argument against the need to change policies.

25 This is not the argument Gauthier gives for the notion of rationality in question. However, we can imagine someone giving such an argument.

26 See Morals by Agreement, 170-7.

27 This may seem to be too extreme a characterization of instrumentalist rationalism. How can instrumentalists possibly avoid appeals to contingencies? This worry can take two forms. (1) The content of one's desires is a contingent matter. Since instrumentalists hold that it's rational to do what best satisfies one's preferences, don't they need to appeal to contingent preferences if they are to argue that it's rational to be moral? No. Instrumentalist rationalists may avoid appealing to the contingent contents of preferences in two ways. (a) They may rest their arguments on claims about the content of agents’ preferences without appealing to contingent contents. They may, after all, limit their discussion to preferences agents necessarily have — e.g., things a rational agent wants no matter what else she wants (as Rawls does in A Theory of Justice). Or (b) instrumentalist rationalists may avoid reliance upon claims about the content of agents’ preferences altogether. They may try to argue that, no matter what agents want, they best satisfy their preferences by adopting moral constraints on their behavior (as does Gauthier in Morals By Agreement). (2) Instrumentalist rationalists conceive of morality as a cooperative venture for mutual advantage. They acknowledge that such a cooperative venture would be either impossible or unnecessary unless certain background conditions — e.g., moderate scarcity of goods and asociality (a.k.a. mutual disinterest) of agents— obtained. But it is a contingent matter whether these conditions, often called the circumstances of justice, obtain. Doesn't this demonstrate that it's wrong to portray instrumentalist rationalists as refusing to appeal to contingencies? No: such appeals to contingencies do not violate the rationalist dictum that a justification of morality must not rely upon contingencies. As Rawls puts it, ‘Unless these circumstances existed there would be no occasion for the virtue of justice, just as in the absence of threats of injury to life and limb there would be no occasion for physical courage’ (A Theory of Justice, 128). We must claim that such contingent circumstances hold if the question of whether or not we should be moral is even to arise. But once the question does arise, the rationalist contends that we may not appeal to contingencies to answer it (for the reason spelled out in section [, 1 of this paper).

28 Richmond Campbell, correspondence

29 Or more precisely, we've seen no argument supporting this worry. But recall Jerry Postema's point discussed above. If considerations about the satisfaction of preferences an agent would have were she to meet certain conditions have no motivational force, then this may be an indication that agents can opt out of conceptions of instrumental rationality which require that these conditions be met. In this case, instrumentalist justifications of this type must be dealt with in the way I describe below for practical non-instrumentalist arguments.

30 I do not want to suggest here that I think that it is always inappropriate to embrace inconsistency. I also do not want to suggest that I believe that the normative force of deductive rationality is beyond question. I do not believe that the normative force of any conception of rationality is beyond question. If we are to demonstrate that deductive rationality has normative force, we must explain why we should reason validly. The answer, it seems to me, is that valid inference preserves truth. Of course, then we must answer the question of why we must care about truth. That is a question I imagine we can usually answer; hence, I feel fairly comfortable accepting (somewhat guardedly, as I do here) the claim that we cannot opt out of deductive rationality.

31 For example, Nagel seems to use such a criterion. See The Possibility of Altruism, 19.

32 This does not mean such an agent can have only self-interested preferences. An agent acts from her own preferences even if her preferences incline her to satisfy the preferences of others. So, to use Kurt Baier's terminology, such an agent needn't act from self-interested preferences, only self-anchored ones (see his ‘Moral Reasons and Reasons to be Moral’ in Values and Morals, Goldman, A.J. and Kim, J. eds. [Dordrecht, Holland: D. Reidel 1978), esp. 242]).CrossRefGoogle Scholar

33 Something like Jerry Postema's point applies again here. Some instrumentalist accounts place restrictions on the bearers of subjective, relative value. Of course, these accounts must show why these restrictions guarantee such value, rather than interfere with it.

34 I am grateful to Walt Edelberg for bringing this example to my attention.

35 Though even on this option we could still maintain that rationality always permits one to be moral, even if it doesn't require it.

36 Some would argue that rationalism is the only legitimate paradigm of justification in moral theory. I argue against this view in my ‘Justification Without Rationalism' (ms).

37 I'd like to thank Annette Baier, Kurt Baier, Alisa Carse, Ann Cudd, Ed Curley, Walter Edelberg, Neal Grossman, Dorothy Grover, Shelly Kagan, Richard Kraut, Geoff Sayre-McCord, Jerry Postema, and Jennifer Whiting for their helpful comments on earlier drafts of this paper. I owe special thanks to David Gauthier for his generous and patient philosophical guidance and moral support. I'd also like to thank the members of the Triangle Ethics Circle sponsored by the University of North Carolina at Chapel Hill and audiences at the universities where I've read versions of this paper for stimulating discussions.