20th WCP: Two (Faulty) Responses to the Challenge of Amoralism

To the question, "Why should I be moral?" there is a simple answer (SA) that some philosophers find tempting. There is also a response, common enough to be dubbed the standard response (SR), to the simple answer. In what follows I show that SA and SR are unsatisfactory; they share a serious defect.

I will interpret "Why should I be moral?" to mean "Why should I habitually perform the outward deeds prescribed by morality? Why, when I’m tempted to cheat or steal, ignore the sufferings of others, or renege on my commitments, should I do what morality calls for, and hence refrain from cheating and stealing, relieve the sufferings of others, and honor my commitments? Why should I go in for such things when so many other lifestyles are possible — for instance, that of a Gauguin or of a master criminal?" Perhaps the question has other meanings, but this is a natural one, and one to which SA and SR are meant to apply.

Interpreting the question this way removes some unclarity from the phrase "be moral." But it removes no ambiguities that might stem from the word "should." SA and SR purport to do this.

SA, briefly put, is this: "Why should I be moral?" is either a request for a moral reason to be moral or a request for another type of reason (or perhaps a motive) to be moral. In the first case it is absurd; in the second it is unreasonable or in some other way illegitimate.

What follows is a representative way of fleshing out SA:

(1) If someone — Alf, say — asks, "Why should I be moral?" his question is (a) a request for a moral reason to be moral, (b) a request for a nonmoral reason to be moral, or (c) a request for just any reason to be moral. Or perhaps it is (d) a request for a motive to be moral.

(2) If Alf’s question is of type (a) or (c), it’s silly. Nothing could be more obvious than that Alf has a moral reason to be moral. It’s a plain fact that being moral is the moral thing to do — it’s the option justified from the moral point of view — and this is clearly a reason for Alf to be moral.

(3) If Alf’s question is of type (b) or (d) it’s unreasonable; hence we need not bother with it. Moral philosophy aims at rational persuasion, not at generating motives to act. Nor does it aim to justify morality in terms of prudence, law, custom, or etiquette. Indeed, such a justification is impossible. But we should not be disturbed about this. The moral life is not called into question by showing that it does not reduce to a concern for custom, prudence, etc., any more than prudence is called into question by showing that it does not reduce to a concern for morality.

(4) Thus, Alf’s question is silly or unreasonable. We need not bother with it.

SR can be set out as follows:

(I) Responses to "Why should I be moral?" that treat the question as illegitimate, SA being one such response, fail to envision all the things the question might express. Very likely, Alf’s question is not of kind (a), (b), (c) or (d). Instead, it’s (e) a request for a justification for giving precedence to moral reasons for action when such reasons conflict with other reasons (e.g., those of self-interest). In other words, Alf is well aware that he has moral reasons to be moral, and that morality cannot be justified by appealing to prudence, law, and so on. His question is this: Whenever moral reasons recommend one course of action and nonmoral reasons recommend another, why should I favor the first set of reasons? What argument could persuade me to do this? This question is neither silly nor unreasonable.

(II) Thus, we must reject the conclusion of SA. Alf’s question is neither silly nor unreasonable. Moral theorists should take it seriously.

At this point, proponents of SR part company, some arguing that Alf’s demand can be met; others arguing that it cannot, that Alf must simply decide between the moral life and its alternatives. But I am not concerned with these arguments; I’m concerned with the first component of SR. My objection to it is this: I see no need to think that Alf’s question is anything other than what it seems to be: a request for a reason — a reason for Alf — to be moral. Indeed, it seems that the move made in SR, that of interpreting Alf’s question as something more complicated than a request for a reason, is simply an ad hoc maneuver designed to resist the implausible conclusion of SA — the conclusion, namely, that Alf’s question is silly or unreasonable. That this conclusion should be resisted I agree entirely. But I think we can find a better way to resist it.

We can do so by challenging premise (2) of SA. For convenience, let’s let M stand for being moral, and F stand for the fact that M is the moral thing for Alf to do. According to premise (2), F is clearly a reason for Alf to do M.

This premise is false, owing to the word "clearly." If anything is clear, it’s the following: First, F is a moral consideration favoring Alf’s performance of M. That is, F is a consideration we might cite in support of the judgment, "It would be wrong of Alf not to do M." (Other moral considerations include: "By not doing X Alf will cause people to suffer"; and "Alf expects others to do Y, particularly in their dealings with him.") Second, most people have desires and interests such that, were they in Alf’s shoes, they would find F to be more than a moral consideration. They would find it to be a reason, if only a prima facie one, for them to do M. Third, we tend to assume that Alf is like most people in the respect just mentioned, and hence that he has, in F, a reason to do M. Indeed, we have a pretheoretical "intuition" that moral considerations count for everyone as reasons for action.

Not one of these facts shows that F is a reason for Alf to do M. For example, the third one fails to do so because even if we can sometimes rest a philosophical conclusion on a single intuition, we cannot do so if we have reason to think that the intuition, far from being a preanalytical insight into the truth, is merely a bias arising from our moral education or from the infrequency with which we encounter amoralists, meaning people for whom morality has absolutely no moving appeal. We have been taught from childhood to treat moral considerations as reasons for action. And we know that for the people we typically encounter each day, moral considerations are reasons for action. These facts suffice to explain our intuition that F is a reason for everyone, including Alf, to behave morally. But they do not guarantee that our intuition is correct. Perhaps Alf’s aims, desires, and psychological make-up are so atypical that he can be aware of F without having even a prima facie requirement of rationality to do M. If so, then F is not a reason for Alf to do M. A reason for Alf to do M, whatever else it is, is something that provides Alf with a prima facie requirement of rationality to do M. More precisely, if F is a genuine reason for Alf to do M, then other things being equal, the rational thing for Alf to do is to act on F and do M. If he fails to do so, he is being irrational (minimally irrational even if not grossly so).

We now can see the thought that lies behind Alf’s question. Alf is aware of F, and he knows that F counts as a moral consideration. He also knows that most people have desires and aims through which moral considerations can rationally motivate them, and hence that for most people, moral considerations are reasons for action. And he knows that owing to the latter fact and to many others, we tend to think that moral considerations are reasons for everyone, including him, to act. Indeed, he knows that this tendency is so strong that we habitually use "moral reasons" and "moral considerations" interchangeably.

But Alf also knows, perhaps from reading Philippa Foot, that "to say that moral considerations are called reasons is blatantly to ignore the problem." The problem, as Alf sees it, is that he differs from other people. He has no goals or desires, or anything else, through which moral considerations clearly can gain a rational purchase on his behavior. For him, moral considerations are not obviously reasons to act. Prudential considerations are, but he is not sure that they give him any reason to do M. So in asking, "Why should I be moral?" he is asking that we provide a reason — a reason of any kind — for him to be moral. (Indeed, in asking the question he is likely to put emphasis on the word "I.") If we answer by citing a moral consideration, Alf will challenge us to show that his failure to be influenced by the consideration is a form of irrationality — that is, that such failure somehow involves a logical contradiction, a failure to fit means to ends, a circularity in his preferences, or something that clearly counts as a failure of rationality. Until he sees something of that kind in his indifference to the consideration, he will question whether the consideration is a reason for him to act.

We now can identify the flaw in SA, and we can see that SR shares it. SA takes for granted that Alf has at least one reason, a moral one, to do M. It treats as trustworthy our intuition that moral considerations count for everyone as prima facie reasons for action. This is to beg the question against Alf, who distrusts our intuition, and sincerely wonders whether he has any reason to do M. Similarly, SR takes for granted that Alf has moral reasons to do M, and that his puzzle concerns cases in which moral reasons conflict with nonmoral reasons. But Alf’s puzzle is of a different sort. He wonders whether he has reasons of any kind to do M. He questions whether those facts the proponents of SR so confidently dub "moral reasons for Alf to act" are indeed just that: authentic reasons for him to act.

Before continuing I will summarize my position: Contrary to SR, Alf is requesting a reason (of any kind) to do M. Contrary to SA, his request is not silly, for it is not obvious that he already has a reason to do M. Given the latter fact, SA and SR each beg the question against Alf, for they assume from the start that moral considerations count for everyone as reasons to do M. The latter assumption derives from a strong intuition, but this settles nothing. There are plausible explanations for the presence of the intuition, many of them consistent with a distrust of it.

These results are significant for two reasons. First, many treatments of the question "Why should I be moral?" share the assumption that moral considerations are reasons for Alf to do M. The error we have identified is not rare. It infects much of the literature on "Why be moral?" Second, and more generally, many philosophers, including many who reject ethical rationalism, continue to assume without argument that moral reasons always count as reasons for acting. Perhaps their thought is that although rationalists have failed to vindicate our intuition that moral considerations are overriding reasons to act, this does not undermine our intuition that moral considerations are always prima facie reasons to act. This is true but beside the point. The point is that in one respect, both intuitions are on the same footing: unless they are supported with arguments, Alf’s question is not silly.

Although I claim significance for these conclusions, I claim no great originality for them. Indeed, most everything I have said about reasons can be found in, or easily extracted from, the "middle" essays of Philippa Foot. This would diminish the importance of this paper were it not for the fact that Foot’s essays, although highly influential, still await the full impact they deserve, especially in the literature on "Why be Moral?" Indeed, one of my central points is that this literature cannot afford to ignore, as it so often does, Foot’s arguments. Foot shows that if the arguments of Kant, Nagel, and other ethical rationalists fail, we have no business taking for granted that moral considerations furnish everyone with reasons to act. The fact that our linguistic practices presuppose such a view of moral considerations is insufficient to show that the presupposed view is correct. Although philosophers routinely cite Foot’s work with respect, they often fail to take it to heart. They continue to assume without argument that Alf clearly has moral reasons — prima facie ones, anyway — to do as morality bids; hence that when Alf asks "Why should I be moral?" he is either confused or in search of something other than moral reasons.

In the short time remaining, I will address a predictable objection. It asserts that I have ignored the ambiguity in the term "reason for Alf to do X." For one sense of that term, the sense in which having a reason involves having a rational requirement, Alf can plausibly doubt that he has a reason to be moral. But for another sense of the term, the sense in which Alf’s having a reason simply amounts to there being some behavioral norms — moral, rational, or what have you — that apply to Alf, he cannot deny that he has a reason to be moral.

This objection fails for two reasons. First, it is not plausible. It’s an example of the all too common practice of postulating, rather than detecting, different "senses" of a term. The term "reason for Alf to do X" differs in meaning from many related items, including "Alf’s reason for doing X," "Alf’s motive for doing X," and "reason for thinking that Alf ought to do X." Perhaps it is sometimes used, imprecisely, in place of one of these items. But once we distinguish it from them we have no grounds for making a further distinction between different senses of the term itself.

Second, and more important, if we assume that "reason for Alf to do X" is ambiguous in the above way, it’s only the first sense of the term — the sense in which having a reason involves having a rational requirement — that is relevant to Alf’s question. When Alf asks "Why should I be moral?" he is rationally deliberating about what to do, and we can meet his request only by giving him something that can enter into his deliberations and influence him, insofar as he is rational, to do one thing rather than another. If we instead give him a "reason" which, even in the absence of any competing reasons and hindrances to acting, he can regard with indifference without being the least bit irrational — that is, if we give him a reason in the second sense of "reason for Alf to do X" — then we have given him something useless, something that does not favor one action more than another.

So the objection fails. But perhaps some will revise it to say this: The term "reason for Alf to do X" sometimes refers solely to agent-centered reasons — reasons that are tied to Alf’s desires or interests. On other occasions it refers to moral, and hence non-agent-centered, reasons. Alf cannot deny that he has, in the second sense of the term, a reason to do M. However, nothing has been said here that would divorce reasons in the second sense from rational requirements; hence nothing has been said that would make such reasons irrelevant to Alf’s question.

This objection is no better than the first. To say that Alf has reasons to do M, in the second sense of "reasons," and to assume that such reasons are tied to rational requirements, is to assume (in a roundabout way) the very thing Alf questions: that moral considerations furnish him with rational requirements to behave morally. Also, the objection runs afoul of an earlier point. There is no reason to think that "reason for Alf to do X" has two senses. Perhaps there are two kinds of reasons for Alf to do X — namely, agent-centered reasons and moral reasons. But if so, this needs to be shown, not assumed. In particular, it needs to be shown to Alf that he has non-agent-centered reasons, not merely agent-centered ones, and that some of the former direct him to do M. But even if this is shown, it does not reveal that "reason for Alf to do X" has two senses. After all, there are two different kinds of butter, salted and unsalted, but this does not show that "butter" has two senses.

In conclusion, SA goes wrong in treating Alf’s question as silly. SR goes wrong in construing Alf’s question as something other than a request for reasons to be moral. Both responses, and many others, go wrong by begging the question against Alf, who has legitimate doubts about whether he has any reason, moral or nonmoral, to do as morality bids.

Notes

(1) For a pristine example of SA see Richard L. Purtill, Thinking About Ethics (Englewood Cliffs, NJ: Prentice-Hall, 1976), 125. See also Peter A. French, The Scope of Morality (Minneapolis: University of Minnesota Press, 1979), 156f; and Robert Arrington, "Ethics (1945 to the Present)," in Routledge History of Philosophy, Volume X: Philosophy of Meaning, Knowledge and Value in the Twentieth Century, ed. J. V. Canfield (London: Routledge, 1997), 182.

(2) For versions of SR see Paul W. Taylor, Principles of Ethics (Belmont, CA: Wadsworth, 1975), chap. 9; Kurt Baier, "Moral Reasons and Reasons to be Moral," in Values and Morals, ed. A. I. Goldman and J. Kim (Dordrecht: D. Reidel, 1978), 231–56; and Kai Nielsen, "Why Should I be Moral? Revisited," American Philosophical Quarterly 21 (1984): 81–91.

(3) For the first sort of argument see Baier, "Moral Reasons and Reasons to be Moral"; for the second see Taylor, Principles of Ethics, 222ff.

(4) A similar objection can be made against premise (1) of QE. For example, there is no reason to think that Alf’s question is sometimes a request for a motive.

(5) Philippa Foot, Virtues and Vices (Berkeley and Los Angeles: University of California Press, 1978), 168 n. 8.

(6) E.g., it’s made by all the authors in note 2; also by John Hospers, Human Conduct, 2d ed. (New York: Harcourt Brace Jovanovich), 29f; James P. Sterba, "Introduction" and "Justifying Morality: The Right and the Wrong Ways," in Contemporary Ethics, ed. J. Sterba (Englewood Cliffs, NJ: Prentice-Hall, 1989), 1–20, 138–54; and Bruce Russell, "Two Forms of Ethical Skepticism," in Ethical Theory, Classical and Contemporary Readings, 2d ed., ed. L. Pojman (Belmont CA: Wadsworth, 1995), 525.

(7) Especially "Reasons for Action and Desires" (1972) and "Morality as a System of Hypothetical Imperatives" (1972), in Virtues and Vices. I follow Gavin Lawrence in dividing Foot’s work into three different phases. The essays I have cited are part of the middle phase. (See Gavin Lawrence, "The Rationality of Morality," in Virtues and Reasons: Philippa Foot and Moral Theory, ed. R. Hursthouse, G. Lawrence, and W. Quinn [Oxford: Clarendon Press, 1995], 89–147.) It’s well know that Foot’s recent work, most of which has not yet been published, runs counter to much of her middle work. But this does not affect the results of this paper. The challenges issued by Middle Foot must be taken seriously, both by Recent Foot (as she well knows) and by all philosophers concerned with "Why should I be moral?"

(8) An example is Sterba (op. cit., 7ff), who sums up Foot’s position on one page, indicates on the next that he intends to contest it, but then, a page later, assumes without argument that altruistic considerations provide everyone with prima facie reasons to act. Understandably, he then treats "Why should I be moral?" as something more complicated than a request for a reason. The trouble is that Sterba’s "altruistic reasons" are among the things Foot calls moral considerations. Thus, he has not engaged Foot’s argument; he has made exactly the assumption her argument challenges.

(9) A similar objection has been used against Foot. See Robert L. Holmes, "Is Morality a System of Hypothetical Imperatives?" Analysis 34 (1973): 96–100. Foot’s reply to it, which differs from mine, is in "‘Is Morality a System of Hypothetical Imperatives?’ A Reply to Mr. Holmes," Analysis 35 (1974): 53–56.

(10) I discuss these and related distinctions in "Motivation and Practical Reasons," Erkenntnis 47 (1997): 105–27.