by Philip L. Quinn

Abstract: This paper is a study of a pragmatic argument for belief in the existence of God constructed and criticized by Richard Gale. The argument's conclusion is that religious belief is morally permissible under certain circumstances. Gale contends that this moral permission is defeated in the circumstances in question both because it violates the principle of universalizability and because belief produces an evil that outweighs the good it promotes. My counterargument tries to show that neither of the reasons invoked by Gale suffices to defeat the moral permission established by the original argument.

Throughout his important book, On the Nature and Existence of God, Richard M. Gale dichotomizes.¹ The material covered by the book is first divided into atheological arguments and theological arguments. The theological arguments are then divided into epistemological arguments and pragmatic arguments. The pragmatic arguments are further divided into prudential arguments and moral arguments. And finally the moral arguments are divided into those that claim religious belief enables us to engage rationally in the practice of morality and those that claim religious belief makes us or our society morally better. In this paper, I examine Gale's discussion of the main argument he considers under the heading of moral arguments that claim religious belief makes believers and their societies better. He tries to show that this argument is flawed. I argue that he does not succeed in doing so.

The argument Gale considers under this heading is constructed out of materials he finds in William James's famous essay, "The Will to Believe." But Gale advances the discussion well beyond the point to which James brought it. Gale's precise and rigorous analysis of the argument he discusses renders its structure perspicuous. For that we are greatly indebted to him.

Gale's views on James's approach to natural theology have changed in the years since he published the material on which I shall focus. I am of course aware of his more recent treatment of it in The Divided Self of William James. I am convinced, however, that the critique of Jamesian natural theology I shall scrutinize has not yet received the attention it deserves. It remains philosophically interesting in its own right, even though Gale has now moved beyond it. Hence, given the space constraints under which I am operating, I have decided to concentrate exclusively on it.

According to Gale, James's moral argument has the following form:

6. Doing X helps to bring about Y;
7. It is morally desirable that Y; therefore,
8. One has a prima facie moral permission to do X (358).

Gale reckons that it is a "valid argument form" (371). However, as he notes, the prima facie moral permission granted in its conclusion is defeasible. Defeat will occur when, by doing X, one brings about some moral evil that outweighs the moral good realized in Y. So, the question to which Gale directs our attention is this: When does religious believing satisfy this condition of outweighing?

Like James, Gale considers the notorious answer proposed by W.K. Clifford. As Clifford sees it, defeat occurs whenever believing, religious or otherwise, occurs in the absence of sufficient evidence. Thus Clifford enunciates the following exceptionless prohibition on believing of this sort:

C: It is wrong always, everywhere, and for anyone to believe anything upon insufficient evidence (355).

Gale ridicules, quite rightly in my opinion, the plague theory of epistemically unwarranted belief with which Clifford supports C. This undercuts Clifford's rationale for C, but it does not show that C is false. And so, again like James, Gale endeavors to formulate a set of necessary and sufficient conditions for counter-examples to C or exceptions to a sensible prohibition on believing in the absence of sufficient epistemic evidence.

The set Gale comes up with defines what he calls a "special genuine option to self-induce a belief" (370). It has seven members. Some of them represent Gale's way of making explicit James's ideas of an option that is live, momentous, and forced. Others represent conditions he extracts from elsewhere in James's discussion. The condition that plays a crucial role in Gale's criticism is the last of them:

15. A knows that she will act so as to help make p true if and only if she believes in advance that p is true (370).

Gale also specifies the religious hypothesis that is to figure in the instantiation of his valid argument form. It is this:

R. Good will win out over evil in the long run (364).

He also indicates that the morally desirable state that believing this hypothesis helps to bring about is that the believer "acts in an altruistic or good-making fashion" (371). So, the instance of his valid argument form Gale would have us focus on is this:

(i) Believing R helps to bring about the believer's acting
altruistically.
(ii) It is morally desirable that the believer acts altruistically.
(iii) Therefore, the believer has a prima facie moral permission to believe R.

So far, so good.

Suppose now that, in the case of a particular believer, this argument is sound. She has a prima facie moral permission to believe R. Suppose also that she is, with respect to R, in the situation of having a special genuine option to self-induce a belief. And suppose she does self-induce the belief that R, even though she lacks sufficient epistemic support for this belief. Since she has the requisite special genuine option, she is a counter-example to Clifford's Principle, C. Hence her prima facie moral permission is not defeated on account of a violation of C. Nor is it defeated because the good she produces is outweighed by the imaginary evils projected by Clifford's ridiculous plague theory. So, is her belief that R in the clear, morally speaking, despite its lack of sufficient epistemic support?

Gale's answer to this question is a resounding no. He supports this negative conclusion with two arguments. According to the first, her permission is defeated because it violates the moral principle of universalizability. According to the second, it is defeated because she brings about an evil outweighing the good she promotes that will occur even though Clifford's plague theory is false. I shall examine and criticize both arguments.

In order to represent the fact that the believer in question induces the belief that R by nonrational means, Gale asks us to imagine that she takes a belief-in-R inducing pill. Appealing to this conceit, his argument from universalizability goes as follows:

Let us imagine two persons, A and A¹, who are exactly alike save for one feature of their psychology. A, being short of courage, will not act so as to help make R true unless she first believes that R will become true, whereas this is not true of A¹, the psychologically stronger member of the pair. It thereby turns out that A, but not A¹, is morally permitted to take in belief-in-R inducing pill. And this seems to violate the principle of universalizability. The reply is that their circumstances are not the same since A satisfies condition 15 while A¹ does not. But is this a morally relevant feature of the circumstances? I think not. It seems wrong to accord a moral privilege to someone but not to another on the grounds that the former is a psychologically weaker person (371).

What are we to make of the line of argument contained in this passage?

I think it is unsuccessful. I take it we are to suppose that A has a special genuine option to self-induce belief that R. Hence A is a counter-example to C, and so her prima facie permission to take the belief-in-R inducing pill is not defeated by Cliffordian considerations. We are also to suppose, at least for the sake of argument, that her prima facie permission is not defeated by any other considerations. Thus A is morally permitted to take the belief-in-R inducing pill. But why should we go along with Gale in further supposing that A¹ is not morally permitted to take the pill. To be sure, he does not need the pill; he will act so as to help make R true even if he does not take it. It does not follow from these facts, however, that A¹ is not morally permitted to take the pill. And Gale does not provide an explicit argument for his claim that A¹ is not morally permitted to do so.

There is, of course, an argument for this claim available to Gale. A¹ fails to satisfy condition (15) with respect to R, which is a necessary condition for having a special genuine option to self-induce belief that R. So, he is not a counter-example to principle C. He therefore is not an exception to the moral prohibition on believing in the absence of sufficient evidence. If he were to take the belief-in-R inducing pill, he would violate this prohibition. Hence his prima facie moral permission to take the pill is defeated. And thus A¹ is not morally permitted to take the pill.

But what this argument shows, as I see it, is that Gale is flatly mistaken in claiming that the difference in circumstances between A and A¹ is not morally relevant. As he sets up the argument, A and A¹ are supposed to be exactly alike except that A satisfies condition (15) while A¹ does not. And this difference alone is supposed to explain why A¹ remains bound by the prohibition on believing upon insufficient evidence while A is an exception to that prohibition. And a difference that by itself explains why one person remains bound by a moral rule while another does not is surely a morally relevant difference. Hence Gale's argument from universalizability fails.

There is a feature of Gale's example that may be an obstacle to appreciating this point. As he describes the situation, A is psychologically weaker than A¹, but he does not explain exactly why this is the case. Suppose A's relative psychological weakness is a result of having been brought up by domineering parents while A¹ was a well raised child. This way of filling out Gale's example casts serious doubt on his intuition that it is wrong to accord a moral privilege to A but not to A¹ on the grounds of A's psychological weakness. On this supposition, after all, the psychological weakness is not A's fault, and so a compensating privilege may be an appropriate way to level the moral playing field. However, Gale characterizes A's psychological weakness as a shortage of courage. So, suppose instead that A has culpably neglected to develop the virtue of courage while A¹ has responsibly built up a courageous character. This way of filling out the example supports Gale's intuition that it is wrong to accord a moral privilege to A but not to A¹ on account of the difference between them. But on this supposition the morally relevant difference between them is that A is vicious while A¹ is virtuous. Hence in the example thus construed A and A¹ are not exactly alike in morally relevant respects except that A satisfies condition 15 while A¹ does not.

An analogy may help to make my point vivid. I have a moral permission to kill a human being and so am an exception to the moral prohibition on homicide. You are not morally permitted to kill a human being and remain bound by the prohibition on homicide. What explains this difference between us? I am being attacked by a maniac and cannot save my own life unless I kill him. If I kill the maniac in self-defense, it will count as a case of justifiable homicide. You are under no such attack. Our circumstances are exactly alike except for this difference. Clearly it is a morally relevant difference in these circumstances.

Gale's second argument mobilizes an objection from personhood. His ingenious idea is to replace the falsified principle C with another exceptionless prohibition that will succeed in doing the work C was meant to do. The argument begins with Gale's firm conviction that "there is an absolute value to personhood" (372). On his view, this means the following:

P.It is always wrong to bring it about that a person becomes less than or less of a person or that a potential person becomes something less than a person (372).

The first disjunct of the principle is what does the work in Gale's argument. His fallback position, meant to appeal to readers unwilling to grant that personhood has absolute value, is that it at least has "a very great value" (372). But what, then, is a person?

According to Gale, having free will is necessary and sufficient for being a person. He does not, however, try to explicate directly the concept of having free will. Instead he claims that "to have free will is to behave as a morally responsible agent" (372). And he then goes on to explicate the notion of morally responsible agency in terms of a moral responsibility game, the playing of which is alleged to be necessary and sufficient for being a person. This strategy of explication strikes me as misguided because it seems to rule out the possibility of morally irresponsible persons. But perhaps it can be defended in terms of Gale's assumption of "a highly normative concept of personhood" (371). In any event, I am willing to go along with his strategy for the sake of argument.

Gale's moral responsibility game is defined by a lengthy list of a dozen rules. Some of them are, according to his terminology, "ontological." For purposes of understanding Gale's argument from personhood, the important ontological rules are these:

R₂.A player is morally responsible for an act only if he did it as a rational agent.
R₃.A player performs an action F as a rational agent only if: (a) he knows what he is doing; (b) he has good reasons for doing F; (c) his reasons are at least a necessary cause of his doing F; and (d) he has no reasons that are both necessary causes of his doing F and not good reasons for doing F (373).

In the course of commenting on these rules, Gale offers an explicit definition of the notion of good reasons employed in R₃. It goes as follows:

D: A reason r is a good reason for an agent A to do an action F just in case it is true both that A is justified in accepting r (even if r is in fact false) and r is logically relevant to his doing F (even if r is not the best reason) (374).

Gale does not provide an account of the concept of logical relevance in this principle. However, he does say that examples of reasons are "desires, wants, intentions, and beliefs" (375). And he offers as an example of good reasons for reaching for a glass of water "that I desire a drink and believe that there is a glass of water in front of me and that water quenches thirst" (374). We may thus work in this discussion with our ordinary intuitive notion of the relevance of reasons to the actions they rationalize.

What Gale calls the "sociological" (373) rules of his moral responsibility game are rather elaborate. But the only rule that is crucial for Gale's argument is this:

R₁₁.No player can opt out of the game (378).

With this background information at our command, we are now in a position to set forth the steps of the argument.

Consider someone, A, who takes the belief-in-R pill. Gale first establishes that "her belief in R constitutes part of her reasons for acting altruistically" (382). He then tries to show that her belief in R fails to satisfy both the conditions specified by D for being a good reason for her to act altruistically. In my opinion, this is the crux of the argument. I shall come back to it, since I intend to concentrate my critical fire on it. But let me first indicate how the argument proceeds to the conclusion Gale wishes to reach. Because the person who takes the pill does not have good reasons for acting altruistically, she "violates her personhood by taking the belief-in-R inducing pill, since she causes herself to perform actions for which she lacks good reasons" (382Ð83). If we take personhood to be an absolute value, we may say that she violates the exceptionless prohibition P. This violation defeats her prima facie moral permission to take the pill based on the good that will result from her doing so. According to Gale, we arrive at a similar conclusion, even if we only assume that personhood is a very great good. For, he claims, "given the very extensive nature of the actions and dispositions caused by taking the pill and the extent to which they are constitutive of A's character and personality, she in effect opts out of the moral responsibility game, which violates R₁₁" (383). She thereby makes herself less of, if not less than, a person. This is the evil that outweighs the good that will result from her taking the pill. And so, in this case too, her prima facie permission to take the pill is defeated.

What is Gale's argument for the claim that A's belief in R, acquired by taking the belief-in-R pill, fails to satisfy both conditions specified in D? It goes as follows:

First, A, ex hypothesis, lacks any epistemic justification for R. Second, and more important, R is logically irrelevant as a justification for her acting altruistically so as to make R true. It would be absurd to give as one's reason for acting so as to make a proposition true that it is in fact true or will turn out to be true. It would be crazy to work to bring about an economic depression because one believes that an economic depression will occur (382).

I shall respond to each of these points in turn.

It seems to me Gale's claim that A lacks epistemic justification for R is correct. However, this claim's truth does not establish the conclusion that the justification condition in D is not satisfied. For if we look closely at D, we will notice that it demands only that A be justified in accepting R, not that A's justification be epistemic. D's justification condition would be satisfied if A's justification for accepting R were moral or pragmatic in some other sense. In the present context, where precisely what is at issue is whether there can be moral justification for religious belief in the absence of epistemic justification, it would be question-begging to assume that A's justification for accepting R cannot be moral rather than epistemic. That is supposed to be the conclusion of Gale's line of argument, and so it cannot be among its premises. Hence, even though we should grant that A lacks epistemic justification for R, this concession falls short of establishing the conclusion Gale wishes to reach. He has not shown that A fails to satisfy the justification condition in D.

Initially anyway, Gale's second point appears to be more formidable. In my opinion, however, this appearance exists only because he has suppressed some subtle qualifications in stating it. Recall that the belief A is supposed to acquire by taking the pill is that good will win out over evil in the long run. This belief, in turn, is supposed to be, not A's reason, but part of A's reasons for acting altruistically. And acting altruistically is supposed to consist in acting, not so as to make it true that good will win out over evil in the long run, but so as to help make it true that good will win out over evil in the long run. So, the question is whether A's belief that good will win out over evil in the long run can be a logically relevant part of A's reasons for acting so as to help make it true that good will win out over evil in the long run. When the question is put this way, it seems clear that A's belief can indeed be a relevant part of A's reasons for acting. Believing that good will win out over evil in the long run but that her contribution is not essential to this outcome, A might nevertheless believe that it would only be fair for her to contribute to this outcome or that it would be shameful for her not to contribute and desire to act fairly or to avoid acting shamefully and thus be motivated to help make it true that good will win out over evil in the long run. But were she not to believe that good will win out over evil in the long run, she might then consider it neither unfair nor shameful not to make an effort and hence not be motivated by her desires to help make it true that good will win out over evil in the long run. In such circumstances, it seems obvious to me, A's belief that good will win out over evil in the long run is logically relevant, intuitively speaking, to her helping to bring about this outcome. So, Gale's argument that the logical relevance condition in D is not satisfied also fails.

Perhaps in this instance too an analogous example will serve to make my point more vivid. Suppose my parish has embarked on a fund drive to raise the money needed to build a new church. If I believe the fund drive will succeed whether or not I contribute and believe it would only be fair for me to do my bit to contribute to its success or would be shameful for me not to do so, then my desires to act fairly or to avoid acting shamefully may motivate me to do my bit. But if I believe the fund drive is doomed to failure whether or not I contribute and believe it would be pointless for me to make a contribution under these conditions, then my desires to act fairly or not to act shamefully may not move me to contribute. So my belief that the fund drive will succeed is, intuitively speaking, logically relevant to my making a contribution. And this example seems to be typical of a large class of cases in which one can make a contribution to the success of a collective enterprise but one's contribution is not essential to its success.

The upshot is this. Both Gale's argument from universalizability and his argument from personhood fail. Neither of them succeeds in showing that A's prima facie moral permission to take the belief-in-R inducing pill is defeated. They are his only arguments for this conclusion. Hence A's prima facie moral permission emerges from Gale's assault triumphantly undefeated. That being so, one question remains to be answered. Will Gale now come to believe that, by constructing a sophisticated Jamesian moral argument for religious belief, he has, like Dr. Frankenstein, inadvertently created a monster?

University of Notre Dame

Notes
1. Richard M. Gale, On the Nature and Existence of God (Cambridge: Cambridge University Press, 1991). Citations in the text refer to this book.

2. I presented a version of this material at a meeting of the Society of Humanist Philosophers devoted to Richard Gale's Philosophy of Religion in Philadelphia on December 28, 2002. Richard Gale was the respondent on that occasion. I am grateful to him and to members of the audience for stimulating discussion.