SCOTT F. AIKIN (Article in PDF format)


 On one side of the ethics of belief debates are the evidentialists, who hold that it is inappropriate to believe without sufficient evidence. On the other side are those who hold that there are legitimate cases of belief without evidence. William James argues in “The Will to Believe” that one may believe ahead of one’s evidence, and so the essay is posed as a defense of religious faith.

To begin, two interpretive stipulations. The first is that James’s preamble delineates the objectives of the essay, the second that James’s argument is what I call an ascending case.

“The Will to Believe” was originally an address to the Yale and Brown Philosophical Clubs. James contrasts the “good old orthodox college conversation” about religious matters at those two universities with the “freethinking … indifference” at his home institution, Harvard. This contrast is full of portent. On the one hand, the orthodox conversation extends to the theological matters bearing on salvation and divine attributes. On the other hand the Harvard students, once they’ve been imbued with “the logical spirit,” reject the “lawfulness” of religious faith and related intellectual matters. James promises to provide “a justification of faith,” a defense of our right to adopt a believing attitude in religious matters, in spite of the fact that our merely logical intellect may not have been coerced.” (WWJ: 717)[1]

A justification of faith, given James’s contrasts, must satisfy two desiderata. First, the commitments saved must be different from those held by Harvard freethinkers. They must be worthy of the appellation “good old orthodox college conversation.” Accordingly, let us call this the orthodoxy requirement.

The second desideratum of the justification of faith is that it defend faith from the objection that it is unlawful. James holds that a defense of faith must concede that there is insufficient evidential backing. So a case for lawful exceptions to the requirement of evidential backing must be made. James makes the case with the doctrine of the will-to-believe, that one may decide an intellectually undecidable matter by one’s passional nature (WWJ: 723). Let us call this the lawfulness requirement.

Any evaluation of James’s essay must answer whether and how James’s justification of faith meets both requirements. Thus is the first interpretive posit – the preamble provides desiderata of orthodoxy and lawfulness.

Second is the ascending case. James identifies the exceptions to the evidentialist rule as friendship, social coordination, and romantic conquest cases. They form a family on the basis of their being genuine, morally momentous, and doxastically efficacious examples of beliefs. They are lawful beliefs because they, in their circumstances, cannot be withheld and they are effective in bringing about their truths, which are preferred. The case ascends, then, to faith. Formally, an argument by analogy:

P1. Prima facie, the evidentialist rule (ER) is acceptable.

P2. Religious faith seems to be ruled out by ER.

P3. Friendship, coordination, and romance cases are lawful
exceptions to ER.

P4. These are momentous, moral and (what I will
call) doxastically efficacious cases, and those properties
explain their lawfulness.

P5. Religious faith is momentous, moral, and doxastically

C.  Therefore, religious faith (despite appearances in P2) is
an acceptable exception to ER.

The argument here depends on P4 and P5 – what doxastic efficacy is and whether it is appropriate to say that religious faith is doxastically efficacious.

Roughly, doxastic efficacy is a property of a belief (p) that if S believes that p, then p is thereby made more likely true. It is clear that this sort of pattern occurs with cases where confidence is a relevant factor (e.g., in making friends or performing a social function), but the question is whether this is appropriate to attribute to religious beliefs. Are religious beliefs the sort of beliefs that are made true by their being believed?

Combining the two interpretive posits occasions a dilemma for James: in short, the commitments constituting faith either (a) satisfy the lawfulness requirement by being doxastically efficacious, but therefore are unorthodox, or (b) the beliefs are orthodox and not doxastically efficacious, but are thereby not, given the argument, lawful. James’s justification of faith in “The Will to Believe,” then, is a failure, at least by the standards his preamble sets forth. That said, “The Will to Believe” is successful in providing a template for subsequent pragmatist accounts of doxastic efficacy and reconstructions of religious views in line with humanist commitments.


James’s justification of faith depends on showing that the assessments of faith as philosophically unlawful are incorrect—specifically, that they are lawful exceptions to those requirements. As such, James’s argument takes the form of a series of acceptable exceptions arranged to ascend from seemingly insignificant occasions of confidence to cases of religious belief. This pattern is James’ ascending case.

The rule for lawful belief James associates with the logical spirit is the evidentialist norm, captured starkly by the English mathematician and philosopher, W.K. Clifford:

[I]t is wrong, always, everywhere, and for anyone, to believe anything upon insufficient evidence (1999 [1877]: 175)[2]

It is important to note that James endorses the rule generally in calling it a “healthy” (WWJ: 721) attitude, so the case he will make against it is not to overturn it, but to show that there are legitimate exceptions. The exceptions, for the case to ascend in significance, must share a handful of features. The cases of belief James takes to be counter-examples to evidentialism are marked, first, by the fact that they are all what he calls “genuine options.” A genuine option is one that is living, forced, and momentous.

The exceptions, further, share a feature of what James identifies as being moral questions (WWJ: 729), that is, decisions wherein it certainly is best to pursue one option over another. Moral questions are those not of what exists, but rather of what would be better to exist.

Finally, with a subset of moral questions, how one believes with regard to the question actually is a determining factor in whether the fact in question comes to exist. That is, with some issues, which option you believe is true is the one more likely to be true. Such cases are doxastically efficacious. James surveys such moral questions:

  • Friendship: Whether or not you like me is a moral question, and it is a doxastically efficacious matter. If I believe you don’t, or I am not inclined to believe you do, I will “stand aloof” and “ten to one, your liking will never come.” Whereas, my antecedent faith in your liking me “is in such cases what makes your liking come” (WWJ: 730).

  • Social coordination cases: How we do our jobs, whether in an army, on athletic team, or in a commercial society, is a moral question and is doxastically efficacious. Our performances in our roles depend not only on the cooperation amongst many individuals in doing their own jobs, but on the shared antecedent trust that we all will continue to do so. Trust keeps it all running.

  • Romantic conquest cases: Whether to pursue a romantic interest is a momentous issue, one that, often, does not yield second chances. The same for marriage proposals. It is infrequent that one knows for sure what the answer will be, but the good things come only to those who make the uncertain proposals. James notes, “How many women’s hearts are vanquished by the mere sanguine insistence of some man that they must love him!” (WWJ 730)

Notice that each of the cases has doxastic efficacy, where believing that p makes p more likely and not believing that p (or believing that not-p) makes p less likely (and not-p more). James captures the point by noting that the “desire for a special sort of truth brings about that special truth’s existence” (736). So the moral cases James takes to be worth posing are those with preferred outcomes that are doxastically efficacious.

Given that these outcomes are good and that the antecedent beliefs contribute to their coming to be, James concludes:

There are, then, cases where a fact cannot come at all unless a preliminary faith exists in its coming. And where faith in a fact can help create a fact, that would be an insane logic would say that faith running ahead of scientific evidence is the ‘lowest kind of immorality’ into which a thinking being can fall. (WWJ: 731)

With doxastically efficacious propositions, evidentialism is inappropriate. They are lawful, and to prohibit them would be an “insane logic.”


The orthodoxy requirement is that what contents of faith that are saved are ones that would be different from Harvard freethinkers. As James turns to religious faith in the closing section of “The Will to Believe,” he acknowledges that there is a question whether religious faith is of the same class as friendship and coordination cases.

But now, it will be said, that these are all childish human cases, and have nothing to do with great cosmical matters, like the question of religious faith. (731)

James demurs. Religious commitments “differ so much in their accidents.” Consequently, we must find what binds them all together as religious commitments, which James identifies as “essentially two things”:

R1: The best things are the more eternal things, and
R2: We are better off believing R1 (from WWJ: 731-2)

Crucially, we are to meet the conjunction of R1 and R2 “halfway” in the same way we’d done in the friendship case. In so doing, we contribute to their truth. James notes:

The feeling, forced on us we know not whence, that by obstinately believing there are gods, . . . we are doing the universe the deepest service we can, seems part of the living essence of the religious hypothesis (WWJ: 733, emphasis added).

Note, first, that in holding the religious hypothesis true, we are doing the universe a service, we are contributing to something. Second, note the way the moral truths must be arranged. If the hypotheses are true, then, James reasons, it would be irrational to have a rule that prevents us from believing them, as, of course, rules of reasoning should not prevent us from believing important and valuable things that are true, but should be conducive of true beliefs.

The question here is whether the faith that James has saved as a doxastically efficacious moral question is recognizable as the orthodox faith announced in the preamble. Would the beliefs salvaged by this strategy be within the realm of the faithful discussion of divine omnipotence or the difference between justification and sanctification? James’s answer is yes, because he, again, holds that R1 and R2 are the essence of the religious hypothesis.

But are they? It certainly seems right that R1 and R2 are parts of the standard religious commitment, but are they essential to it?

Notice that with James’s identification of the essence of religion with this view about value, there is no mention of god, the divine, or the role of either in the world. Only “the more eternal.” R1 is diagnostic, in that James is identifying the essence of religious impulses as a move from recognizing that goods that last are preferred to positing an ontology that offers that permanence. The core of religion is the articulation of how what one values most can last.

The trouble with this diagnostic take on religion, certainly from the perspective of believers, is that it identifies the positive views within a religion as coping mechanisms in the face of contingency and uncertainty. Consequently, the identifying components of religions, and specifically the doctrines of the nature of the divine and the proper ways for humans to relate to it, are no longer essential. These issues are, then, not saved by the doctrine of the will to believe, but rather explained away.


One may pose the problem for James’s “The Will to Believe” as a dilemma. Either (a) faith is doxastically efficacious or (b) it is not. If (a), faith may pass the lawfulness requirement, but in making religious belief so efficacious, it no longer can pass the orthodoxy requirement. The faith that is lawful is no longer the faith of the faithful. This is the trajectory of the later pragmatist reconstructive tradition, but this is not what James himself promises to open the essay.

If (b) faith is not doxastically efficacious, then James may pass the orthodoxy requirement, as the commitments to trinitarianism, virgin births, and so on may yet count as faith (and not be doxastically efficacious), but the case for the lawfulness of these commitments falls. That is, if James allows faith in miracles in the distant past (n.b., precisely what Cliffordian evidentialism prohibits), he may keep the faith as recognizable to the faithful, but all the apparatus arranged to demonstrate its acceptability becomes otiose.

The failure of “The Will to Believe” on its own terms is not, however, a failure überhaupt. First, in moral truths, James has identified a class of propositions that have a surprising evidential pattern, that of doxastic efficacy. This is an important and useful observation, as we see that in this doctrine our ideas and attitudes make a difference. Second, in identifying the “essence of religion” as the view that the best things are more eternal and that it is better to believe that, James has contributed a blueprint for subsequent pragmatist reconstructions of religious belief as consistent with naturalism. Such a reconstruction of religious belief yields nothing orthodox, but for the later pragmatist tradition, this was not a serious concern. For these two positive contributions, James’ essay deserves our respect and praise. 

Vanderbilt University



James, William. 1896 [1977] “The Will to Believe” In The Writings of William James. Ed. John J. McDermott. Chicago: University of Chicago Press.

James, William. 1979-1882 [1977]. “The Sentiment of Rationality.” In The Writings of William James. Ed. John J. McDermott. Chicago: University of Chicago Press.

Dewey, John. 1934. A Common Faith. New Haven: Yale University Press.

Hodges, Michael. 1990. Transcendence and Wittgenstein’s Tractatus.Philadelphia: Temple University Press.

Peirce, Charles Sanders. 1960. “How to Make our Ideas Clear.” In The Collected Papers of Charles Sanders Peirce. Volume 5. Ed. Charles Hartshorne and Paul Weiss. Harvard University Press.

Phillips, D.Z. 2001. Religion and the Hermeneutics of Contemplation. Cambridge University Press.

Rorty, Richard. 2006. “Religious Faith, Intellectual Responsibility, and Romance.” In The Cambridge Companion to William James. Ed. Ruth-Anna Putnam. Cambridge University Press. 84-102.

West, Cornel. 1989. The American Evasion of Philosophy. Madison: University of Wisconsin Press.


[1]References to James’s essays here will be to John McDermott’s comprehensive edition, The Writings of William James. Chicago: University of Chicago Press, 1977, in the following format (WWJ: p#).

[2]References to Clifford’s essays will be to The Ethics of Belief and Other Essays, Ed. Timothy Madigan. Amherst, NY: Prometheus Books, 1999 [1877].

Leave a Reply