Michael Pace on Aaron Rizzieri’s Pragmatic Encroachment, Religious Belief, and Practice

According to a recent Pew poll, the proportion of adults in the U.S. who claim to believe that God exists has dropped from 92% to 89% since 2007. Though statistically significant, the decrease is relatively modest, and one might well be struck by how overwhelmingly theistic the U.S. has been and remains. More dramatic is the drop (from 71% to 63%) in the share of the population that claims to be absolutely certain that God exists. More people than ever before (including one out of three millennials) claim to believe that God exists without being certain.

One might conjecture that what motivates many of those who self-identify as uncertain theists is a desire to be intellectually honest about their evidence. Perhaps they think that their overall evidence favors God’s existence somewhat, but not nearly enough to warrant certainty. Whatever the merits of this psychological conjecture, it suggests related normative questions about the strength of evidence required for religious beliefs to be epistemically appropriate: Must one have evidence that makes God’s existence certain in order for theism to be rational, or is rational theism compatible with evidence that leaves room for doubt? Is the evidential standard for rational religious beliefs especially high, like the “beyond a reasonable doubt” standard familiar in criminal murder trials? Or is it lower, perhaps more like the “preponderance of evidence” required for convicting in civil court?

These questions are central to Aaron Rizzieri’s book. One of his main conclusions is that many people must meet a higher evidential standard in order to have rational religious beliefs than they must meet in order to have rational beliefs about other, more trivial matters. On Rizzieri’s view, a rational belief in God may even require evidence that warrants being certain of God’s existence. If this is correct, there may be no safe harbor for uncertain theists who aim to be rational while acknowledging that their evidence does not warrant anything close to certainty.

Although Rizzieri sets the bar high for the rationality of religious belief, he does not set the bar especially high for the rationality of engaging in religious practices. A second main conclusion of the book is that one can reasonably and meaningfully engage in religious practices with less evidence by hoping (without believing) that religious claims are true. If both of Rizzieri’s main conclusions are correct, the religious practitioners who are most in line with the dictates of rationality are a group much rarer in the American religious landscape than uncertain theists: religiously devout agnostics.

Rizzieri argues for these conclusions by defending a relatively new epistemological theory (the “pragmatic encroachment” theory of Rizzieri’s title), which he then applies to religious epistemology. Pragmatic encroachment is a general theory about how much warrant or evidence is required for knowledge or rational belief. The best way to begin to understand the theory is to consider a non-religious case of the sort used to motivate it:

I am at a restaurant considering what to have for dessert. The chocolate brownie on the menu looks good, but I’m not sold yet. I ask the waiter whether it has nuts, explaining that I have a slight preference for the kind of brownie without nuts. The waiter tells me that there are no nuts in the brownie. “Okay,” I reply. “Now that I know it doesn’t have nuts, I’ll have a brownie.”

Unbeknownst to me, Virginia is sitting at a nearby table and overhears the waiter tell me that there are no nuts in the brownies. Virginia is also considering dessert, and brownies are her absolute favorite. She was hoping that the brownie doesn’t have nuts because she has a very bad nut allergy; ingesting even a small amount of nuts is life-threatening.

You probably did not balk at my claim to know that the brownie lacks nuts based on the waiter’s testimony. The testimony of waiters is usually sufficient evidence for us to know things about menu items that we didn’t know before. One may also judge, however, that Virginia is not in the same position to know that the brownie lacks nuts, even though the story portrays her as having the very same evidence I have. Pragmatic encroachment theorists take the example to show that two people can have the very same quality of evidence for a proposition while only one of them is in a position to know that the proposition is true. They hold that practical or moral considerations — such as the fact that whether the brownies have nuts is a matter of life or death for Virginia — can “encroach” on whether you know, influencing the evidential standards for knowledge or rational belief. Thus, even though the testimony of a waiter is usually good enough evidence for knowing that a brownie is nut-free (as it is for me in the example), it is not good enough evidence for Virginia, for whom so much is at stake.

Rizzieri’s version of the pragmatic encroachment theory appeals to the following general principle: If your evidence is not good enough to make it reasonable for you to act on the assumption that a proposition is true, then your evidence is not good enough for knowledge either. Applied to our example, Virginia does not know that the brownie is nut-free, since her evidence is not good enough for her to reasonably act on the assumption that it is nut-free. He also defends an encroachment principle linking knowledge to the moral category of culpable negligence: If it would be morally negligent for you to act as if some proposition is true, then your evidence is not good enough to know the proposition. Suppose, for example, that Virginia explains her severe nut allergy to the waiter and then asks him if the brownies are nut-free. The waiter would be morally negligent if he were to answer “yes” based on evidence that leaves room for even a slight doubt, and so, given Rizzieri’s theory, he will not know that the brownies are nut-free unless his evidence is of a very high caliber.

I have so far described Rizzieri’s version of pragmatic encroachment in terms of how much evidence is required for knowledge. He frames his theory in terms of “justified belief,” which is a term of art in epistemology that signifies something similar to what I have called “rational belief.” Knowledge differs from justified (or rational) belief in that one can only know things that are true, whereas it is possible to have false but justified beliefs. For example, people living in an age before science were justified or rational in believing that the earth was flat (since this fit their available evidence), even though they did not know that the earth is flat (since it is not). Despite this difference between justification and knowledge, Rizzieri argues that the evidential standards are the same for each. To be justified requires having “knowledge-grade” evidence, that is, evidence that would be strong enough for knowledge if it were not misleading. Given this, it is fairly straightforward to apply the rationale for pragmatic encroachment sketched above to justified belief. Since Virginia’s evidence for the claim that the brownies lack nuts is not knowledge-grade evidence, she neither knows nor is justified in believing this claim.

What, then, are the implications of pragmatic encroachment for religious epistemology? The answer is complicated. The very point of the pragmatic encroachment theory is that details about what you care about and what practical decisions you face can affect which beliefs are and are not justified for you. Without knowing these details, we cannot say much about what you are justified in believing. This makes it tricky to make general pronouncements about the implications of pragmatic encroachment for religious epistemology.

For example, at one point Rizzieri advertises that he will give a “Stakes-Based Argument against Miracles.” The key idea of the argument is that believing that a miracle has occurred (or will occur) often involves serious practical and moral risks. (A particularly tragic recent example involved members of a Liberian Pentecostal church who believed, with disastrous effects, that a miracle would result from their praying and laying hands on a visitor suffering from Ebola.) Given pragmatic encroachment, these high practical and moral risks raise the evidential standards for having a rational belief in a miracle. Rizzieri develops a version of the argument that targets the Christian claim that Jesus miraculously rose from the dead. Many people, he says, are in a high-stakes situation with respect to this central Christian doctrine. Indeed, the Apostle Paul seems to acknowledge this when he claims that he and his followers are “of all people to be most pitied” if the resurrection of Jesus never occurred. Rizzieri contends that Christians such as Paul must meet a higher evidential standard to rationally believe in the resurrection, similar to the higher standard Virginia must meet in order to rationally believe that the brownies lack nuts.

Notice, though, that this argument is not fully generalizable, since it only has traction against someone for whom the practical or moral risks of believing a miracle are high. As Rizzieri himself acknowledges, not everyone faces the same high stakes with respect to miracle beliefs, including a belief in Jesus’ resurrection. Consider, for example, Rizzieri’s description of a fictional character, Clive:

Clive has hoped and lived as if Jesus was physically resurrected from the dead as long as he can remember. Clive’s hope grounds his optimistic sense that God cares about humanity and has played a significant role in shaping his own spiritual/moral character. Clive views Jesus as an exemplar of self-sacrificial giving and obedience to the divine. Furthermore, Clive has cultivated a variety of attendant spiritual attitudes and practices that are both meaningful to him and beneficial to others, but would make much less sense to Clive if he were forced by strong arguments to believe that [Jesus did not rise from the dead] which would undermine his hope [that Jesus rose from the dead].

The practical and moral consequences of believing that Jesus rose from the dead seem not to be bad for Clive, even if the belief is mistaken. (Rizzieri stipulates that Clive’s believing in the resurrection would not tempt him to hold any morally dangerous or repugnant beliefs, such as sexist or homophobic beliefs that are often associated with Christian fundamentalism.) Clive is thus immune from the “Stakes-Based Argument against Miracles,” and the standard of evidence required for him to rationally believe the resurrection will be no higher than “ordinary standards” that apply to beliefs about relatively trivial matters.

The most pressing practical and moral risks for Clive, as Rizzieri describes him, do not stem from the possibility that he will have a false belief about the resurrection, but rather from the possibility that he will fail to have a positive attitude toward the resurrection that will allow him to engage meaningfully in the religious practices that are practically and morally beneficial to him. This might lead one to wonder whether, given pragmatic encroachment, Clive’s circumstances actually make it easier for him to have a rational belief in the resurrection by lowering evidential standards below what is required for trivial beliefs. A version of pragmatic encroachment (not endorsed by Rizzieri) that has this consequence might take its inspiration from William James’s discussion of the truth goal in “The Will to Believe.” James points out that we have two competing intellectual goals. One goal is to avoid believing what is false, which is achieved easily by remaining agnostic about whether a claim is true or false. The second goal is to believe what is true, a goal which is impossible to achieve by being agnostic. The two goals pull in opposite directions, and James says that which goal we prioritize “may end by coloring differently our whole intellectual life.”

Inspired by James, a pragmatic encroachment theorist might argue that practical and moral considerations are relevant to striking the appropriate balance between the goals of avoiding errors and believing important truths. In Virginia’s case, it would be a disaster if she were to believe in error that the brownies lack nuts, whereas the potential benefits of believing this if it turns out to be true are comparatively slight. Adopting very high evidential standards in such cases protects against the high cost of error. Clive’s situation with respect to the resurrection of Jesus is just the opposite. For him, the practical and moral cost of erroneously believing that Jesus rose from the dead would be relatively modest; but if it is true that the resurrection occurred, remaining agnostic about it seems to carry a significant practical and moral cost for Clive, depriving him of an important true belief that is central to the religious practices he values. Might these considerations significantly lower the evidential standards for Clive, so that, for example, he could rationally believe the resurrection based on evidence that makes the resurrection barely more likely than not?

Not according to Rizzieri. In an interesting last chapter, he gives several objections to William James’s reasoning, some of which are relevant to the James-inspired argument just given.

A first objection stems from Rizzieri’s claim that one must have knowledge-grade evidence in order to have a justified or rational belief. The evidential standards for knowledge are arguably very high (perhaps humorously high), even for trivial beliefs, and it does not seem that they can be significantly lowered. Our ordinary concept of knowledge seems specifically designed to protect against errors in belief without tolerating much risk. Thus, it seems Clive would need evidence that makes the resurrection very likely if rational belief requires knowledge-grade evidence.

Why think that rational belief requires knowledge-grade evidence? According to Rizzieri, this is a consequence of the very nature of belief. To believe something, he argues, involves a commitment to the claim that you know it. This is supposed to explain why it sounds strange to say that one believes what one denies knowing (e.g., “I believe that the brownies are nut-free, but I do not know whether the brownies are nut-free”) and also why rational people automatically cease to believe things that they discover they do not know. According to Rizzieri, it is therefore appropriate to evaluate beliefs with respect to a fundamental rule: You should not, rationally, believe something unless you have knowledge-grade evidence for it. It follows that Clive should not believe that Jesus was resurrected or that God exists unless he has knowledge-grade evidence for these claims.

Rizzieri is not idiosyncratic in holding that one should only believe claims for which one has knowledge-grade evidence. Other contemporary epistemologists have defended these or other closely-related “knowledge norms” for similar reasons. These views, however, are resistible. They seem to me to overlook the important category of reasonable or rational opinion, which we arguably use to positively evaluate beliefs that we explicitly do not claim to know. We let people have and express their opinions when we are talking politics, philosophy, or sports, for example, and there seems nothing strange about claiming that we believe things in these domains that we deny knowing. The standards of evidence for rational belief in these domains are not as high as standards for knowledge, although this is not to say that there are no evidential standards for them. Although believing a proposition does not always commit one to the claim that one knows the proposition, believing does plausibly commit one to the claim that the proposition is more likely than the relevant alternatives. This suggests that the evidential standard for rational belief may sometimes be as low as a mere “preponderance of evidence,” rather than the higher, “beyond a reasonable doubt” standard that is arguably required for knowledge.

Rizzieri raises a further objection (related to the second main conclusion of the book) to the idea that the evidential standards for rational religious belief are significantly lower for Clive. The James-inspired argument above depended on the idea that Clive must believe in the resurrection of Jesus in order to secure the practical and moral benefits associated with engaging in Christian religious practices. Rizzieri challenges this idea, arguing that Clive could secure the same benefits by hoping, without believing, that the resurrection occurred. More generally, Rizzieri claims that agnostic hope for religious claims can support being “fully committed” to a spiritual life, allowing one to “pray, meditate, fast, give alms, and pursue other staple spiritual practices.”

Although I am sympathetic to Rizzieri’s claim that religious hope (without belief) can serve as a basis for reasonably engaging in some religious practices, I am doubtful that it can ground anything close to full participation in religious practices. Suppose Clive tries to engage in Christian practices with agnostic hope in the truth of Christian claims. Notice that any practice that requires sincerely asserting religious claims will be off-limits to Clive, since sincere assertion — whether in one’s own private mental life or to others — requires believing what is asserted. Many Christian practices, including singing hymns and reciting creeds, involve sincere assertions. Perhaps most troubling, sincere assertion plays a large role in what is perhaps the most important practice in religion: prayer. For example, the most central prayer in Christian practice — the “Lord’s Prayer” (or “Our Father”) — begins by asserting that we have a Father in heaven whose name is hallowed, and it asserts a lot more besides. Clive cannot sincerely pray it “as Jesus taught.” He might attempt to modify it, although doing so would require some crazy verbal gymnastics. (He might try this: “Our Father, who might exist and might be in heaven, may your name — if it refers to someone real — be hallowed…”)

Even setting aside the aesthetic poverty of the resulting liturgy, Clive’s prayers will lack many of the rich emotions toward God that are commonly expressed through prayer: praise, gratitude, lament, penitence, love, etc. Such “reactive emotions” essentially involve beliefs about the good or ill will of others. To feel gratitude toward someone, for example, involves believing that the person has benefited you in some way. It is difficult to see how one could rationally feel or sincerely express gratitude toward someone without thereby believing that the person exists and has benefited you.

Again, I am sympathetic to Rizzieri’s idea that it can be reasonable to pray and engage in other religious practices in the face of serious doubts, or even despite concerns that one is praying into the void. Sincere engagement in many religious practices, however, requires a commitment to the truth of religious claims that goes beyond a mere hope. A more promising way forward in defending the rationality of engaging in religious practices would be to defend the idea that religious beliefs can be held without much certainty and can be rational even when one’s evidence leaves considerable room for doubt.