1.1 Proposed Defeaters for Perceptual and Memory Beliefs

Skeptical objections are potential defeaters. Arguments for skepticism concerning our perceptual beliefs about the external world are of this sort. I will focus not on rebutting objections, which argue that our perceptual beliefs are false, but on undercutting ones, which argue that our perceptual beliefs aren't formed in a trustworthy way or that we have good reason to doubt that they are formed in a sufficiently trustworthy way. Consider the underdetermination objection to our perceptual beliefs. It begins by noting that these beliefs are not based on arguments or inferences. Instead, they are noninferential beliefs based on sensory experiences that don't guarantee the truth of these beliefs. It seems that we could have these sensory experiences whether or not there is an external physical world and whether or not the external world is the way these experiences incline us to think that world is. If our perceptual evidence could be the same whether or not the external world is as we are inclined to think it is (on the basis of that evidence), then that evidence underdetermines the truth of our perceptual beliefs. One natural response is to seek out arguments—ones that don't rely on our perceptual evidence—for the conclusion that our perceptual evidence makes it probable that our perceptual beliefs are true. There are good reasons for thinking that such arguments aren't available. I won't rehearse those reasons here.Footnote ¹

Exactly parallel worries plague our memory beliefs, which are noninferential beliefs based on memory impressions or seemings. These memory seemings don't guarantee the truth of those beliefs: we could have the same memory seemings on which we base our memory beliefs whether or not there was a past and whether or not the past was the way this memory evidence inclines us to think it was. Our memory evidence, therefore, underdetermines the truth of our memory beliefs, and we don't have any good arguments (that don't rely on memory evidence) for thinking that our memory evidence makes it probable that our memory beliefs are true.Footnote ²

These underdetermination objections to our noninferential perceptual and memory beliefs are potential undercutting defeaters for those beliefs. The question, thus, arises: How can these perceptual and memory beliefs avoid being defeated, given that no good arguments are available for defending those beliefs inferentially against these underdetermination objections?

1.2 Responding to Proposed Defeaters for Perceptual and Memory Beliefs

In my view, the best response to this sort of skeptical worry about perception and memory is a two-step Reidian response that both relies on and responds to epistemic intuitions. The first step in this response is to highlight the strong epistemic intuitions we have that contradict the skeptical conclusions in question. I develop this first step using the work of Thomas Reid and a more recent philosopher, William Tolhurst. The second step is to point out that both the skeptical objections and the Reidian response rely on epistemic intuitions, but the epistemic intuitions on which the skeptical objections rely are much weaker than and, for that reason, outweighed by the stronger epistemic intuitions employed in the Reidian response to the skeptical objection.

Let's start by considering what Reid says on the topic. He notes that it is a first principle that our faculties are reliable and that first principles are properly believed noninferentially.Footnote ³ Just as we have noninferential beliefs about our immediate physical environment by means of sense perception and about our past by means of memory, so also we have a faculty of common sense by means of which we form noninferential beliefs in first principles.Footnote ⁴ How exactly does this faculty of common sense produce beliefs in first principles? According to Reid:

The suggestion is that when you entertain the contrary of a first principle (e.g., the principle that your faculties are reliable) you experience the emotion of ridicule, which involves the contrary of that first principle seeming absurd. On the basis of this seeming you dismiss as absurd the contrary of the first principle and believe the first principle itself. Thus, noninferential common sense belief in the reliability of your faculties is like noninferential perceptual belief in that both are based on experiential evidence—a claim's seeming absurd in the former case and sensory experience in the latter case.Footnote ⁶

Similar points have been developed in a different context by William Tolhurst who tries to capture the essence of seemings as follows:

Tolhurst calls this feel of a state whose content reveals how things really are its ‘felt veridicality’. It is the distinguishing feature of seemings. He goes on to speak of a higher-order awareness of this felt veridicality:

So, according to Reid, when a normal person considers the suggestion that all her beliefs formed on the basis of her strong perceptual and memory seemings are unreliably formed, it will first seem absurd to her that they are unreliably formed and it will then seem that they are reliably formed and she'll believe that. (Or at least this is what I take Reid to be suggesting. What follows in the remainder of this section is one way of developing these ideas. Whether this development captures what Reid says in his own writings is not of primary concern to me.) Tolhurst says something similar: when we consider, for example, a memory seeming that p, we become aware of its felt veridicality and this makes it seem to us that the memory seeming that p is veridical; on the basis of this higher-order seeming, we believe that the memory belief that p, based on that memory seeming that p, is formed in a reliable and nonmisleading way, contrary to the skeptical scenario that, according to the objector, should make us have doubts.

These higher-order seemings that our first-order beliefs are reliably formed or that our first-order seemings are veridical are examples of epistemic intuitions. Epistemic intuitions are like moral intuitions, which are seemings about moral matters, the difference being that epistemic intuitions are seemings about epistemic matters. Both are normative seemings. Just as one can have moral intuitions about particular cases, mid-level principles, or higher-level principles, so also one can have epistemic intuitions about particular cases, mid-level principles, or higher-level principles. The Reidian response to the underdetermination objection to our perceptual and memory beliefs relies on epistemic intuitions. The skeptical objection raises doubts about the reliability of our noninferential perceptual and memory beliefs or about the veridicality of our sensory experience or our perceptual or memory seemings. The Reidian reply is to reject those doubts and affirm the reliability of those beliefs and the veridicality of the seemings on which they're based. In doing this, the one endorsing the Reidian reply is forming a higher-level belief about the trustworthiness of our faculties and this higher-level belief is based on epistemic intuitions. Responses to skepticism in this Reidian or commonsensist vein have been championed by many since the time of Reid, including G.E. Moore and Chisholm, but also contemporary epistemologists such as Jim Pryor and John Greco.Footnote ⁹

(I should note here that not all who support a Reidian response to skepticism would focus on seemings or evidence in the form of epistemic intuitions (Greco, for example, wouldn't). I happen to think such a focus is a natural and plausible Reidian way to account for what in fact happens when we humans hear about and then rationally dismiss radical skeptical hypotheses and objections. However, it is not my view that having such seemings or intuitions is the only way any possible cognizer could rationally respond to skeptical objections. Nor is it my view that beliefs can't be justified unless they're based on seemings of some kind (e.g., memorial, perceptual, intuitional, etc.). It's just that I think rational human beliefs very often happen to be based, at least in part, on seemings of some kind.Footnote ¹⁰ )

Chisholm famously cites the Reidian response to Humean skeptical worries as a paradigm instance of a view he called ‘particularism’.Footnote ¹¹ The Humean skeptic believes an epistemic principle saying that we can't know anything about the external world or the past unless we have strong arguments for our conclusions—arguments beginning with premises about our current experience. Reid recognized that such a principle implies that we don't have any knowledge about the external world or the past, since no strong arguments of that kind are available. But he thought it was clear that we do have knowledge of the existence of the external world and the past. As a result, Reid and Moore and others in this tradition reject the epistemic principle on which the Humean skeptical argument depends. This is called a ‘particularist’ response to Humean skepticism because it gives more weight to intuitions about the epistemic status of particular beliefs than to intuitions concerning general epistemic principles about what is required for justification or knowledge—principles of the sort employed in skeptical objections.

An important thing to notice is that, in making this response, the Reidian particularist is doing the same sort of thing that the skeptic is doing. After all, how does the skeptic know that her epistemic principles are true? Presumably the epistemic principles, on which the skeptic relies in proposing her skeptical objections, are believed on the basis of epistemic intuitions. (Indeed, just as it's hard to see how moral beliefs are in fact justified without relying in some way at some point on moral intuitions, it's also difficult to see how skeptical arguments—for epistemic conclusions about what we epistemically shouldn't believe—are in fact justified without relying in some way at some point on epistemic intuitions.Footnote ¹² ) The Reidian can concede that there may be some initial appeal to the epistemic intuitions on which skeptical objections rely. But she goes on to insist that the strength of these skepticism-inducing epistemic intuitions pales in comparison to the strength of the epistemic intuitions in support of the rationality of beliefs about particular cases and in support of the reliability of the belief-forming faculties that produced them.

Putting all this in terms of defeaters, what we have is the following. I form my noninferentially justified perceptual and memory beliefs. I then become aware of the underdetermination objection—a skeptical objection suggesting that I should be doubtful of the reliability of these perceptual and memory beliefs. This is a proposed defeater for these beliefs and it is based, at least in part, on epistemic intuitions (e.g., the intuition that I should withhold judgment about the reliability of my beliefs based on evidence that doesn't guarantee their truth—at least until I've got an argument that those beliefs are made probable by that evidence). But when I consider the higher-level claim that my perceptual and memory beliefs aren't formed in a trustworthy way, I have the strong higher-level seeming that this is false and that, in fact, my perceptual and memory beliefs are formed in a trustworthy way. On the basis of this epistemic intuition, I deny the claim that they're unreliably formed and I also deny the skeptic's premise, which says that, in light of the underdetermination objection, I epistemically should be doubtful about or withhold my perceptual and memory beliefs. In short, this proposed defeater is deflectedFootnote ¹³ by my reliance on epistemic intuitions, ones that are stronger than the epistemic intuitions supporting the defeater. As a result, my justified perceptual and memory beliefs avoid being defeated by this skeptical objection.

Moreover, although epistemologists can explain in some detail how this happens (by mentioning the sorts of things I've said above and by developing those remarks at length), this Reidian response works just as well for the philosophically naïve. For example, an ordinary person might watch the movie The Matrix for the first time and, as she walks out of the theater, she might consider the possibility that her own perceptual beliefs are massively unreliable and she might wonder how she could rule out this possibility. She might even wonder if it is sensible for her to continue trusting her sensory experience. But soon after she considers this worry, she dismisses it as implausible and affirms that it would be silly to be skeptical in that way, on the basis of such concerns, about her perception. This ordinary person is responding to the underdetermination objection in basically the same way that Reid is, despite the fact that she wouldn't describe her response in the way I've been describing the Reidian response.

1.3 Satisfying the Skeptic

Notice that the claim here is not that this sort of response will satisfy the skeptical objector by offering her what she'll consider to be a successful proof that the objection fails. The claim, rather, is that these perceptual and memory beliefs can remain rational, in the way noted, in the face of these potential defeaters. This is because even if the skeptical objector is not satisfied by these responses, it's simply false that, in such a situation, rationality requires me to give up those perceptual and memory beliefs.Footnote ¹⁴ And what matters for epistemic defeat is not what the skeptic finds satisfying but what rationality in fact requires.

Given that this reply won't satisfy the skeptic, how is this a helpful response to skeptical objections? It can be helpful in a couple of ways. First, it's helpful for people who share the epistemic intuitions most non-skeptics share but who aren't sure how best to respond to skeptical objections, which seem to carry some weight. This response helps these non-skeptics to see that although there is a tempting appeal to the skeptic's epistemic intuitions, they are ultimately non-veridical and overridden by considerably stronger epistemic intuitions to the contrary. Second, it's also helpful for skeptical objectors for two reasons: (i) It helps objectors to see (if they haven't already) that those to whom they're objecting have an internally coherent response that fits their non-skeptical epistemic intuitions, which shows that the situation for the non-skeptic isn't as problematic as skeptical objectors sometimes think. (ii) It helps objectors to come face to face (if they haven't already) with the charge that although the epistemic intuitions supporting their objection are understood and to some degree appreciated by non-skeptics, these skeptical intuitions are viewed with skepticism by non-skeptics as non-veridical epistemic illusions; being faced with this charge might go some way toward helping these objectors to see that their skeptical objections aren't as strong as they may have supposed.

1.4 Internal Rationality, External Rationality, and the Unconvinced Skeptic

I've said that our perceptual and memory beliefs remain rational and justified in the face of the skeptic's underdetermination objection. For the most part, I've been using the terms ‘rationality’ and ‘justification’ as synonyms referring to what we might call ‘internal rationality’. Internal rationality differs from external rationality as follows: a belief is internally rational if the belief formation process is going as it epistemically should downstream from experience (i.e., in response to the subject's conscious mental states, which constitute her evidence); a belief is externally rational if the belief formation process is going as it epistemically should both upstream and downstream from the experience.Footnote ¹⁵ For example, a perceptual belief is internally rational if it is an epistemically appropriate response to the subject's conscious mental states, in particular her sensory experiences. But a perceptual belief can be internally rational even if these sensory experiences were not produced in an epistemically appropriate way but were, instead, artificially produced by a deceptive demon or a computer, as in the movie The Matrix. In that case, the perceptual belief based on these artificially produced sensory experiences would be internally rational but not externally rational.

(Notice that one can be either an internalist or an externalist about internal rationality. For example, if you think that what makes a belief an epistemically appropriate response to your evidence base is that it is a proper functioning response to it or a reliable response to it, then you are an externalist about internal rationality. But if you think that what makes a belief an epistemically appropriate response to your evidence base is that your belief fits that evidence base, where this fittingness relation holds of necessity between that belief and that particular evidence base, in virtue of some of the intrinsic consciously accessible features of the relata, then you are, presumably, an evidentialist and an internalist about internal rationality. My own view is that externalism is true. So although I think we humans typically form beliefs on the basis of conscious mental states, I also think (i) that it's possible for beliefs to be justified even if they aren't based on conscious mental states and (ii) that justification supervenes not on our mental states but on facts about what proper function requires.Footnote ¹⁶ )

With the distinction between internal and external rationality in mind, consider what I've said about satisfying the skeptic. I've said that epistemic intuitions in support of the justification of our perceptual and memory beliefs are much stronger than and outweigh the epistemic intuitions in support of the premises used in the underdetermination objection. But what if the skeptic's epistemic intuitions are different? What if her epistemic intuitions in support of the premises used in the underdetermination objection are as strong as or stronger than those indicating that her perceptual and memory beliefs are justified and reliably formed? Perhaps in that case, it won't be internally rational for the skeptic to join the Reidian respondent in treating the epistemic intuitions in support of the underdetermination objection as illusory; and perhaps that means that the skeptic can't save the internal rationality of her perceptual and memory beliefs in the Reidian way I suggest above. So perhaps the skeptic is internally rational to endorse the underdetermination objection on the basis of her epistemic intuitions, even though the non-skeptic is internally rational to endorse the Reidian response on the basis of her epistemic intuitions. What's internally rational can differ between people if their evidence (in this case, the epistemic intuitions they have) differs.

I've also said, in the previous subsection, that what rationality in fact requires is not that we give up our perceptual and memory beliefs in response to the underdetermination objection. The point here is that having epistemic intuitions in support of the underdetermination objection that are as strong as or stronger than those in support of the justification of our perceptual and memory beliefs is not in accord with external rationality. So even if the skeptic is internally rational in (i) endorsing the underdetermination objection on the basis of her epistemic intuitions that support it and (ii) withholding (or holding less firmly) her perceptual and memory beliefs in the face of that objection, she is not externally rational in doing so. The externally rational thing for the skeptic to do is to have weaker epistemic intuitions in support of the underdetermination objection and to have them outweighed by much stronger epistemic intuitions in support of the justification of her perceptual and memory beliefs, leading her to treat the skeptical intuitions as illusory. External rationality will result in this person continuing firmly to hold her perceptual and memory beliefs, despite the skeptical objection.

What is my basis for this view about what external rationality requires? Reliance on epistemic intuitions. Of course, the skeptic won't view things this way, so she will think she is being misjudged as irrational by the non-skeptic. (I've conversed with perceptual skeptics who feel this way.) But a similar point holds in the other direction: the non-skeptic won't view things the way the skeptic does, so she will think she is being misjudged as irrational by the skeptic. It's not easy to see a way forward that will guarantee a resolution between these two perspectives, a way for the skeptic and non-skeptic to engage in a philosophical conversation that is likely to bring about a shared point of view on the matter.

1.5 Concerns about Epistemic Circularity

Various objections can be raised against reliance on epistemic intuition. One kind of objection is based on work in experimental philosophy that raises concerns about the reliability of epistemic intuition. Unfortunately, I don't have the space to enter into a discussion of this challenge in this paper.Footnote ¹⁷ Another sort of objection to reliance on epistemic intuition asks the following question: What happens if the skeptic takes the underdetermination objection and uses it against beliefs formed on the basis of epistemic intuitions? (After all, these epistemic intuitions don't guarantee the truth of the beliefs based on them.) Won't there be trouble if one tries to use the Reidian response to defend epistemic intuitions against this objection? I.e., won't it be viciously circular to respond to the skeptic about epistemic intuitions by relying on epistemic intuitions to confirm the trustworthiness and rationality of relying on epistemic intuitions?

Before addressing that concern, let's consider a related worry that this skeptic about epistemic intuition faces—the worry that using the underdetermination objection to support skepticism about epistemic intuitions is self-undermining. This objection itself relies on premises that are based, ultimately, on epistemic intuitions (e.g., the intuition that I epistemically should withhold judgment about the reliability of my beliefs based on evidence that doesn't guarantee their truth—at least until I've got an argument that those beliefs are made probable by that evidence). So the skeptic appears to be guilty of relying on epistemic intuitions to argue that we can't trust epistemic intuitions. The skeptic might respond by pointing out that it's possible to learn that a belief source is problematic by discovering that it indicates its own unreliability.Footnote ¹⁸ That is a response worth taking seriously. But the question is whether, in making such a discovery, the belief source should be mistrusted wholesale or whether, instead, our mistrust should be directed more narrowly at the alleged indicators of the source's supposed untrustworthiness. The Reidian will be happy to acknowledge that our belief sources are imperfect. But in this particular case, she'll be inclined to think that the problem is with the epistemic intuitions behind the underdetermination objection, not with epistemic intuition as a whole.

Let's return, briefly, to the circularity problem that afflicts the Reidian anti-skeptic who relies on epistemic intuitions in forming the belief that her epistemic intuitions are veridical. Is that a problematic kind of epistemic circularity? The short answer is that relying on epistemic intuitions in forming the belief that one's epistemic intuitions are veridical needn't be problematic even though it does manifest a kind of epistemic circularity. (A belief, B, is epistemically circular if (i) it is a belief in the trustworthiness of a belief source S and (ii) the person holding B depends on belief source S in forming or holding B.) Sometimes epistemic circularity is benign. In the case of epistemic intuition, it is benign when the person with such intuitions neither is nor (epistemically) should be seriously questioning or doubting the trustworthiness or veridicality of these intuitions. I don't have the space here to give a detailed account of this way of defending benign epistemic circularity, but I do so elsewhere.Footnote ¹⁹

3.1 p-evidence, Rp-evidence, and R~p-evidence for Theistic Belief

First, let's consider the p-evidence, the evidence concerning theism itself. I will be ignoring theistic arguments and focusing instead on evidence on which noninferential theistic belief is based. There are various kinds of experiential evidence for noninferential theistic belief. One kind is dramatic religious experience, including in particular perceptual experience of God. In his Perceiving God, Alston focuses on mystical perception, in which God seems to reveal himself in some vivid or shocking or overwhelming way. Alston thinks of it as perception because God is (allegedly) perceived as doing something vis-à-vis the perceiver or as having some perceivable property such as goodness, power, or kindness.Footnote ²⁸ Because this kind of perceptual experience of God is rare, I will set it aside and focus instead on a different kind of experiential evidence for noninferential theistic belief, namely, theistic seemings.

Just as noninferential memory, moral, and mathematical belief are often based on memory, moral and mathematical seemings, so also, noninferential theistic belief is often based on theistic seemings.Footnote ²⁹ Many things can trigger ordinary theistic seemings. They might be triggered by things upstream from experience, including things such as the direct activity of God (this is one way of thinking about at least some instances of what the Christian tradition calls ‘the testimony of the Holy Spirit’). But they can also be triggered by things downstream from experience such as feelings of guilt or being forgiven or desperate fear or gratitude; other triggers can be experiences of awe at the grandeur and majesty of oceans, mountains, or sky.Footnote ³⁰ Another way theistic seemings can arise is in response to the spoken or written testimony of others: we encounter the testimony and what is said simply seems right.Footnote ³¹ Theistic seemings can also result from ruminating upon what we have learned about the immensity, complexity, mysteriousness, and possible origins of nature and of the human mind.Footnote ³² Likewise, a consideration of the apparent design in nature (e.g., in the biological world and in Big Bang cosmology) can prompt a seeming that God designed these things, a seeming that isn't based on any argument from design and that is compatible with believing in evolution.Footnote ³³ These theistic seemings aren't the results of simply considering the proposition God exists and finding that it seems true; nor are they conclusions of arguments. They are more like what Audi calls ‘conclusions of reflection’, which are not based on inferences from premises but instead emerge noninferentially from an awareness of a variety of observations, experiences, and considerations over a (perhaps long) period of time.Footnote ³⁴ Thus, in a certain sense, the evidence I'm thinking of doesn't consist solely of theistic seemings. It also includes many recent and long past observations, experiences, testimony, considerations, and the traces of these retained in memory out of which these theistic seemings emerge noninferentially upon reflection. It is often the case that we are unable to trace the origins of our theistic seemings. But, as Ernest Sosa and Graham Oppy point out in other contexts, the fact that we can't trace the origins of our seemings doesn't show that the beliefs based on those seemings aren't rational or trustworthy.Footnote ³⁵ And just as memory, moral, and mathematical seemings can be weak or very strong, so also theistic seemings can be weak or very strong. The kind of p-evidence that I have in mind consists of strong theistic seemings. These theistic seemings might not be as strong as our strongest perceptual and memory seemings. But this doesn't mean they aren't strong enough to be used in making theistic belief rational in the face of disagreement.

The accounts given of what justifies noninferential theistic beliefs so based will be similar to accounts given of the justification of perceptual, memory, mathematical, or moral beliefs. Evidentialists will say that theistic beliefs so based are justified if the beliefs fit the evidence on which they're based.Footnote ³⁶ Reliabilists will say that beliefs so based are justified if the processes by which such beliefs are formed are reliable.Footnote ³⁷ Proper functionalists will say that beliefs so based are justified if they are produced by properly functioning cognitive faculties in the appropriate cognitive environment.Footnote ³⁸ Insofar as theistic beliefs can be fitting responses to theistic seemings and insofar as theistic beliefs based on theistic seemings can be reliably formed or produced by properly functioning faculties, noninferential theistic beliefs formed on these bases can (according to these familiar accounts of epistemic normativity) be justified or rational.

As I suggested earlier, in order for theistic beliefs based on theistic seemings to be internally rational, things must be going as they epistemically should downstream from experience (i.e., they must be epistemically fitting responses to triggers consisting of other current or previous mental states of the subject). Thus, on my view, it often isn't merely strength of seeming that is required, where that strength is understood subjectively. For example, in cases where theistic seemings are conclusions of reflection, then, even though they aren't inferential, they must be epistemically appropriate responses to previous mental states of the subject (if such seemings are to be sources of justified belief). But in cases where the theistic seemings aren't formed in response to previous mental states, what matters for internal rationality is mainly that they aren't failing to be appropriately responsive to any defeating evidence one might have.Footnote ³⁹

The relevant Rp-evidence is the higher-order seeming that the theistic seemings, on the basis of which one believes that God exists, are veridical. It's not that the theist finds it absurd that any seeming on a religious topic could be nonveridical. Rather, when she ponders the particular theistic seemings on which her own theistic beliefs are based, it seems strongly to her (because of an awareness of their felt veridicality) that these theistic seemings are veridical. These higher-order seemings are epistemic intuitions about particular theistic beliefs. They are the very same sort of evidence that makes us think our strong memory, moral, and mathematical seemings are veridical. The kind of Rp-evidence I have in mind, then, is strong epistemic intuition in support of the veridicality of the strong theistic seemings on which our theistic beliefs are based. Again, these epistemic intuitions in support of our theistic seemings might not be as strong as our strongest epistemic intuitions in support of the veridicality of our perceptual and memory seemings. But this doesn't mean they aren't strong enough to be used in making theistic belief rational in the face of disagreement.

Our overall Rp-evidence is more strongly supportive when it also includes another kind of Rp-evidence, namely, the recognition that many other people are theists, including many who are exceedingly intelligent and aware of potential defeaters for theism, whose moral character is extremely admirable, and who are very matureFootnote ⁴⁰ and practically wise in ways that enable them to flourish in their environments. These considerations have to be balanced against the fact that many non-theists are also like this. But they still make our Rp-evidence more strongly supportive than it would be in the absence of such awareness of numerous other intellectually and morally impressive theists.Footnote ⁴¹

The theist's R~p-evidence is her evidence bearing on whether the belief that theism is false, held by those who disagree with her, is formed in a reliable and nonmisleading way. I don't think the theist has very strong evidence that the atheist's belief on this matter is reliable. Of course, educated theists are aware that both atheists and theists have developed arguments for their positions and responded to arguments against their positions. They are also aware that, just as with other topics in philosophy, there is a lot of variation in the assessments of these arguments and responses. If one considered only the philosophical literature on theistic and atheistic arguments, it would be very controversial indeed to say that it gives us strong evidence that the atheist's beliefs on this matter are reliable (it would be about as controversial as saying that the philosophical literature on incompatibilism about free will gives us strong evidence that the incompatibilist's beliefs about incompatibilism are reliable). Focusing only on atheistic arguments, the strongest seem to be arguments from evil and, although some of those arguments are more plausible than others, it seems fair to say that none of them are knockdown arguments and none of them are strong enough to rationally require consent from all informed intelligent people.Footnote ⁴²

Moreover, even if a theist thinks that theistic arguments aren't strong enough to justify theistic belief and that atheistic arguments have some force, that isn't sufficiently strong evidence for thinking that the atheist's beliefs on this matter are reliable. Consider again the Jury Case. In that example, you've been accused of a crime and you agree that the jury has very strong evidence for thinking you are guilty. Nevertheless, your overall evidence for thinking that the jurors’ beliefs about your guilt are reliable is not strong. This is because you think their evidence is deficient in an important way in which your own evidence (which includes your vivid memories) is not. Something similar might be true of the theist who is thinking about the atheist who relies on atheistic arguments that have some force. Although the theist might have good reason to think that the most sophisticated atheists are as good as anyone else at formulating valid arguments with somewhat appealing premises on the topic of God's existence, the theist doesn't have good reason to be equally impressed with the atheist's capacity or tendency to have appropriate theistic seemings and to respond properly to them. The theist has her theistic seemings and her higher-order seemings about these theistic seemings, which assure her of their veridicality. But when she considers the atheist, what she notices is that the atheist apparently lacks these seemings, or at least that she has them only weakly and doesn't have or trust any higher-order seemings about the veridicality of her theistic seemings. In noticing these things about the atheist, the theist isn't thereby getting strong evidence that these atheistic beliefs are being formed in reliable and nonmisleading ways. If anything, she's getting evidence to the contrary, given her own theistic seemings and their felt veridicality. The theist's assessment of the atheist who has somewhat forceful atheistic arguments but who lacks or doesn't trust theistic seemings is like your assessment (in the Jury Case) of the jurors who have strong evidence for your guilt but who lack the memories you have of your innocence: you don't have strong evidence for the reliability of the jurors about your guilt and the theist doesn't have strong evidence for the reliability of the atheist about whether God exists. Importantly, evidence for the atheist's overall intelligence, virtue, etc. doesn't count as strong evidence for her reliability on the topic of theism, since such intelligence and virtue are compatible with (and not made improbable by) being not very good at all at forming accurate beliefs about theism.Footnote ⁴³ So these theists don't have strong R~p-evidence in support of the reliability of the atheist's beliefs on whether theism is true.

In sum, I think there are many fairly ordinary cases of theistic belief where the theist's p-evidence and Rp-evidence are strongly supportive, and her R~p-evidence is at best weakly supportive. As a result, these cases are more like the Jury Case or the Math Conference Case, where rational demotion occurs and the belief remains undefeated, than the Restaurant Case, where the belief is defeated by the recognition of peer disagreement. This is why I think theistic belief, at least of the sort I described, can be rational in the face of disagreement.Footnote ⁴⁴

3.2 Internal Rationality, External Rationality, and Satisfying Objectors to Theism

Will this response satisfy objectors to theism? Probably not. But, as with the response to skepticism about perception and memory, that isn't the goal. The goal is to consider what rationality requires in certain cases of disagreement about theism. And for the reasons given above, I think there are cases where rationality does not require theists to give up their theism. It may be that the objector to theism has different epistemic intuitions or epistemic intuitions of different strengths—about whether theistic seemings have evidential value and are veridical or about what rationality requires in the face of disagreement. But from the perspective of the theist I have in mind, the objector is mistaken not only in rejecting theism but also in endorsing epistemic intuitions that support the objection and conflict with the theist's response—a response based in part on the theist's stronger epistemic intuitions in support of her theistic beliefs. In this way, the defense of theistic belief in response to the objection from disagreement over theism is similar to the Reidian defense of perceptual and memory beliefs in response to underdetermination objections targeting those beliefs: in each case the epistemic intuitions driving the objections—intuitions about principles concerning what sort of belief retraction is rationally required in certain circumstances—are ultimately overridden by stronger epistemic intuitions (in the form of higher-level seemings) about the veridicality, reliability, or rationality of first-order perceptual, memorial, and theistic seemings or beliefs.

Notice that, just as in the dispute between the radical skeptic and the Reidian defender of perception and memory, here too both sides might be internally rational. The theist could be internally rational in adopting the Reidian-style response to skeptical objectors, since such a response could be the rational response to have if one has strong theistic seemings and strong epistemic intuitions supporting the veridicality of those theistic seemings, and these are formed in epistemically appropriate ways downstream from experience. At the same time, if the objector to theism doesn't have theistic seemings or doesn't have epistemic intuitions supporting the veridicality of such seemings (or if she lacks both or if both are very weak) and these things are epistemically appropriate responses to her experience, then the internally rational response for her is skepticism about theism.

As for external rationality, if the theist's strong theistic seemings and her strong epistemic intuitions supporting their veridicality (and the things upstream from experience that produce these two kinds of seemings in the theist) are in accord with what external rationality requires, then her Reidian-style response to objectors to theism is externally rational as well, just as the Reidian response to radical skepticism is externally rational.

But what about the objector to theism? Is she doomed to external irrationality, according to the defense of theistic belief offered in this paper? Perhaps not. Granted, it may be that the objector to theism has something going epistemically wrong upstream from experience that prevents her from having epistemically appropriate theistic seemings (of sufficient strength) or from having epistemically appropriate epistemic intuitions supporting the veridicality of such theistic seemings. If so, then her skepticism would be externally irrational. But perhaps the situation of the objector to theism is more like the jurors’ situation mentioned in the Jury Case. The jurors in that example weren't externally irrational in their skepticism. Instead they were quite rational. But they were mistaken because, through no epistemic deficiency on their part, they didn't have the (very good) evidence that the accused had. Moreover, they were rational (though mistaken) not to believe that the accused had good evidence. If the situation of the objector to theism is like this, then she might be both internally and externally rational in her skepticism while the theist is also internally and externally rational in her theism (just as the jurors and the accused are both internally and externally rational in their different beliefs about the guilt of the accused).

I don't mean to suggest that things are as simple in the dispute between theists and nontheists as they are for the jurors and the accused in the Jury Case. I mention that jury example here merely to emphasize that there are a variety of ways for theists and nontheists to interpret each other when considering whether the recognition of their disagreement should lead them to view themselves, or the other, as irrational.Footnote ⁴⁵ My main purpose in this paper has been to explain how it is that—by relying on epistemic intuitions in the way Reidians do in response to skeptical challenges to perception and memory—theists can rationally maintain their theistic beliefs in the face of disagreement-based skeptical worries.Footnote ⁴⁶