Religious disagreement and epistemic commitment

The interlocutors I have in mind are committed religious believers who deny the epistemic-potential claim. Some might deny that religious disagreement has any positive epistemic value. Others might agree that it may have some, but think nonetheless that suppressing it is epistemically safer. Like Plantinga, they might reason that, although a ‘heightened awareness of the facts of religious pluralism could bring about a reappraisal of one's religious life, a reawakening, a new or renewed and deepened grasp’ (Plantinga (Reference Plantinga, Meeker and Quinn2000a), 190) of your religious beliefs, it will usually have the opposite effect: ‘For many or most exclusivists, I think, an awareness of the enormous variety of human religious responses . . . directly reduces the level of confidence or degrees of belief in the proposition in question’ (ibid., 189, italics added; cf. McKim (Reference McKim2012), 155; Gellman (Reference Gellman2000) ). Moreover, ‘[s]ince degree of warrant depends in part on degree of belief’, disagreement ‘can also deprive the exclusivist of the knowledge that [his religious beliefs] are true, even if they are true and he believes that they are’ (Plantinga (Reference Plantinga, Meeker and Quinn2000a), 189).

In defending the epistemic-potential claim, then, I must reassure committed believers not only that engaging in religious disagreement stands to deliver epistemic benefits, but also that the risks are less grave than one might fear – both on their own and in balance against the risks of not doing so.

One might worry that my proposal is incoherent: if disagreement is epistemically beneficial, this is because of its potential to correct or change your belief system – and readiness to alter your belief system seems incompatible with epistemic commitment to it.Footnote ³ But this objection fails to recognize how multifaceted epistemic commitment can be. Belief systems are not monoliths, but rather ‘webs’ with numerous elements of graded centrality. We may grant that there are some sorts of belief-system revision that epistemic commitment is incompatible with. But it is perfectly compatible with others. For example, being strongly committed to your belief system's central components is compatible with discovering that these components do not, after all, epistemically necessitate certain other attitudes to which you had been committed on their account. It is also compatible with the discovery of new attitudes that you should be committed to in light of them.

One might also worry that, in tolerating any epistemic commitment, I am licensing epistemic entrenchment, dogmatism, or fundamentalism. These attitudes are characterized by an unwillingness to assimilate evidence against one's view and are both epistemically and socially dangerous; religious epistemology must surely strive to eliminate rather than accommodate them. I have two responses. One is that these attitudes only accompany the most extreme forms of epistemic commitment, so tolerating commitment does not entail tolerating them too. Indeed, most of us are epistemically committed, for example to certain ethical beliefs, but this does not mean that we would not change them given enough evidence that they are mistaken. Second, my intended audience includes the entrenched, dogmatic, and fundamentalist, though is not exhausted by them (cf. Dormandy (Reference Dormandy2018c) ). If I can convince believers of any stripe that religious disagreement contains epistemic benefits, then, even if they are only willing to countenance small epistemic adjustments, this is far better than nothing.Footnote ⁴ Religious disagreement could issue in drastic belief revision, but it need not. For the sake of fostering what benefits we can, we must not expect interlocutors to cede more, epistemically, than they are prepared to.

So although the epistemic-potential claim is inspired by the philosophy of science, I will not be assuming that the parties to a religious disagreement take a neutral, scholarly attitude towards their belief systems – on the contrary.

Religious belief systems and disagreement

A core concern of religious epistemology is epistemically ‘central’ religious beliefs, such as ‘God exists’. But this narrow focus ignores the array of other elements of a belief system, including more peripheral beliefs and things other than beliefs, that can be fruitfully disagreed over (Dormandy (Reference Dormandy2018a) ). No wonder, then, that committed believers might be wary of religious disagreement. I will thus begin by zooming out to a broader picture of the objects of religious disagreement. At the same time, I will lay some cognitive-psychological groundwork for defending the epistemic-potential claim.

Three epistemological worries

Each type of epistemic influence can promote or impede accuracy. Evidence, arguably constitutively, is supposed to promote true beliefs. In many cases it does. However, whether it does in a given case depends, first, on how good a sampling of reality it is: it is highly unlikely that a given person can have all of the epistemically relevant experiences that there are to be had, so it is highly likely that everyone is missing out on relevant information. Whether a body of evidence promotes the truth depends, second, on whether the psychological and social factors that influence one's evidence and evidence-evaluation are themselves sensitive to the truth; this affects the content of a person's individual pieces of evidence as well as how representative of reality as a whole her evidence is.

And many psychological and social factors are not truth-sensitive. Some are sensitive to other considerations, such as survival. For example, we may be wired to believe, of any snake that we encounter, that it is poisonous, even if statistically this is highly unlikely. We may also be wired to believe that social hierarchies are just even when they are not, because this promotes social cohesion (Hundleby (Reference Hundleby, Brownstein and Saul2016), 241–243). And we are wired to be far more optimistic about our own capabilities than the evidence warrants; this arguably ‘motivate[s] us to get out of bed and optimistically take up challenges’ (Chabris & Simons (Reference Chabris and Simons2011), 126). Even many so-called ‘epistemic values’, such as a preference for simplicity over explanatory breadth, are not truth-sensitive as such, but are rather a priori preferences reflecting little more than aesthetics or pragmatic needs (Solomon (Reference Solomon2001), 16–20; Antony (Reference Antony, Brownstein and Saul2016), 175–178).

Other psychological and social influences, by contrast, are sensitive to the truth: think of expert dispositions developed from years of training. But even these are highly unlikely to be perfectly responsive to the way things are: experts are not immune from error in their domain of expertise. And in many cases experts commit certain errors because of their expertise. Given a single set of symptoms, a cardiologist may be biased towards cardiological explanations whereas an endocrinologist may favour endocrinological ones, but the real explanation may be nutritional. So even truth-sensitive psychological and sociological influences are prone to generating certain sorts of false belief even as they promote other true ones; I will call a belief brought about by an otherwise truth-sensitive influence a false-positive belief (Dormandy (Reference Dormandy2018b), 60).

Groupthink is worth special mention. This social influence is not truth-sensitive: whether your group believes something is not thereby an indication of its truth. A given case of groupthink is only truth-sensitive if the group's thought patterns are themselves truth-sensitive – as in a group of experts. But false-positive beliefs may arise here too. For example, the recorded proportion of diagnoses of a given disease in some countries is much larger than in others, when other indicators would predict similar proportions. Groupthink in the countries’ respective medical professions is arguably a culprit.

These three influences on our belief systems (evidence, psychological factors, and social ones) are largely implicit: they are active below our fully conscious awareness.Footnote ⁷ This holds especially for psychological and social influences, some of which are patently unconscious. But even evidence, though understood to be available on sufficient reflection (Conee & Feldman (Reference Conee and Feldman2004), ch. 4), is often implicit: we have too much evidence to reflect on simultaneously or even in sequence, and for many questions we never even try (ibid.). Because of this, we are often in the dark about the origin and justification of our beliefs, concepts, and values; even cogent reasoning conceals hidden influences. Many of these will not be truth-sensitive. Many others will be, but will nonetheless be fallible and even prone to generating false-positive beliefs. And many of the epistemically good and bad features of our reasoning will be shared by our community.

Three epistemological worries arise from these considerations.Footnote ⁸ The circularity worry says that a group of like-minded people is in a bad position to epistemically self-evaluate. Group members will be strongly disposed to conclude that their belief system is in epistemically good shape whether or not it is. One reason is that they will evaluate their belief system by its own epistemic standards, which will be biased to self-favour. Another reason is that groupthink makes it likely that most members will exhibit similar thought patterns and hence be in a bad position to offer independent criticism.

Second, what I'll call the distortion worry says that a group's belief system is prone to misrepresent some aspects of reality. One reason is that perspectives by nature reveal some things but occlude others. Another reason is given by the prevalence of truth-insensitive influences, which have a higher probability of distortion than truth-sensitive ones and may compete with them. A third is given by the likelihood that even truth-sensitive influences are apt to generate certain false-positive beliefs. Fourth, the circularity worry itself promotes distortion, for it hinders group members in becoming aware of any distortions or risks thereof in their belief systems, leaving them apt to build greater distortions upon them.

Whereas the circularity and distortion worries apply to reasoning about any topic, the exacerbation worry (as I'll call it) arises specifically with respect to reasoning about existentially and personally important issues such as religion. It says that, when such topics are at issue, the circularity and distortion worries are exacerbated: we become less adept at critically evaluating our own belief systems, and those belief systems are more likely to be subject to distortion. The reason is that thinking about these matters activates certain psychological and social defences. Religion creates an overall coherent picture of life and our place in it, so that we tend to fear (and thus be biased against) anything that may call ours into question (Solomon et al. (Reference Solomon, Greenberg and Pyszczynski1991) ). On top of this, a person's religious belief system tends to be tightly linked with his sense of identity. The circularity worry is thus worsened, because psychological and social factors are in full gear to promote positive evaluations of our own views and negative evaluations of alternatives. The distortion worry is worsened because, first, if we are in error, strong psychological and social influences make us more likely to stay that way; but second, even if we are largely on track, they will make us more susceptible to false-positive beliefs.

The epistemic-potential claim defended

I will now argue that engaging in disagreement can help mitigate these three worries. My argument will not, for the most part, differentiate between disagreement about religion and about other matters: engaging in disagreement is epistemically beneficial in general. The only way in which I will argue that religious disagreement can be of particular benefit is in mitigating the exacerbation worry, at least when the topic of religion is what caused the exacerbation. The first three benefits address the circularity and distortion worries, and the fourth addresses the exacerbation worry.

1. Engaging in religious disagreement can provide constructive external criticism of your belief system, of a sort that you would be ill equipped to generate alone. A disagreeing outsider is in a good position to notice unexamined auxiliary beliefs or concepts that miss important aspects of a phenomenon, or to call attention to values that your belief system ignores. External criticism can help mitigate the circularity worry by alerting us to aspects of our belief systems that we took for granted and thus failed to evaluate, and it can provide a less sympathetic source of evaluation with which to triangulate our usually favourable self-evaluations. Granted, we will evaluate others’ criticism through the filter of our own belief system, but such criticism nonetheless provides a different sort of datum to run through the filter. As for the distortion worry, external criticism can provide clues about what our belief system might miss or misconstrue, and in prompting us to nuance our belief systems, it can rein in truth-insensitive biases and the false-positive effects of truth-sensitive ones.

2. Engaging in disagreement can expand your evidential basis. It delivers, first, the evidence that someone disagrees with you, which, depending on certain epistemically relevant characteristics of your interlocutor (e.g. whether he is an epistemic peer, epistemically or personally admirable, and so forthFootnote ⁹), might prompt you to rethink some aspects of your belief system. Second, disagreement provides you with information about your interlocutor's evidence, which will probably differ from yours in two respects: it will arise from different types of event from those you have experienced (since she will probably travel in different social and cultural circles than you do), and her psychological and social influences will prime her to experience some of the same types of event differently than you do. Additional evidence of these sorts can help mitigate the circularity worry, for such evidence can be harder to accommodate within your belief system than evidence more typical to your community members, and hence can challenge facile self-evaluations. It can mitigate the distortion worry by offering correctives to incomplete (and thus potentially misleading) evidence of your own.

3. Engaging in disagreement can provide epistemic alternatives. In helping you understand your interlocutor's belief system and patterns of reasoning, it may create new conceptual or epistemic possibilities. This can mitigate the circularity worry, since alternatives can present other means of epistemic evaluation. And it can mitigate the distortion worry, since new epistemic alternatives can complement or improve on aspects of your own belief system or help you understand, by direct contrast, why your belief system is epistemically preferable.

4. Whereas the first three benefits address the circularity and distortion worries, the fourth helps mitigate the exacerbation worry. This problem (to recall) says that reasoning about religion strengthens the negative effects of our biases, thus exacerbating the other two worries. Disagreement can help here by supplying counterinstances (Saul (Reference Saul2012), 259). A counterinstance is a person or thing that does not correspond with what a bias would lead one to expect; exposure to counterinstances can help weaken bias and can improve one's ability to engage in nuanced and differentiated reasoning. For example, if fear of dogs biases us against them, then the more friendly dogs we meet, the less we will tend to fear dogs and the more we will assess the dangerousness of each dog individually. Similarly, the more intelligent and admirable people we meet among those who disagree with us about religion, the more likely these interactions are to defuse some of the defence mechanisms that close us off from constructive discourse. Of course, if our interactions with our interlocutors are adversarial, things could go the other way: our biases may be reinforced and our nuanced thinking diminished. But if our discourse is cordial and constructive, engaging in disagreement can work against the factors that exacerbate the circularity and distortion worries.

One might think that these epistemic benefits arise primarily from disagreements between epistemic peers;Footnote ¹⁰ if so, this would limit my argument (though how much it did so would depend on how demanding our notion of peerhood is). For epistemic peers are a mere subset of the people we might encounter – on some accounts of peerhood, a vanishingly small one (King (Reference King2012) ). In response to this concern, peerhood bears little relevance to my overall argument. An interlocutor need not be a peer to call attention to weak spots in your belief system, including values that it neglects. Nor does her not being a peer undermine the epistemic value of her evidence; indeed, if her evidence differs significantly from yours (which many definitions take to exclude peerhood), this would be an epistemic advantage, since (as I've argued) her evidence can supplement yours. Finally, someone need not be a peer in order for aspects of her belief system to open up alternatives that are invisible on yours.

The committed believer might also wonder how the epistemic openness needed for engaging in disagreement is compatible with the ‘error theories’ that religious belief systems sometimes deploy to explain the disagreement of nonbelievers; think of Plantinga's (Reference Plantinga2000b, ch. 7) claim that non-Christians are tainted by the noetic effects of sin, or Martin's (Reference Martin1990, ch. 6) claim, in support of atheism, that religious beliefs result from wishful thinking. Such error theories can add coherence to one's belief system, and there may be evidence (especially internal to the belief system in question) to think them true. But in order to reap the greatest epistemic benefits from disagreement, interlocutors must not wield these theories as ‘discrediting mechanisms’ (McKim (Reference McKim2001), 136) that excuse us from taking our interlocutor seriously.

But how can it be epistemically appropriate to maintain an error theory without using it as a discrediting mechanism? After all, if you reasonably believe, because of evidence internal to your belief system, that sin has noetic outworkings on your interlocutor, epistemic responsibility might seem to require discrediting his opinions. We must recognize in response that, first, even if an error theory is true, it does not follow that it explains every disagreement you encounter – a given disagreement might arise from genuine vulnerabilities in your view. For example, even if atheists are prevented from perceiving signs of God in the world, they can and do perceive horrendous evil and construct powerful arguments from it. Second, even if your interlocutor's disagreement is explained by (say) the noetic effects of sin, it does not follow that she is not on to something or that she has not really spotted a weakness in your view. After all, false beliefs can rest on good evidence and true ones on bad evidence. Third, even if your interlocutor's criticism is mistaken (due to the noetic effects of sin or some other cause), engaging in disagreement with her can still yield important epistemic benefits. For example, it can show you that your belief system stands to be better articulated and supported.

Finally, one might object that my argument falters because of the differences in epistemic standards from one religious belief system to another. Recall that religious belief systems include auxiliary beliefs about epistemic standards. Unlike science, which relies on intersubjectively available evidence and on premises that do not beg the question of one's interlocutor, there are few if any interreligious epistemic standards that two parties disagreeing about religion even could employ. Is not the whole endeavour of engaging in religious disagreement doomed from the start? It is not. First, one might engage in disagreement within one's own tradition. Second, and more importantly, the fact that there are few tradition-external standards is a major reason for epistemically valuing religious disagreement. Comparing and contrasting the deliverances of markedly different epistemic standards themselves will help interlocutors take a critical view on their own, where such a critical view – precisely because there is no interreligiously accepted method – would otherwise be hard to foster.

The already-accurate objection

The committed believer may grant that the epistemic-potential claim applies to many disagreements, but she might deny that it applies to religious disagreement – at least, when the religious belief system in question is hers. For her religious belief system, she will say, is mostly accurate already. The best way for her religious community to achieve epistemic aims is thus to sit tight with its belief system as it has been passed down. Disagreement is not only epistemically otiose in her case, but threatens to mislead or confuse.

But even if her belief system is accurate, sitting tight is not a viable option for the committed religious believer. For religious communities exist in times and cultures, and these bring constant epistemic challenges and opportunities (Wynn (Reference Wynn, Abraham and Aquino2017), 126–127). As history advanced, for example, societies developed destructive political ideologies, many of which infiltrated religious belief systems. Because of the less than fully conscious way in which epistemic influences work (see above), this can catch religious communities unaware. A dissenting outside perspective can help call attention to worrying trends like this. Times and cultures also bring new insights that the historical founders of one's belief system may not have been in a position to appreciate. Think for example of human rights, or of scientific developments which – when accepted by religious communities – led to a deeper understanding of the relationship between divine and physical reality. Disagreement with those outside your religious community can help you assimilate positive developments that ultimately benefit your religious belief system. Even largely accurate religious belief systems, then, must guard against infiltration by outside error and watch for positive outside developments. Disagreement can help them do both.

The proper-function objection

The committed believer may still deny that the epistemic-potential claim applies to her religious belief system. For her belief system (she may say) is powered by divine inspiration, insulating it against the malign outside epistemic influences mentioned above and helping it assimilate new ones – without needing disagreement.Footnote ¹¹

I cannot discuss all accounts of divine epistemic inspiration, so I will focus on Plantinga's influential sensus divinitatis account (Plantinga (Reference Plantinga2000b), 171). The sensus divinitatis is a cognitive faculty; it resembles a sensory mechanism, but instead of yielding sensory experiences, it yields experiences of God's traces in the world, including accurate intellectual seemings concerning religious matters (ibid., 172–174). These experiences non-inferentially give rise to religious beliefs, including the belief that God exists. That not everyone believes in God might seem to refute this view. But Plantinga, as we saw, has an explanation: the noetic effects of sin cause everyone's sensus divinitatis to malfunction (ibid., 203, 212), and the Holy Spirit repairs those of some but not all (ibid., 278).

The committed believer may argue that the person with a properly functioning sensus divinitatis stands to gain nothing from religious disagreement. First, such a person does not need disagreement to reveal whether his belief system has absorbed malign epistemic influences from new contexts: the sensus divinitatis will prevent this from occurring to begin with. Second, he does not need disagreement to reveal new insights: the sensus divinitatis provides all the inspiration he needs.

To respond, let us look more closely at the cognitive profile of the believer with a properly functioning sensus divinitatis. We may divide his experiences into two (mutually exclusive) subsets. The first, which I'll call SD experiences, are those which his properly functioning sensus divinitatis helps bring about. These experiences purport mainly to represent religious aspects of reality and can include intellectual seemings. I take it that religious believers also have other experiences, without religious content or character, that arise from the usual sensory and intellectual mechanisms, and that these mechanisms do not need the help of a properly functioning sensus divinitatis to deliver accurate experiences. I will call this second subset mundane experiences. Both SD and mundane experiences provide experiential evidence, which is non-inferential, and on its basis beliefs, from which inferences can be drawn.

With this picture in mind, let's return to the first of the proper functionalist's reasons for denying that religious disagreement can benefit believers whose sensus divinitatis functions properly: that the properly functioning sensus divinitatis keeps out epistemically malign influences.

To assess this claim, let's consider each factor that might influence a belief system undergirded by a properly functioning sensus divinitatis. I'll start with evidence, focusing on experiential evidence, which I have just subdivided into SD and mundane experiences.

As for SD experiences, we will grant the objector that, due to their origin in the sensus divinitatis, these exert an epistemically good influence. They provide non-inferential bases for many true religious beliefs and for at least some true auxiliary beliefs (e.g. about scriptural hermeneutics; Plantinga (Reference Plantinga2000b), 203–204). Although Plantinga only discusses the way in which the properly functioning sensus divinitatis produces religious beliefs, we may grant the objector that this mechanism also promotes the formation of useful and accurate religious concepts. SD experiences might even shape the believer's values in a way that yields important religious insights. It is conceivable that a person's SD experiences might be affected by unreliable psychological or social influences, but for argument's sake we will charitably grant the objector that the sensus divinitatis prevents this. So far, so good for the proper functionalist.

But let's turn to the evidence of mundane experiences. Consider their influence on religious beliefs. Mundane experiences are unlikely to have any direct evidential impact on religious beliefs, since these beliefs will be based largely on SD experiences. But mundane experiences can still exert an indirect evidential influence on religious beliefs: they can influence auxiliary beliefs that lack religious content, such as those pertaining to metaphysics, epistemology, or empirical matters. Mundane experiences can also influence concepts that are not religious as such, such as those of love, justice, or forgiveness. And they can influence the believer's values; for example, someone whose parents emphasized punishment over compassion may grow up valuing justice over mercy. Since mundane experiences are not governed by the sensus divinitatis, there is no guarantee that they will be accurate. So there is a live possibility, even with a properly functioning sensus divinitatis, that inaccuracies will creep into one's religious belief system indirectly, via mundane experiences. This of course is problematic for the proper-functionalist.

What about psychological and social influences on religious belief systems? Recall that these can be direct or indirect, where direct ones bypass evidence altogether, and indirect ones operate through evidence.

Consider first the direct influence of psychological and social factors. Some may be epistemically good. For example, an agent's religious affections may directly bring about true religious beliefs. But direct psychological and social factors can also exert negative influence, for example when the stereotypes with which the believer grew up cause him to conceive of nonbelievers as (say) inferior. It is hard to see how the sensus divinitatis could shield a religious belief system from epistemically malign direct psychological or social influences, since this faculty's remit is limited to generating accurate SD experiences.

Let's turn to indirect psychological or social influence. When it comes to SD experiences, it is reasonable to suppose that the sensus divinitatis shields them from epistemically undesirable influences. But mundane experiences pose a problem for the proper functionalist. Since these experiences are not within the remit of the sensus divinitatis, this faculty cannot guarantee that their epistemic influence on the believer's mundane experiences will be good. Once more, inaccuracies in one's belief system are a live possibility.

In summary, epistemically malign influences can affect even a belief system that is underwritten by a properly functioning sensus divinitatis, by the influence of mundane experiences and indirect psychological and social workings on them, and by the influence of direct psychological and social factors. Since belief systems are interlocking networks, any inaccuracies will probably spread through them. And since individuals together engage in groupthink, at least some inaccuracies will probably infect the religious community – even a community powered by inspired cognitive functioning.

The proper functionalist might resist this claim by expanding the remit of the sensus divinitatis beyond Plantinga's account. Rather than claiming that this faculty is responsible only for generating accurate experiences of God's traces in the world, she might construe it as a global faculty that regulates all of our cognitive mechanisms.

But this ‘global sensus divinitatis’ hypothesis cannot be maintained. It is ad hoc, in at least the following three (negative) senses. (a) It is subject to abuse, since there are few constraints on it (Boudry (Reference Boudry2013), 6): religious believers might suppose that anyone who disagrees with them must be mistaken. (b) It is highly implausible on general background knowledge (Howson & Urbach (Reference Howson and Urbach1993), 157–158). For it predicts the existence of some group of religious believers with generally accurate experiences and intellectual seemings, not only on religious topics, but on non-religious ones too. Yet this prediction falls flat: no religious group appears to have higher epistemic success rates in maths, humanities, science, and so forth. (c) The ‘global sensus divinitatis’ hypothesis is highly implausible even given most religious belief systems themselves (McGrew (Reference Mcgrew2014), 495). Religious scriptures and traditions do not tend to teach that believers are generally cognitively superior to nonbelievers (and indeed Plantinga's Calvinist tradition teaches that nonbelievers can be just as reliable on non-religious matters as believers; Moroney (Reference Moroney1999), ch. 1).

Another move for the proper functionalist would be to concede that individual believers might be susceptible to epistemically malign influence, but deny that this poses a problem for the religious community, let alone one that disagreement would help alleviate. A robust religious doxastic practice (Alston (Reference Alston1991) ), she might argue, should prevent outside errors from spreading through the community. But this response will not do. For there is no reason to think that epistemically malign influences will limit themselves to isolated individuals; on the contrary, groupthink phenomena typically yield the opposite effect.

So even when a believer's sensus divinitatis functions properly, his religious belief system remains vulnerable to epistemically undesirable influences that this faculty does not control. Disagreement can ameliorate matters.

Let us now consider, more briefly, the proper functionalist's second reason for denying that disagreement can epistemically benefit a belief system underwritten by a properly functioning sensus divinitatis: that this faculty will home in on epistemically desirable features of one's context without the help of disagreement. The objector is right that disagreement is not necessary for noticing and appropriating epistemically good developments from outside one's belief system, at least as long as those developments are compatible with it – though it is unclear how one would naturally discover such developments if engaging in disagreement with outsiders is discouraged. What about developments that are incompatible with aspects of one's belief system? There seems little other way for a religious believer to appropriate these than to engage in disagreement over them. But more importantly, it is not clear what the objector stands to gain by claiming that the properly functioning believer does not need recourse to disagreement. If the sensus divinitatis equips believers to sense traces of the divine in a starry sky (Plantinga (Reference Plantinga2000b), 172), then why not trust it to sense them in some interlocutor's disagreeing viewpoint?

I conclude that, even if you have a properly functioning sensus divinitatis, epistemic infelicities can creep into your religious belief system, and disagreement can help uproot them. Moreover, there is good reason to welcome religious disagreement as a source of possible epistemic benefit, confident that this faculty will help you see what religiously relevant insights it might yield.

The objection from religious commitment

The objector may grant that divine guidance leaves room for inaccuracy, even on religious matters. And she may grant that disagreement can help root it out. But she may still deny that fostering disagreement, as opposed to suppressing it, is the best epistemic move for a religious community.Footnote ¹² The reason is that engaging in disagreement may threaten the believer's commitment to his religious belief system. If our only two options are cleaving to a largely accurate if imperfect religious belief system or destroying believers’ commitment by striving for perfection, then we should surely pick the first option.

There are two dangers that the objector will highlight. The first is normative. Religious disagreement, she will say, is likely to furnish counterevidence that threatens the epistemic status of your religious or auxiliary beliefs. Indeed, McKim endorses something like the epistemic-potential claim, yet draws from it a pessimistic conclusion about the epistemic status of religious beliefs: because much of the evidence provided by disagreement speaks against one's religious view, religious beliefs are unjustified unless held tentatively (McKim (Reference McKim2012), 155–156). And tentative belief, the objector will note, is not committed belief. So it would seem that engaging in religious disagreement will force a choice between the believer's epistemic status and his epistemic commitment.

But this worry can be dismissed. The epistemic-potential claim entails nothing about epistemic justification or any other kind of epistemic status. It is compatible with an evidentialist view, according to which positive epistemic status requires respecting your evidence, or a fideist view, on which positive status can tolerate a measure of belief ‘beyond’ your evidence (James (Reference James1921); Bishop (Reference Bishop2007), 107–112). More than this, even evidentialism need not threaten the epistemic status of religious beliefs in light of disagreement. For evidentialism is a general claim, and there are many ways in which it can be spelled out (Dougherty (Reference Dougherty, Callahan and O'Connor2014), 101). For example, the evidentialist epistemologies of Plantinga (Reference Plantinga2000b) and Dougherty (Reference Dougherty, Callahan and O'Connor2014) assign a large role to the evidence of one's own experiences (including religious ones) and to the beliefs of one's religious tradition, over and above the evidence of intersubjective arguments; this will tend to allow religious beliefs to remain justified even given the counterevidence supplied by disagreeing interlocutors (Dormandy (Reference Dormandy2018b)). On the other hand, the evidentialisms of Schellenberg (Reference Schellenberg2007, 190), McKim (Reference McKim2001, ch. 10), and Philipse (Reference Philipse2014, pt I) assign greater weight to intersubjective evidence (including the opposing viewpoints of disagreeing interlocutors), which may make positive epistemic status for religious beliefs more elusive. So it is not McKim's endorsement of the epistemic-potential claim that yields his pessimism about the epistemic justification of religious beliefs; it is the strict form of evidentialism that he endorses alongside it. The epistemic-potential claim by itself entails no position on epistemic justification or status more broadly. It does not matter for my purposes which epistemic norms committed believers abide by, as long as those norms acknowledge the epistemic goods outlined above.

The second worry is that engaging in disagreement may psychologically undermine the believer's commitment to her belief system, by shaking her confidence (Plantinga (Reference Plantinga, Meeker and Quinn2000a), 189). In response, we must acknowledge that disagreement can pose psychological and therefore (assuming that one's belief system is accurate) epistemic risks. But I contend that discouraging it is riskier still. First, foregoing disagreement forfeits you the benefits that it can bring (see above). Second, discouraging it may court the suspicion that the religious belief system that you are protecting would not hold up to outside scrutiny – and this is a greater, if more subtle, psychological threat to epistemic commitment than the sanguine encouraging of disagreement would be. Third, discouraging disagreement can create the impression that acceptance and approval in your community are linked to epistemic conformity. This might prompt some to remain epistemically committed in spite of nagging doubts – but such commitment is unhealthy, maintained for fear of rejection rather than for good epistemic reasons. And unhealthy commitment is epistemically dangerous: such believers risk forming a warped view on which God values strict doxastic obedience more than understanding powered by freedom. Hence discouraging disagreement runs epistemic risks of its own. I contend that the risks of engaging with disagreement are smaller, and the epistemic rewards greater, than those of eschewing it.

One might object, from the other direction, that I am conceding way too much to committed religious believers. Am I not letting them uncritically ring-fence as much of their belief system as they please, and does this not contravene the whole point of engaging in disagreement? My first response is that nothing about my approach prevents people from changing their religious belief system, even drastically, should an earnest dialogue lead them to regard this as the epistemically and personally best course. Second, commitment to a belief system, as I have said, is compatible with critical engagement with many of its components. Third, insisting (like McKim (Reference McKim2001), 142) that you should be prepared to abandon your religious belief system is epistemically counterproductive: it will put off committed religious believers from engaging in any disagreement, thus forfeiting them any of its benefits.