Peer disagreement

In recent years, a thriving debate has taken place around the question of what rationality requires in situations of an apparent peer disagreement. A peer disagreement is simply defined as a disagreement between epistemic peers (that is, between epistemic peers regarding the disputed issue), given a standard definition of the latter, according to which an epistemic peer of some person A regarding a certain issue fulfils both of the following two conditions. First, she has the same or equally good evidence as A regarding a certain issue; and second, she is equally competent as A in making judgements on the basis of the kind of evidence in question.Footnote ¹

In order to avoid reiterating some of the often-cited examples circulating in the debate, I ask the reader to imagine the following scenario, which may also highlight the possible practical aspect of the matter. Imagine two gamekeepers responsible for leading two groups of elephants across the Nile river delta, where each of the two gamekeepers is responsible for her own group of elephants. Given the geography of the terrain, the task involves leading the elephants across numerous arms of the river. Given that some of the bridges they encounter are stable enough not to break under the weight of the elephants, while others are not, an important aspect of the gamekeepers’ task consists in judging whether to use a given bridge or whether to search for the next one. Imagine further that the two gamekeepers think that they have an equal past record in making this kind of judgement, that is, the percentage of instances in which they have been right is roughly equal, and the two are aware of this. However, at some point during their journey, gamekeeper A reaches a bridge that, after some inspection, she considers as stable enough not to break under the weight of her elephants. Gamekeeper B, who has already inspected the bridge and has come to a different conclusion, sees that A is about to lead her elephants across the bridge, and returns to her in order to tell her that she deems it not to be stable enough. The two have examined the bridge equally carefully, thus basing their opposing judgements on the same evidence (given, of course, some relatively loose conception of the identity of evidence). Furthermore, this particular occasion is one in which none of the two appears any more tired, drunk, etc. than the other. The question relevant regarding the debate on peer disagreement is now the following: to what extent is A, in the face of the fact that B disagrees with her regarding whether the bridge would hold, rationally required to lose confidence in her initial opinion that the bridge would hold?

David Christensen (Reference Christensen2007) argues for the claim that A is rationally required not only to lose confidence in her initial judgement to some significant degree, but even to give B's judgement (roughly) the same epistemic weight as her own, such that the rationally required degree of confidence in the proposition that the bridge would hold lies in the middle between A's and B's initial degree of confidence. Accordingly, this position has been coined the ‘Equal Weight View’. (See Elga (Reference Elga2007) and Cohen (Reference Cohen, Christensen and Lackey2013) for further defences.) The view applies to the gamekeeper scenario in virtue of the latter's fulfilment of the following three conditions. First, A regards B as possessing equally good evidence as herself regarding the truth or falsity of the proposition in question (i.e. the proposition that the bridge in question would hold). Second, A regards B as being in general as competent as herself in making judgements on the basis of the kind of evidence in question. And third, A regards the situation in question as one in which B's cognitive abilities are no more impaired by being tired, drunk, etc. than her own. The Equal Weight View applies if and only if all of these three conditions are met. As already indicated, the view consists in the claim that in the kind of situation in question, the rationally required degree of confidence in the truth of the proposition in question, that is, the degree of confidence one should arrive at when getting to know of the disagreement, is the arithmetic mean of the initial degrees of confidence of the two interlocutors. More precisely, it is the arithmetic mean of one's own initial confidence and the initial confidence one believes one's interlocutor to have had. Thus, if A had an initial degree of confidence of about 0.9, and she believes B to have had an initial degree of confidence of about 0.3, then the recommended degree of confidence after getting to know of the disagreement is about 0.6.Footnote ²

The major rationale behind the view is the following symmetry consideration: given the fulfilment of the three conditions mentioned above, why should A consider the possibility that B is wrong in this particular case as any more probable than the possibility that A herself made a mistake in assessing the evidence? According to defenders of the Equal Weight View, favouring one's own initial opinion over that of the disagreeing peer simply on the basis of the very reasoning that led to one's initial opinion is not rationally permitted. See, for instance, Cohen:

Given these considerations, according to which the reasoning on the basis of which one is rationally permitted to favour one's own initial opinion over that of the disagreeing peer cannot be part of the reasoning through which one had arrived at one's initial opinion, the situation may indeed seem to exhibit the kind of symmetry in question.

It should not be surprising, however, that the correctness of the view has not been unanimously accepted. According to Wedgwood (Reference Wedgwood2007, 261ff.), for instance, the kind of situation in question is one in which one is always rationally permitted to have an ‘egocentric epistemic bias’ in favour of one's own initial opinion. Kelly (Reference Kelly, Feldman and Warfield2010) and (Reference Kelly, Christensen and Lackey2013) proposes a different alternative, the so-called ‘Total Evidence View’, according to which the rationally required degree of conciliation depends on which of the disagreeing peers evaluated the initial evidence correctly, such that the person who did evaluate the initial evidence correctly is permitted (and even required) to give less than equal (but still some) weight to the interlocutor's opinion, while the person who did not evaluate the initial evidence correctly is required to give more than equal weight to the other's opinion.Footnote ³

Despite the differences among the positions just mentioned, there is what may be called a consensus that, at least in standard kinds of peer disagreement,Footnote ⁴ including the one described above, one is rationally required to lose at least a bit of one's confidence in one's initial opinion.Footnote ⁵ This latter view has, following Elga (Reference Elga, Feldman and Warfield2010), been called the ‘Conciliatory View’ (CV). It applies to the same kind of case as the positions mentioned above (though proponents of CV are of course also committed to the same claim regarding cases of disagreement with supposed epistemic superiors; and given the relative inclusiveness of the formulation ‘at least a bit’, there seems to be no good reason not to apply the view also to at least some cases of disagreement with supposed epistemic inferiors). Christensen describes endorsing CV as no more than

Christensen even goes on to describe CV as ‘little more than epistemic common sense’ (ibid.).

The problem I will be concerned with in the present article is the fact that even though the debate on peer disagreement has been quite vivid in recent years, a different kind of disagreement has hitherto been neglected – namely, cases of disagreement in which the two parties do not have the same or equally good evidence, and do thereby not count as peers regarding the issue in question. The practical impact of the fact just stated concerns especially those cases of disagreement in which the interlocutors do not only base their opinions on different pieces of evidence, but in which it is not even possible to share the respective evidence by conveying the relevant kind of information to each other. If, for instance, in the above-described scenario of the two gamekeepers, a certain bridge has only been inspected by A, and not by B, but A nonetheless wishes to hear B's informed opinion about whether the bridge would collapse under the weight of A's elephants, then A can (at least in principle) describe to B all the relevant aspects of the construction of the bridge, thereby transforming B into a peer regarding the issue at hand. And even if A's inspection depends on some hard-to-describe tactile or visual impressions, B could still just travel back to inspect the bridge by herself. In contrast, there are cases of disparity of evidence, and even cases of disagreement involving disparity of evidence, in which the relevant evidence cannot be shared, and which thereby cannot be transformed into peer situations, no matter how cooperative the disagreeing parties are.

In the present article, I will be concerned with a paradigm example of this practically troublesome kind of disagreement – namely, with cases of disagreement in which the opinion of at least one of the parties is based on a private religious experience. For a recent statement of the supposed inapplicability of arguments in the debate on peer disagreement to the kind of disagreement in question, see for instance Bogardus:

From now on, I will use the term ‘deep disagreement’ to denote the practically troublesome kind of case described above – namely, cases of disagreement in which the relevant evidence cannot be shared, no matter how cooperative the disagreeing parties are.Footnote ⁶ In the following section, I will specify the kind of religious experience involved in deep religious disagreements.

Deep religious disagreement

The term ‘religious experience’ may be used to denote experiences of crucially different sorts. Not all of them are such that the respective disagreement has to be regarded as a deep one. That is, the information gathered by having some so-called religious experiences may just as well be conveyed by (convincingly) reporting the experience to someone else – analogously to the above-described situation in which one of the two gamekeepers informs the other one about all the relevant aspects of the construction of a certain bridge. Among the many reports about religious experiences cited by William Alston in his influential book Perceiving God, the following may serve to exemplify the kind of religious experience in question:

Imagine that the reported event turned the protagonist into a believer, mainly basing his or her belief in the existence of God on the experience described. Imagine further that the protagonist subsequently enters into a conversation with someone disagreeing with him or her about the existence of God. Does the fact that only the protagonist, but not the interlocutor, had the kind of experience in question have to prevent the two from talking as peers about the issue? No. Rather, the information gathered by the protagonist by having the experience in question can simply be conveyed to the interlocutor by reporting the experience. The two might thereby end up sharing all the relevant evidence (given that the interlocutor believes in the reported auditory event).

Compare this with the following report cited by Alston:

Imagine, analogously to the scenario contemplated above, that the protagonist mainly bases his or her belief in God on the experience in question, and subsequently enters into a conversation with someone disagreeing with him or her about the existence of God – someone who did not have this kind of experience. Can the conversation be transformed into a peer situation if the protagonist simply reports the experience to the interlocutor? No. In contrast to the above-described case involving an auditory experience, reporting the experience in question does not convey all the relevant information gathered by having the experience. Rather, while in the case involving an auditory experience the interlocutor can imagine the event concerning all relevant aspects, reporting the experience in the second case would leave the interlocutor puzzled, unable to grasp what the protagonist is really talking about. Obviously, this difference between the two cases is due to the fact that in the second case, the experience in question is being had through a perceptual channel allegedly foreign to those who did not have experiences of the same peculiar kind.

I call those experiences that are falling – like the one figuring in the second case – under none of the standard senses ‘perceptually unique’. One criterion for the deepness of religious disagreements has thus been identified: the disagreement is a deep one if some of the experiences on which at least one of the opinions is based are perceptually unique, where the interlocutor has not had the same kind of experience and can therefore not wilfully reproduce it in imagination. I call this criterion the ‘uniqueness criterion’.

The uniqueness criterion is sufficient for the deepness of a religious disagreement. For all that has been said so far, though, this criterion may not be a necessary one, for there may still be other reasons for the deepness of a given religious disagreement – that is, other reasons for there being an irresolvable disparity of evidence regardless of the degree of co-operation by the disagreeing parties. However, in the present article, I will merely be concerned with the question of how to deal with those deep religious disagreements that are deep due to their fulfilling the uniqueness criterion. Accordingly, in the following section, I will start my investigation of how to deal with this kind of deep religious disagreement. I will start this investigation by introducing a new notion of peerhood that will be applicable even to the troublesome kind of case in question.

Meta-peerhood

In what follows, I shall call the party to the disagreement who (at least partly) bases the religious belief in question (e.g. the belief in the existence of a divine being) on a perceptually unique experience the ‘experiencer’. And I shall call the other party to the disagreement, who has not had the kind of experience in question, the ‘sceptic’. The crucial questions to be asked here are the following: Are the two parties rationally required to be conciliatory? If so, why? And if not, why not? As already stated, the debate on peer disagreement, which has been thriving in recent years, stays silent regarding these questions. For as also already stated, the disagreement in question is not a peer disagreement (since it violates the equal-evidence condition), and also cannot be transformed into one, regardless of the degree of cooperativeness of the interlocutors. That is, it is what I have called a deep disagreement. (Of course, the parties to the disagreement may falsely believe that the situation is that of a peer disagreement – namely, if one of them is dishonest regarding the evidence on which the respective opinion is based. In this kind of case, the views sketched above in the opening section indeed recommend a certain degree of conciliation. But I take it that the practically more interesting kind of case is that in which both parties are cooperative regarding the disclosure of the evidence on which their opinions are based. At least, this is the kind of deep disagreement with which I am concerned.)

In the present kind of case, the sceptic might be tempted not to be conciliatory, due to a dismissal of the perceptually unique experience in question. And on the other hand, the experiencer might be tempted not to be conciliatory, due to a belief that the sceptic, who has not had the kind of experience in question, is thereby not well informed enough in order to count as competent regarding the issue at hand. Given that none of the views on peer disagreement applies to this kind of case, are the two parties rationally permitted to be steadfast for the reasons just mentioned?

I will now argue that there is still a way to apply the notion of peerhood to the kind of case in question, so as to be able to apply some of the considerations from the debate on peer disagreement to at least a subset of the kind of case in question. The way to do this is the following. Imagine two atheists who have often discussed religious matters and who found themselves to agree most of the time. In particular, imagine that each of the two regards the other as equally competent when it comes to judging whether a given piece of evidence regarding the respective religious question is a relevant piece of evidence, that is, whether it is a piece of evidence that should be given considerable weight in answering the respective question. Imagine further that someday, one of the two has a perceptually unique experience as of a divine being, which makes her turn into a theist. Her interlocutor, the sceptic, first hears about the other's experience, and then gets to know that this experience has turned the experiencer into a theist.

The crucial part of this story concerns the fact that until getting to know about the disagreement,Footnote ⁷ the two parties regard each other as equally competent in judging whether a given piece of evidence is relevant regarding the respective religious matter. To be a bit more precise, they regard each other as equally competent in recognizing relevant evidence regarding religious matters when having the evidence. Even though, right beforeFootnote ⁸ getting to know about the disagreement, they already know that they now base their religious views on different bodies of evidence, and therefore no longer regard each other as peers, the fact just stated still constitutes a relation of similar epistemic significance. That is, even though, right before getting to know about the disagreement, they no longer regard each other as peers, they still have no reason not to regard the other as equally competent at the meta-level of recognizing relevant evidence when having it. Accordingly, I call those equally competent in recognizing relevant evidence regarding a given subject matter ‘meta-peers’ regarding this subject matter.

What consequence arises from the fact that even though, right before getting to know about the disagreement, the two parties do not regard each other as peers, they still regard each other as meta-peers? In the following section, I will specify the degree of conciliation rationally required in the face of this latter fact.

Conciliation in deep religious disagreement

In order to specify the degree of conciliation rationally required in cases in which the two parties in a deep religious disagreement regard each other as meta-peers, it may be helpful first to recall why conciliation appears to be rationally required in cases in which two disagreeing parties regard each other as peers concerning the disputed matter before getting to know about the disagreement. In the latter kind of case, the supposed peerhood rationally forces the subject to acknowledge the possibility that it is herself whose initial opinion on the disputed matter results from a faulty assessment of the evidence. In cases of supposed meta-peerhood in a deep religious disagreement, matters are different. The sceptic, for instance, in regarding her interlocutor as a meta-peer, is not thereby rationally forced to acknowledge the possibility of a faulty assessment of the evidence. For the crucial piece of evidence – namely, the religious experience her interlocutor has had – is one that she has not had a chance to assess at all. Rather, the situation is as follows. In regarding her interlocutor as a meta-peer – that is, in regarding her as equally competent in recognizing relevant evidence concerning the issue in question – the sceptic is prima facie rationally required to expect (or at least to acknowledge the possibility) that if she had had the religious experience, she would have given it the same weight as given by the experiencer and would thereby also not have dismissed it as bogus.

Does this already mean that the sceptic is rationally required to be conciliatory? No. The mere fact that one expects that a certain experience would lead one to adopt a certain opinion does not by itself rationally force one to have a credence higher than 0 in this latter opinion, because the experience in question might be notoriously misleading. Imagine, for instance, a hypnotist who is known for being able to make anyone believe some absurd proposition p by speaking about p to the respective person in a certain tone of voice. Obviously, one's acknowledgement that the hypnotist's voice would make oneself believe p does not at all rationally force one to have any credence higher than 0 in p. Analogously, being rationally forced to expect that having the religious experience in question would lead one to adopt the opinion of the disagreeing meta-peer by itself does not at all rationally force the sceptic to be conciliatory.

Before proceeding, the following remark about the debate on peer disagreement is in place. The theories that have been proposed in the course of this debate are not to be understood as theories about the degree of conciliation that will lead one to a degree of credence in p that is overall rational. Rather, these theories are merely to be understood as theories about the right reaction to the specific piece of evidence constituted by the occurrence of the disagreement. Whether the resulting credence in p is overall rational depends on whether one's initial credence in p was overall rational. That is, theories about peer disagreement are supposed to specify the degree of conciliation that is overall rational if one's initial degree of credence was already overall rational. In those cases in which one's initial degree of credence in p has been arrived at irrationally, one's reaction to the occurrence of the disagreement cannot result in an overall rational degree of credence anyway. Rather, it can at best not worsen the irrationality of one's degree of credence in p. Accordingly, Cohen, in his defence of (something close to) the Equal Weight View, notes that the view ‘simply tells me not to make a bad situation worse. My .8 credence is irrational on e, and to simply remain at .8 in the face of Peer being at .2 would make me even more irrational’ (Cohen (Reference Cohen, Christensen and Lackey2013), 111). Relatedly, the rationally required reaction to the occurrence of a disagreement should be regarded as independent, not from the overall rationality of the resulting degree of credence, but from the rationality of one's initial beliefs.Footnote ⁹

The point just made regarding the debate on peer disagreement has the following implication for the correct theory of meta-peer disagreement. Such a theory cannot specify the credence in a given proposition that is overall rationally required in the face of a meta-peer disagreement. Rather, it can take into account the initial degrees of credence of the subject and then specify what kind of reaction to the occurrence of the disagreement does not reduce the overall rationality of the subject's belief. One of the initial degrees of credence of the sceptic is her degree of credence in the proposition that the kind of religious experience in question misleads those who make this experience (like the hypnotist's voice). I call this proposition ‘m’. The rationally required reaction to the occurrence of the religious disagreement with the experiencer depends on the sceptic's initial degree of credence in m. Whether her initial degree of credence in m is rational is an issue about which a theory on meta-peer disagreement has nothing to say. (After all, the probability of m is mainly an empirical matter.)

Thus, the following scenarios are possible. First, the sceptic might hold the initial belief that the kind of experience in question is analogous to the hypnotist's voice, in having a judgement-distorting effect on anyone who has the experience. Given this initial belief, the meta-peer situation is one in which the sceptic is rationally required not to be conciliatory. For given her initial belief about the misleading character of the kind of religious experience in question, her expectation that she would react to this kind of experience in the same way as her interlocutor merely amounts to the expectation that she would be misled when having the experience.

Second, the sceptic might hold the initial belief that all her meta-peers are reliable in assessing the kind of perceptually unique experience in question. (It might seem odd both to be a religious sceptic and to think so positively of the epistemic value of the kind of experience in question. Well, one may just imagine that the sceptic believes that none of her meta-peers has ever had the kind of experience in question, and that she is willing to abandon her religious scepticism when some of her meta-peers credibly report having had the kind of experience in question.) This belief that the kind of experience in question is truth-conducive for her meta-peers, together with her initial belief that the experiencer is a meta-peer, rationally requires her – when getting to know that the experiencer has become a believer in reaction to having had the experience – to acknowledge the possibility that the experiencer, in having had the religious experience, has an advantage over her when it comes to assessing the disputed issue. Thus, this distribution of initial credences is one in the face of which the rationally required reaction to the disagreement is to be conciliatory to some significant degree.

Finally, the third kind of scenario to be mentioned covers all those cases that lie somewhere between the two extreme cases just described – that is, cases in which the sceptic merely acknowledges the possibility that the perceptually unique experience in question would be truth-conducive if one of her meta-peers had it. These cases are arguably such that the rationally required reaction to the disagreement is to revise one's initial opinion at least a bit.

The second of the three above-described kinds of scenario (the one that calls for some significant degree of conciliation) might as well be described as one in which the sceptic regards the experiencer as an expert (relative to herself) regarding the issue in question – in virtue of regarding her as possessing more non-misleading evidence than herself regarding the issue in question while not regarding her as any less competent than herself in evaluating this evidence. This kind of scenario thereby carries with it all the theoretical problems discussed within the philosophical literature on expertise. One of these problems occurs in an extended version of the kind of scenario in question – a version in which the sceptic does not only meet an experiencer who takes the kind of experience in question as good evidence for theism, but also someone who claims to have had the very same kind of experience while claiming that the latter is not at all good evidence for theism. Given that the sceptic has hitherto regarded both of these persons – the converted and the non-converted experiencer – as meta-peers regarding the issue in question, the situation exemplifies what Goldman (Reference Goldman2001) has called the ‘novice/two-experts problem’. This problem consists in the sceptic's being confronted with two experts disagreeing with each other, and in the corresponding fact that the sceptic's status as a layperson (that is, as a non-expert) makes it particularly difficult for her to assess the experiencers’ testimonies. Goldman examines several ways for a layperson to adjudicate between the opposing experts (cf. ibid., 93ff.), some of which (i.e. searching for additional experts, consulting meta-experts regarding the alleged experts’ expertise, and investigating possible biases on the part of the alleged experts) are applicable to the case of the sceptic and the two disagreeing experiencers. At any rate, unless these ways of adjudicating clearly speak in favour of the non-converted experiencer, the rationally required way for the sceptic to adjust her credences in this case is – as in the third kind of scenario mentioned above – to revise her initial opinion at least a bit.

Concentrating on the simpler one-novice/one-expert relation, the contemporary philosophical literature on expertise hardly contains any disagreement on whether regarding someone as an expert (relative to oneself) gives one a reason to defer to that person, or, for that matter, a reason significantly to revise one's opinion in the face of disagreement with that person. A rare recent defence of the claim that expert testimony should not be trusted can be found in Mizrahi (Reference Mizrahi2013). According to the latter, ‘[there] is a growing body of research which shows that experts are wrong more often than one might expect’ (ibid., 63), which in turn shows that ‘the fact that an expert says that p does not make p significantly more likely to be true’ (ibid., 58). However, the empirical data Mizrahi relies on regarding the surprising degree of fallibility of those who are given the label ‘expert’ regarding certain professions – ‘academics, economists, policymakers, and journalists’, ‘clinical psychologists’, ‘attending physicians’, etc. (ibid., 64) – merely suggest that the term ‘expert’ is generally used in an inflationary manner, and that being a professional should not be conflated with being an expert. That is, the data merely suggest that many professionals have turned out not to be experts concerning the respective subject matter. Accordingly, Seidel, providing a detailed critique of Mizrahi's points, notes that ‘we need to distinguish between somebody being taken to be an expert and somebody being an expert’ (Seidel (Reference Seidel2014), 199), and that Mizrahi's considerations merely support ‘the obvious fact that the argument from “S is taken to be an expert by another person” to “S is an expert” is fallacious’ (ibid.).Footnote ¹⁰

Given these remarks on Mizrahi's conclusion, how do his considerations relate to the second kind of case described above? As already emphasized, the rationally required reaction to a disagreement (as opposed to the degree of credence overall required) depends on the interlocutors’ initial beliefs, regardless of whether these beliefs are true or justified. And since in the second kind of case described above, the sceptic regards the experiencer as an expert in a substantial sense, by regarding her as possessing crucial, non-misleading evidence that the sceptic herself does not possess, Mizrahi's considerations do not call into question that the rationally required reaction on the part of the sceptic is to be conciliatory to some significant degree.

A more controversial issue in the literature on expertise is whether the requirement to trust experts is always governed by epistemic principles. For instance, John (Reference John2011) holds that sometimes – i.e. when deciding whether an MMR vaccine for one's own child is safe and whether one should therefore allow the child to be vaccinated – the requirement to defer to expert testimony is to be explained by a non-epistemic, fairness-based account.Footnote ¹¹ However, regardless of whether one accepts John's argument for the claim that the case in question does not involve an epistemic requirement to defer to the experts, the specific kind of example he discusses can be neglected for present purposes, for it does not call into question the claim that in the second kind of case described above (to which John's considerations are not at all applicable) there is an epistemic requirement to be conciliatory to some significant degree.

What about the requirement posed by the disagreement for the experiencer? The fact that she regards the sceptic as her meta-peer amounts to an expectation that the sceptic would, if she had also had the kind of experience in question, be equally convinced by it as the experiencer herself. Thus, the occurrence of the disagreement does not provide her with a reason for being conciliatory. But does this mean that her resulting, undiminished credence regarding the religious issue in question is overall rational? Of course not. As with the sceptic, the overall rationality of the experiencer's resulting credence depends on her initial stance regarding whether or not the kind of experience in question is misleading (as well as on the rationality of her reasons for this initial belief). If, for instance, she held the initial belief that the kind of experience in question should be regarded as analogous to the hypnotist's voice, then there would be a rational tension between this initial belief and her reaction to the experience. However, this tension would already arise at the moment of trusting the religious experience, and not at the later moment of getting to know that the sceptic disagrees on the issue in question. Thus, the experiencer's initial stance regarding the trustworthiness of the kind of experience in question comes into play earlier than the sceptic's initial stance regarding the trustworthiness of the kind of experience in question. For while the experiencer's initial stance regarding the trustworthiness of the kind of experience in question comes into play when she changes her religious beliefs in response to having the experience, the sceptic's initial stance regarding the trustworthiness of the kind of experience in question comes into play when she hears about the experiencer's reaction to having the experience, namely, at the moment of getting to know about the disagreement. Thus, the sceptic is actually rationally permitted to remain steadfast in response to the disagreement, because the relevant challenge to the overall rationality of her resulting belief (that is, the challenge of rationally accommodating her initial stance regarding the trustworthiness of the kind of experience in question) already occurs while having the experience.Footnote ¹²

Of course, the kind of case to which the notion of meta-peerhood has just been applied in order to assess the respective rational requirements is only a subset of deep religious disagreements. One may, in particular, still ask whether the sceptic is rationally required to be conciliatory in those cases of deep religious disagreement in which she does not regard her interlocutor as a meta-peer (or even as a meta-superior), but instead as less competent than herself in recognizing trustworthy evidence. The obvious stance to take here is the one according to which the degree of competence ascribed to the interlocutor corresponds to the required degree of conciliation, where – once again – the overall degree of rationality of the resulting degree of credence partly depends on the rationality of the sceptic's initial opinion regarding her interlocutor's degree of competence.

Let me close with some remarks about the extent of applicability of the above observations to actual cases of religious disagreement. The conciliationist claims stated above regarding what is rationally required from the sceptic do not only concern situations in which someone whom the sceptic has hitherto regarded as a meta-peer (or as at least somewhat competent in recognizing trustworthy religious evidence) turns out to have become a theist (or to disagree with the sceptic on some other religious issue) as a reaction to having the kind of experience in question. Rather, they also concern situations in which someone who dismissed someone else's trust in some private religious experience wonders whether this dismissal resulted in a rational attitude. If, for instance, the sceptic retrospectively realizes that she had – apart from the other's trust in the respective private experience – good reasons for regarding the other as competent in recognizing trustworthy evidence regarding the issue in question, then she should also retrospectively question the dismissal, as well as her respective religious attitude.

Thus, the question of how practically relevant all this is has also a lot to do with the question of how often there are good reasons for a religious sceptic to regard someone who will later turn out to trust some private religious experience as competent in recognizing trustworthy evidence regarding religious matters. Even though it is not my aim here to answer this partly empirical question, I wish to make the following related remark. In very many cases, it seems to be completely acceptable for religious sceptics to dismiss someone's trust in a private religious experience – for instance in all those cases in which the respective person has been subject to some form of religious indoctrination, resulting in a credulous attitude towards any experience that seemingly supports the religious beliefs of the respective community. However, the vast number of such cases does not strongly diminish the significance of those fewer cases in which the sceptic has good disagreement-independent reason to regard the other as a meta-peer – just as the fact that many people believe some theory on completely inadequate grounds does not provide a good reason for dismissing the opinions of those who believe this theory on different grounds.

Moreover, it should be noted that the supposed rarity of cases in which the sceptic has good disagreement-independent reason to regard the other as a meta-peer does not entail that a rational requirement for conciliation on the part of the sceptic is rare as well. As argued by King regarding the traditional notion of peerhood, ‘peer disagreement is rare, and [. . .] we rarely have reason to think it obtains in a given case’ (King (Reference King2012), 250), especially in cases of disagreement concerning issues people tend to genuinely care about (cf. ibid., 263). Nevertheless, King also notes that there are still many cases in which, even though there is no good reason to think that the dissenter is one's peer, there is also no good reason to think that the dissenter is in a worse epistemic position than oneself regarding the issue in question. For instance, ‘it may be unclear which of us (if either) is better disposed to respond to evidence in a rational way’ (ibid., 267). And this unclarity can arise ‘even if one is better than one's dissenter at assessing evidence’ (ibid., 268). As King plausibly claims, a rational requirement for conciliation does not only obtain in cases of acknowledged peerhood, but also in cases in which it is unclear which of the dissenters is in a better epistemic position regarding the issue in question (cf. ibid.). Analogously, even though the sceptic may rarely have good disagreement-independent reason to regard the other as a meta-peer (or as a meta-superior), there may still be many cases in which it is unclear to her who is more competent in recognizing relevant evidence concerning the issue in question. And, again analogously, as long as one does not initially hold the firm belief that the kind of evidence in question is misleading (as in the above-mentioned example of the hypnotist's voice), it may plausibly be assumed that a rational requirement for some degree of conciliation does not only obtain in cases of acknowledged meta-peerhood, but also in cases in which it is (independently of the disagreement) unclear to the sceptic which of the dissenters is more competent in recognizing relevant evidence concerning the issue in question.

At any rate, the sceptic should expect that among the vast number of cases in which people's theistic beliefs are based on religious experiences of an allegedly unique kind, there are at least some in which she should – independently of the disagreement in question – either regard the respective person as a meta-peer or at least acknowledge that it is unclear whether it is herself or the dissenter who is more competent in recognizing relevant evidence concerning the issue in question. Therefore, in all those cases in which the sceptic does not initially hold the firm belief that the kind of experience in question is misleading, she should, when confronted with the disagreement, acknowledge that complete steadfastness is not the most rational attitude to take.