Islamic Insights on Religious Disagreement: A New Proposal

Abstract

In this article, I consider how the epistemic problem of religious disagreement has been viewed within the Islamic tradition. Specifically, I consider two religious epistemological trends within the tradition: Islamic Rationalism and Islamic Traditionalism. In examining the approaches of both trends toward addressing the epistemic problem, I suggest that neither is wholly adequate. Nonetheless, I argue that both approaches offer insights that might be relevant to building a more adequate response. So, I attempt to combine insights from both by drawing a distinction between inferential and noninferential reflective responsibility. Given this distinction, I argue that it may be possible for a theist to remain steadfast in upholding their tradition-specific theistic belief, without having to hold that belief by way of inference; but nevertheless, having to be sufficiently reflectively responsible in forming their theistic belief noninferentially.

1. Introduction

Religious believers of differing religious traditions disagree as to the nature of God, His revelation, and His purposes for humankind. Similarly, religious believers disagree with non-believers as to whether God exists at all. Yet, in many such cases of religious disagreement, it often appears that either party to the dispute is just as intellectually virtuous and evidentially familiar as the other. The presence of religious disagreement between people of similar epistemic standing (i.e., epistemic peers) is thus often considered by contemporary philosophers to be a potential threat to the epistemic rationality or justification of one’s theistic belief (including atheistic belief). Let us call this the epistemic problem of religious disagreement.1

It is my aim in this article to contend with this problem. Specifically, I seek to consider the problem in reference to how it has been viewed within the Islamic tradition. In doing so, I will refer to two broad religious epistemological trends within Islamic thought: Islamic Rationalism and Islamic Traditionalism. I will point to how the former corresponds to a broadly theistic evidentialist perspective, whereas the latter amounts to a reformed epistemological view. I argue that Muslim rationalists can be understood as adopting a limited form of conciliationism and Muslim traditionalists as upholding steadfastness. In considering both perspectives as it bears on the issue of religious disagreement, I contend that neither is wholly adequate but that both contain important insights that enable us to put forth a more adequate response. In drawing on the insights gleaned from both Islamic Rationalism and Islamic Traditionalism, I offer a response to the epistemic problem of religious disagreement based on a distinction between inferential reflective responsibility (IRR) and noninferential reflective responsibility (NIRR). I argue that this distinction can enable us to see how believers may remain steadfast in upholding their tradition-specific theistic beliefs, even if they do not base such beliefs on theistic arguments or inference.

My approach then will be as follows. In Section 2, I sketch out the main preliminaries of the discussion concerning epistemic peerhood, responding to peer disagreement, and the epistemic problem of religious disagreement. In Section 3, I outline the Islamic Rationalist view, highlighting both reasons for and against this perspective. In Section 4, I do the same for Islamic Traditionalism. In Section 5, based on the previous two sections, I attempt to combine the central insights of both views toward addressing the epistemic problem of religious disagreement. I also consider an objection to the view. In Section 6, I conclude.

2. Two Preliminaries

I want to initiate our discussion then by laying out two important preliminaries before turning toward the relevant Islamic theistic perspectives on the epistemic problem of religious disagreement.

2.1. Epistemic Peerhood and Disagreement

The matter of disagreement among epistemic peers prompts two immediate questions: (a) the characterization of an epistemic peer, and (b) the epistemological consequences when such peers disagree. In attending to the first question, I take it that epistemic peers roughly denotes the following (cf., Simpson 2013):

EP: epistemic agents similarly well-qualified—in terms of epistemic virtue and evidential familiarity—to form judgements with respect to some issue.

It is essential to clarify the intended meaning here. First, epistemic agents being similarly well-qualified in terms of epistemic virtue and evidential familiarity does not necessarily mean that they possess the exact same degree of virtue or evidential familiarity. For if we interpret EP as referring to epistemic agents who are exact equals in this respect, then it would lead to an implausible consequence that disagreement bears too little epistemic impact upon one’s beliefs in the actual world (cf., Lackey 2014, pp. 302–3). This is because it is hard to imagine cases in the actual world where persons do possess the exact same degree of virtue and evidence (cf., King 2012, pp. 253–58).

Second, the fact that epistemic agents are similarly well-qualified in virtue and evidence does not necessarily mean that epistemic peerhood should be defined too broadly either. For if we were to define epistemic peers as agents similarly well-qualified in a very broad sense, it would lead to the implausible consequence of disagreement bearing too much epistemic impact upon one’s beliefs in the actual world. For in such a case, even individuals who are one’s epistemic inferiors might be treated as epistemic peers. Thus, EP indicates that only epistemic agents similarly well-qualified in virtue and evidential familiarity with respect to the specific issue under dispute count as epistemic peers.

It is also crucial to clarify the notions of epistemic virtue and evidential familiarity deployed here. By epistemic virtue, I mean to denote a kind of virtue that might enable one to acquire true beliefs, e.g., sound reasoning, accurate perception, intellectual carefulness, truth-seeking, open-mindedness, honesty, humility, and courage (cf., Roberts and Wood 2007). By evidential familiarity, I mean to refer to a familiarity with the relevant evidence that bears on the truth of some proposition. This may include arguments or inferences, as well as private experiences or rational insights. What is relevant here is not possessing the exact same kind of such evidence but rather, that the epistemic quality of such evidence is on a par.

I turn my attention now toward question (b), regarding the epistemological implications of EP on disagreement. Let us suppose there are two epistemic agents, S₁ and S₂, who are epistemic peers in the sense of EP. Suppose that both S₁ and S₂ disagree about some issue by forming contrary beliefs that p and that not-p, and both are aware of their disagreement. A potential implication of this disagreement is that both S₁ and S₂ should either decrease their confidence in or completely withold their beliefs. Philosophers who accept this implication uphold a position referred to as “conciliationism” (cf., Christensen 2009). According to this position, S₁ and S₂ should be conciliatory towards each other by adjusting their original beliefs, given their disagreement. In contrast to this position, however, the position referred to as “steadfastness” (at least in its strongest version) stipulates that epistemic peers are not required to reduce their confidence levels or abandon their original belief, and they may remain steadfast in upholding it (cf., Kelly 2005).2 Between these two opposing positions, there are more dynamic or hybrid views that see the issue as context-relative and varied from case to case.3

2.2. The Epistemic Problem of Religious Disagreement

Having explained EP and explored the broad spectrum of responses to peer disagreement, I now shift focus toward examining how disagreement between epistemic peers unfolds in the context of religious disagreement. In doing so, I want to consider the following argument from religious disagreement (cf., Dormandy 2023, p. 209):

(1)

There are people who hold theistic beliefs incompatible with mine, and yet they appear roughly as epistemically virtuous and evidentially familiar as me (RD-evidence).

(2)

RD-evidence, when considered against my total evidence, speaks against my theistic belief.

(3)

If premises (1)–(2) are true, then I ought to reduce my confidence in my theistic belief or if not, give it up completely.

(4)

Therefore, I ought to reduce my confidence in my theistic belief or if not, give it up completely.

Let us analyze this argument. Premise (1) asserts that concerning the truth of one’s theistic belief, one has epistemic peers who uphold beliefs that are incompatible with one’s own. In the context of this article, I will take this to refer to different forms of theistic belief, such as Islamic, Christian, and Jewish theistic belief, and I shall also take it that oneself and one’s peers refer to “reflective theists” (i.e., intellectually sophisticated adult theists).4 The apparent fact that one has epistemic peers who hold contrary theistic beliefs to one’s own constitutes one’s RD-evidence. By evidence, I mean some reason (i.e., argument, intuition, seeming, etc.), that justifies a certain doxastic attitude taken toward a proposition. Premise (2) states that one’s RD-evidence, when considered against one’s total evidence (i.e., all the relevant first order/higher order evidence), speaks against one’s belief (at least to some degree). Premise (3) suggests that the combination of premises (1) and (2) means that one should either lower one’s confidence or give up one’s belief altogether, given one’s RD-evidence. Hence, (4) concludes that one indeed ought to reduce their confidence or simply give up their belief, given the truth of (3).

The above argument constitutes the basis of what I refer to as the epistemic problem of religious disagreement. In what follows, I want to consider how Muslim thinkers have engaged with this sort of problem. In doing so, I’ll first consider what I take to be the conciliationism of Islamic Rationalism, and second, the steadfastness of Islamic Traditionalism. By Islamic Rationalism, I am referring to the strand within Islam known as kalām (i.e., ʿilm al-kalām). By Islamic Traditionalism, I mean the strand within Islam referred to as atharī (i.e., ahl al-āthār). This distinction, however, deserves some qualification. As Sherman Jackson (2009, pp. 31–32) points out, it is not as if Muslim traditionalists were anti-reason or that Muslim rationalists were anti-scripture or tradition; rather, each had different conceptions of reason and its place in theology. In my view though, I think there might be a distinction between the two in terms of one laying greater emphasis on reason or rational arguments, at least in the context of religious epistemology. Nevertheless, let us begin by considering the Islamic Rationalist engagement with the epistemic problem of religious disagreement.

3. Islamic Rationalist Conciliationism

3.1. Islamic Rationalist Religious Epistemology

The Islamic Rationalist tradition of the kalām theologians in the religious epistemological sense is essentially theistic evidentialist (cf., Doko and Turner 2023).5 Roughly, this means that such theologians took something like the following to be true:

TE: S’s theistic belief that p is epistemically justified/a piece of knowledge for S only when S’s belief that p is based on sufficiently supporting evidence e.

However, many Muslim rationalists also seem to have taken a more specific or qualified version of theistic evidentialism to be true (i.e., strong theistic evidentialism), where TE is qualified in something like the following way (Doko and Turner 2023, pp. 154–56; Frank 1989):

STE: S’s theistic belief that p is epistemically justified/a piece of knowledge for S only when S’s belief that p is based on sufficiently supporting evidence e, where (i) e is propositional and (ii) S is aware that e supports p.6

Let us gain some clarity on STE. According to this theistic evidentialist principle, for a theistic belief (i.e., “that God exists”) to be epistemically justified or indeed a piece of knowledge for someone, they must base that belief on propositional evidence which sufficiently supports the belief in question—by which I simply mean some theistic argument (e.g., cosmological argument) or combination of such arguments (i.e., a cumulative case theistic argument)—and where the person is aware of how the relevant propositional evidence supports their belief (i.e., aware of how or simply that the premises support the conclusion). But how exactly did Muslim rationalists arrive at this view?

First, they drew a distinction between noninferential knowledge (ʿilm ḍarūrī) and inferential knowledge (ʿilm muktasab), (cf., Ibrahim 2013, p. 102).7 This distinction thus carved out a foundationalist epistemology in that some of our knowledge was deemed to be noninferential. Crucially, however, theistic knowledge (or justified belief) was not considered to be among what we can know (or justifiedly believe) noninferentially (Heer 1993, pp. 187–88); rather, only beliefs that are evident to the senses (al-ḥissiyyāt), logically self-evident (al-badīhiyyāt), or mass reported (mutawātir) were considered by Muslim rationalists to be among our noninferential knowledge (cf., Abrahamov 1993). In this sense, they roughly upheld what Dewey Hoitenga (1991, p. 181) refers to as “medieval foundationalism”. The result of this foundationalism meant that theistic belief could not in principle be thought to comprise part of our noninferential knowledge (or justified belief). Second, and related to their foundationalism, Muslim rationalists stipulated an epistemic duty concerning theistic belief: the duty or obligation to acquire theistic belief through discursive reasoning (i.e., wujūb al-naẓar), (cf., Spevack 2020, pp. 237–42; Adamson 2022, pp. 7–11). This epistemic duty amounts to another way of stating STE, insofar as justification and knowledge with respect to theistic belief consists in fulfilling a certain epistemic duty, which is fulfilled when one reasons discursively and grounds their theistic belief on sufficiently supporting evidence in terms of (i) and (ii). Hence, the Muslim rationalists’ arrival at theistic evidentialism.

This religious epistemological outlook embraced by Muslim rationalists, particularly in relation to this epistemic duty to reason discursively, bears an important relation to our issue concerning the epistemic problem of religious disagreement. For one thing, this epistemic duty appears to eliminate the possibility of being justified in upholding one’s theistic belief by way of private religious experience (e.g., by spiritual inspiration—ilhām; or spiritual unveiling—kashf) or through testimonial evidence gleaned from following authorities in one’s socio-religious milieu (i.e., taqlīd). This implies that settling religious disagreement over theistic belief must in principle be settled on argumentative grounds or by way of publicly accessible evidence. But it also seems to me that it is not simply the foundationalism or theistic evidentialist principle of STE that motivated Muslim rationalists to uphold this view; in fact, it seems to me that the very problem of religious disagreement itself had important bearing on why they formed this religious epistemological outlook.

3.2. Religious Disagreement and Limited Conciliationism

Muslim rationalists did consider the epistemic problem of religious disagreement, but in a more restricted or limited form than that captured in the above argument in Section 2.2. Consider, for instance, the kind of epistemic problem of religious disagreement spoken of in the following passages of two early Muslim rationalists. The first is from Abū Manṣūr al-Māturīdī (d. 944 CE), and the second is from Qāḍī ʿAbd al-Jabbār (d. 1025 CE):

[1] We find that all people, with their different religious opinions and sects, agree on one statement, namely, that whatever one holds to be true, is valid, and, as a result, that whatever others than him hold, is invalid … Therefore, it is taken for granted that taqlīd [i.e., blind social imitation/following] excuses its embracer from holding the opposite view on the same question. This, however, only accounts for the multiplicity of the number. The only way out of this is if one of them has his ultimate argument based on intellect (ʿaql) by way of which his truth can be known and if he has proof by way of which he can persuade fair-minded people to accept his truth

(quoted in Ceric 1995, pp. 67–68; cf., Kam 2022, pp. 36–37).8

[2] The learned man knows that there are people who are wrong in their speculation (yukhṭiʾu) and others who are right (yuṣību), however each of them claims that he is right. Why, then, is the taqlīd of one of them better than the other? Why is the taqlīd of the believer of God’s unity better than the taqlīd of the unbeliever? … What is the proof that the learned man must speculate [i.e., reason] in order to obtain the knowledge of God? … [because he knows] there are systems which are right and others which are wrong.

(quoted in Abrahamov 1998, pp. 58–59)

What these two passages appear to illustrate is the importance that the epistemic problem of religious disagreement has played in Muslim rationalist thinking about the epistemic duty to reason discursively. Particularly instructive is the first passage from al-Māturīdī. This passage highlights the centrality of reason (ʿaql) as the ultimate guarantor amid religious disagreement, as well as evidence which may persuade others, i.e., publicly accessible evidence. Indeed, al-Māturīdī (Kam 2022, pp. 40–41) goes on to add that private religious experience in the form of “spiritual inspiration” (ilhām) is unfit as a sure basis for one’s theistic belief, just as is blindly following one’s socio-religious community (i.e., taqlīd). The second passage from ʿAbd al-Jabbār seems to emphasize the issue of epistemic luck in the context of disagreement. In this case, the one who fails to reason discursively lacks some form of “symmetry breaker” by way of a rational argument to tip the scales of truth in one’s favor. For without such a rational argument, where one merely follows his or her socio-religious milieu, acquiring a true theistic belief is a matter of the sheer luck and contingency of one’s upbringing. But such epistemic luck precludes knowledge.

It seems to me that the main epistemic problem Muslim rationalists were concerned with relate to the social contingency of religious belief, private/public evidence, and epistemic luck. In other words, they were concerned that basing theistic belief on one’s religious socio-milieu or private religious experiences would prevent one’s theistic belief from receiving positive epistemic appraisal because its truth seems to be a matter of mere epistemic luck (Adamson 2022, pp. 7–11). This is apparently so because of the evidential symmetry or epistemic parity shared between various religious believers of different stripes holding conflicting theistic beliefs, with only one of them being right. So, Muslim rationalists sought to resolve this problem by stipulating an epistemic duty which amounted to acquiring a symmetry breaker in terms of some publicly accessible evidence (i.e., theistic argument), tipping the evidential and epistemic scales in one’s favor and eliminating any concern over epistemic luck. For if one believes justifiedly through some sound theistic argument, then one is rightly appraised or credited with the epistemic success of forming a true theistic belief.

What this suggests, however, is that Muslim rationalists affirmed a limited form of conciliationism with respect to religious disagreement. Unlike the argument from religious disagreement presented above, these thinkers had in mind a different version of the argument where premise (1) concerned theistic belief being based on something other than theistic arguments. In other words, where that premise looked something more like the following:

(1*) There are people who hold theistic beliefs incompatible with mine, and yet they appear to be my epistemic peers because they similarly hold their beliefs on the basis of either (a) socio-religious imitation (i.e., taqlīd) or (b) private religious experience/evidence (i.e., ilhām; kashf). (RD-evidence).

It is the truth of this premise, in combination with something like (2) and (3), which Muslim rationalists would appear to take as entailing something like the conclusion (4). In essence, I think that they appeared to have upheld conciliationism but in a limited way, where one ought to be conciliatory in cases where there is something like symmetry in private or testimonial evidence (broadly construed). Hence, there is an apparent epistemic need for a symmetry breaker through means other than these forms of evidence that are publicly accessible. Clearly, Muslim rationalists had a rather straightforward solution to the epistemic problem of religious disagreement they had in mind as per premise (1*), for they proposed that one can simply reject the premise by wielding a theistic argument. In this case, one’s theistic belief would not be held similarly on the basis of either (a) or (b); rather, it would be based on some evidence that breaks the symmetry, thus doing away with the RD-evidence.

Admittedly, there does seem to be something right about the idea of needing a symmetry breaker in terms of publicly accessible evidence. As Hamid Vahid (2014, p. 91) explains in the context of peer disagreement, if one is to remain justified, then “the initial epistemic symmetry between peers has to be broken at some point along the way”. Similarly, Kirk Lougheed (2018) considered remaining steadfast in the face of peer disagreement in terms of possessing some “evidential asymmetry” between oneself and one’s peer. So, the Muslim rationalists seem to have been onto something here. For, plausibly, being justified in holding steadfast to one’s initial theistic belief requires that one has some evidential asymmetry that tips the evidential scales in one’s favor, sufficient to demote one’s epistemic peer. It might also be plausible to defer solely to publicly accessible evidence because it precludes the concern of either party to the dispute harking bias by deferring to one’s socio-religious authorities or favoring one’s own private evidence over that of their peers, especially when the disagreement itself gives one prima facie grounds to question one’s private or community-specific evidence.

Perhaps, however, there is also another central epistemological insight here that underlies something close to Muslim rationalist thinking, which concerns epistemic luck and epistemic credit.9 Consider that if one forms a true belief by way of luck, then one does not deserve any epistemic credit for the success of acquiring that true belief (cf., Greco 2003, p. 123; Pritchard 2016, p. 5). By contrast, if one forms a true belief by exerting the relevant intellectual effort in acquiring a true belief, then one deserves epistemic credit for the success of having acquired it (Greco 2003, p. 123; Greco 2010, pp. 71–75; Pritchard 2016, p. 64). It seems to me that Muslim rationalists were concerned that failing to reason discursively in acquiring theistic belief, even where that belief is true, still precludes one from epistemic credit because it is merely a matter of epistemic luck that one acquires that true belief. So, it is only when one has exerted intellectual effort or responsibly exercised one’s reflective capacities in considering the relevant evidence at hand that one can be epistemically credited when acquiring some true belief. It is just this that is perhaps what is at the root of the problem in cases of holding theistic belief based on social imitation or private religious experiences. For in these cases, one has not exercised reflective responsibility in terms of logical reasoning or argumentation, sufficient to deserve epistemic credit. I think this view has something going for it, particularly in the context of philosophical or religious beliefs held by reflective thinkers. So, we will return to this key epistemological insight of Muslim rationalists later in Section 5.

3.3. The Limits of Limited Conciliationism

I have argued that the conciliationism of Muslim rationalists is of a limited kind. Indeed, Muslim rationalists did not believe that we should be conciliatory once we have fulfilled our epistemic duties and appropriately acquired (Islamic) theistic belief. However, they did seem to think being conciliatory is the epistemic consequence in cases where one fails in their epistemic duties due to grounding one’s theistic belief in something like socio-religious imitation/community-specific testimony or private religious experiences. For here, the problem of evidential symmetry and epistemic parity between similarly positioned theists rids them of their epistemic justification (and by extension knowledge).

The trouble with this limited conciliationist view, though, is that Muslim rationalists seem to think we can escape the impasse of private religious experience or socio-religious imitation through publicly accessible evidence; however, it is not so clear that they have established that the symmetry breaks down once the dispute moves to such evidence. In other words, it fails to address the epistemic problem of religious disagreement encapsulated in the argument of Section 2.2. Perhaps one thinks it is unfair to criticize a view that did not explicitly critique the argument I am considering, but it is not unreasonable to point out that some of the thoughts they left us with are clearly limited in application to new and similarly relevant problems. So, it seems a legitimate concern.

To see the problem more clearly, consider two reflective theists with conflicting theistic beliefs. Suppose S₁ satisfies the STE principle and so is at least prima facie justified in upholding her Islamic theistic belief on relevant publicly accessible evidence, and thus also well positioned to reject premise (1*). Let us also suppose that S₂ satisfies STE and so justifiedly rejects premise (1*), but forms a contrary Christian theistic belief (e.g., that God is one but tri-personal). Let us further suppose that both believers are aware of each other’s disagreement, the publicly accessible evidence upon which both beliefs were formed, and that they are epistemic peers as per EP (see Section 2.1). The question now remains: do S₁ or S₂ display evidential symmetry and epistemic parity? At least prima facie, the answer seems to be yes. But if that is right, then the mere fact that S₁ forms her belief on some publicly accessible evidence (i.e., a selection of Islamic theistic arguments) does not in itself appear to constitute an evidential symmetry breaker required for her to demote S₂. This then is the problem of the limited conciliationism of Muslim rationalists: whence the need for theistic arguments? In other words, what significant epistemic difference do they make when one is aware of religious believers who also form conflicting theistic beliefs by way of argument? When considered in terms of evidential asymmetry, it seems the epistemic difference such arguments make is not as significant as Muslim rationalists seem to have thought.

So, we are left with a problem concerning the Islamic Rationalist view, which in effect means that the epistemic problem of religious disagreement we are concerned with remains. Perhaps the Islamic Traditionalist view can offer some resources to help.

4. Islamic Traditionalist Steadfastness

4.1. Islamic Traditionalist Religious Epistemology

Unlike in the case of Islamic Rationalism, Islamic Traditionalism is less systematic in that it does not consist of a clearly unified religious epistemological trend within Islamic thought; instead, it consists more of a spectrum of views which share important commonalities concerning reason and tradition. I want to offer two examples of Islamic Traditionalist approaches in religious epistemology, one which essentially defends taqlīd or socio-religious following, and the other private evidence in the form of religious experience, at least broadly construed. Despite this slight difference between these two traditionalist approaches, there is an important sense in which they converge: both are reformed epistemologically. By this, I simply mean that they uphold something like the following (cf., Moon 2016):

RE: S’s theistic belief that p may be epistemically justified/a piece of knowledge for S, even where S’s belief is not based on some argument that supports p.

Insofar as Muslim traditionalists uphold something like RE, they stand opposed to the Islamic Rationalist principle of STE. The two thinkers I want to draw on as examples of Islamic Traditionalism are Ibn Hazm (d. 1064 CE) and Ibn Taymiyya (d. 1328 CE). In referring to the former, I will briefly illustrate how Islamic theistic belief might be construed as potentially justified or known on testimonial grounds. In reference to the latter, I will try to show how Islamic theistic belief might be considered justified or known based on religious experience.

Ibn Hazm (2014, pp. 4:28–36; cf., Izutsu 2006, pp. 155–62) takes aim at those who chastise Muslim believers for upholding their Islamic theistic belief on grounds other than by way of argument or inference (istidlāl). In brief, he puts forward several counterarguments against the Islamic Rationalist view which amount to arguing for why a Muslim believer may be justified in upholding their belief on noninferential grounds. I will consider just two of the points most relevant to our discussion. First, he makes a distinction between socio-religious imitation (taqlīd) condemned by the Qur’an on the one hand, and religious faith (īmān) on the other. In my view, his distinction amounts to one between good socio-religious imitation vs. bad socio-religious imitation (cf., Ibn Hazm 2014, pp. 4:29–30). It is only the latter which is to be condemned because it amounts to following those who God has not sanctioned us to follow. Second, he essentially rejects the foundationalism of the Muslim rationalists by arguing that they are wrong to restrict knowledge to either noninferential knowledge by way of the senses and inferential knowledge by way of rational argument (Ibn Hazm 2014, p. 4:32). Instead, he suggests that there are many ways of acquiring theistic knowledge, including intellectual intuition (badīhat ʿaql) or God directly bestowing it upon a person (Ibn Hazm 2014, p. 4:32).

Turning now to Ibn Taymiyya, he also upheld something like the principle of RE above, but primarily by arguing that we can acquire theistic knowledge through a natural (fiṭrī) theistic disposition/faculty that we have by virtue of our God-given nature (fiṭra) as humans (cf., Turner 2021). Roughly, he states that we have a natural ability to know God, such that when it works as it ought to, we will naturally form belief in God and submit to Him (cf., Ibn Taymiyya 2004, p. 6:73). This natural theistic knowledge is activated through one’s recognition of theistic signs in the natural world (i.e., āyāt al-fiʿliyya), as well as in the Qur’an (āyāt al-qawliyya) (cf., Ibn Taymiyya 1979, p. 7:302; Turner 2022). As I pointed out in my (Turner 2021), the view is externalist insofar as the acquisition of theistic knowledge is principally the result of our cognitive nature and its related faculties working properly, as it was designed to by God, in the appropriate cognitive environments. Though Ibn Taymiyya (1979, p. 6:271) does acknowledge circumstances in which our fiṭrī theistic disposition/faculty may fail to work as it ought to, still when it does, theistic knowledge will ensue.

4.2. Islamic Traditionalist Steadfastness

Muslim traditionalists, whether adhering to views similar to those of Ibn Hazm or Ibn Taymiyya, appear to dismiss the epistemic problem of religious disagreement as understood by Muslim rationalists and by extension, the contemporary version we are concerned with in this article. According to Ibn Hazm, faith in the Prophet and, in effect, the testimony of one’s Islamic religious community is sufficient to have knowledge of Islamic theistic belief, while Ibn Taymiyya stipulates that one can acquire Islamic theistic knowledge by way of our fiṭrī theistic disposition/faculty, at least when situated in the right cognitive environment. In both cases, it does not appear that they regarded the mere fact of religious disagreement as a threat to the epistemic status of one’s Islamic theistic belief. This is especially so given their rejection of any need to engage in discursive reasoning to validate one’s theistic beliefs in the way that Muslim rationalists have suggested one ought to. So, Muslim traditionalists will want to say that the epistemic problem of religious disagreement, as per the argument from (1)–(4), is in some way unsound. But how would they go about arguing this?

It seems to me that they might challenge the second conjunct of premise (1), that other religious believers appear roughly as epistemically virtuous and evidentially familiar as me. Perhaps one strategy that they may be seen to deploy is what Katherine Dormandy (2023, p. 211) refers to as “divine-help epistemology”. This idea consists of a positive thesis, to the effect that God has especially enabled one’s personal cognition or that of one’s community to form true beliefs about Him, and a negative thesis, that one’s interlocutor or their community in some sense lacks that divine help. Concerning the negative thesis, one idea might be that a person is suffering from some form of noetic impediment brought about by the sinful actions of the person or the broader community. Therefore, just as one is epistemically advantaged by divine help (broadly construed), one’s interlocutor is epistemically disadvantaged by the absence of such help. It is possible to see this idea in both Ibn Hazm’s and Ibn Taymiyya’s thinking. For instance, Ibn Hazm suggests that when one follows or imitates the person or community that God has not commanded one to follow, this is condemnable; but when one follows God’s prophet, it is praiseworthy. In Ibn Taymiyya’s case, he speaks of socio-religious environments as acting as veils preventing the truth from being known, as well as individual actions which cause one’s fiṭrī theistic disposition/faculty to become blocked or impaired (cf., Ibn Taymiyya 1979, p. 6:271; 2004, p. 4:257). So, in essence, divine help is lacking. This makes at least some sense, in that it may, in fact, be the case that the difference between knowing God or not is essentially divine-help and the avoidance of certain noetic impediments. So, as long as one has that help and avoids such impediments, maybe they would know God and gain some reason to demote their peers.

Another way to think about this steadfast approach is to see it as taking the opposite strategy to what we have associated with Muslim rationalists about giving greater weight to certain kinds of evidence. The Muslim traditionalists might be thought of as giving epistemic priority to “partialist” forms of evidence over “impartialist” forms.10 By the former, I mean evidence which is private, not wholly shareable, specific to one’s community, and/or affected by one’s religious biases (Dormandy 2018, p. 66). By the latter, I mean evidence which is public, shareable, and analogous to evidence in science (Dormandy 2018, pp. 63–64). The result of this prioritization may be that the impartialist evidence in the form of RD-evidence does not, all things considered, count against one’s total evidence given the greater weight allotted to one’s partialist evidence. So, traditionalists can reject premise (2) of the above argument. This arguably makes sense, for it may be that the most salient forms of theistic evidence are of a more experiential or community-centered type, for these are the sorts that are acquired through special divine help. So, it would be wrong-headed to prefer weaker forms of impartialist evidence.

4.3. The Problems of Traditionalist Steadfastness

Despite some plausibility in the steadfastness of Islamic Traditionalism, it faces its own difficulties.

First, there is the problem of evidential symmetry between theists who hold their belief either on broadly testimonial grounds or through religious experience that cannot be so easily dismissed in the way that the above steadfastness suggests. Consider again two theists, S₁* and S₂*. Suppose that S₁* forms her Islamic theistic belief by forms of partialist evidence, e.g., her readings of scripture, instruction from her religious teachers, her sense of God’s presence during acts of worship. Suppose similarly that S₂* forms her Christian theistic belief by the same kinds of partialist evidence. Let us further suppose that they are both aware of their disagreement and report their evidential insights to each other. Imagine that they both also adopt a divine-help epistemology and give greater weight to their partialist evidence. The problem seems to be that the levels of evidential symmetry and epistemic parity between the two preclude both from being justified. For even if they have some incommunicable evidential insight, each is at least aware that one’s peer has such an insight as well and can recognize it as bearing evidential force. So, the worry is that one has reason to question one’s own insights or experiences as sufficient indicators that one’s belief is true (cf., Feldman 2007).

Second, this approach also falls prey to the problem of neglecting the importance of exercising our reflective capacities responsibly in forming theistic belief. As we will see in the next section, and have already alluded to above, receiving epistemic credit particularly concerning controversial philosophical issues like theism seem unlike our ordinary sensory beliefs in the demand for responsible reflective thinking, including engaging theistic arguments. So, as in the case of Islamic Rationalism, it seems to me that the Islamic Traditionalist response to the epistemic problem of religious disagreement is not wholly satisfactory. However, I think that both retain important insights which, when taken together, might enable us to construct a more adequate response to the problem.

5. Combining Insights

In light of our discussion above, I want to carve out a way of responding to the epistemic problem of religious disagreement, and hence the argument laid out in Section 2.2, by drawing on important insights present in both the limited conciliationism of Muslim rationalists and the steadfastness of Muslim traditionalists. When taken together, I’ll argue that these insights provide a plausible way of responding to the epistemic problem of religious disagreement that combines the two approaches.

5.1. Inferential and Noninferential Reflective Responsibility

In combining the central insights of Islamic Rationalism and Islamic Traditionalism, I first want to draw a distinction between two different kinds of reflective responsibility. Recall that a central insight for Muslim rationalists was the idea that theistic belief must be the result of exercising our reflective capacities responsibly. In their view, this can only be done when one grounds one’s theistic belief in inferential terms (i.e., on some theistic argument). By contrast, Muslim traditionalists reject the necessity for inference; rather, a central insight of their view is that we can have justified theistic belief or knowledge noninferentially. In order to bridge these two opposing views in the context of religious disagreement, I offer the following distinction between (a) inferential reflective responsibility and (b) noninferential reflective responsibility. When applied to some subject S, we can render them in the following terms:

IRR: Inferential reflective responsibility is exercised where S has carefully considered the relevant evidence that directly bears on p, and where S forms her belief that p by a conclusion of inference.

NIRR: Noninferential reflective responsibility is exercised where S has carefully considered the relevant evidence that directly bears on p, and where S forms her belief that p by a conclusion of reflection.

I take it then that the main distinguishing factor between inferential and noninferential reflective responsibility is principally as to whether S forms her belief that p by way of a “conclusion of inference” or a “conclusion of reflection”.

The distinction between these two types of belief formation has been drawn out by Robert Audi (2004, pp. 45–48). According to Audi (2004, p. 45), conclusions of inference refer to beliefs formed on the basis of premised propositions noted as evidence. An example of a conclusion of inference then would just be some belief based on an argument (e.g., a simple modus ponens/tollens). A conclusion of reflection, though, refers to a belief grounded in a kind of intuitive judgement “formed through rational inquiry or searching reflection” (Audi 2004, p. 45). This form of reflection emerges from thinking about the matter in question as a whole or in a more global sense, as opposed to from specific evidential premises. As Audi puts it, “drawing a conclusion of reflection is a kind of wrapping up of the question, akin to concluding a practical matter with a decision. One has not added up the evidence and inferred their implication; one has obtained a view of the whole and thereby broadly characterized it” (Audi 2004, pp. 45–46).

Audi (2004, pp. 45–46) gives the example of reflecting upon a poem or a letter and wrapping up that reflection with a proposition about the poem, which emerges from thinking about the poem as a whole but not from one or more relevant evidential premises. For instance, it may be that upon one’s reading and reflection, one senses a subtle artificiality or stilted quality in the language of the poem and thus concludes or wraps up this reflection with a belief in some relevant proposition (e.g., the language is artificial).11 Epistemically speaking then, whereas conclusions of inference are inferential, conclusions of reflection are noninferential. However, that is not to say that the latter is non-reflective; on the contrary, both forms of conclusions result from an exercise of one’s reflective capacities, and both are formed responsibly when S carefully considers the relevant evidence bearing on the truth of some proposition p. By “careful consideration”, I mean exhibiting the virtue of intellectual carefulness (i.e., being mindful of the salient aspects of the relevant evidence and thoroughly considering the evidence as fairly as possible, but not being overly scrupulous).12

To see the difference between these two forms of belief formation in the context of our discussion, let us consider some examples concerning theistic belief. Examples of conclusions of inference abound in philosophy of religion. Take any version of the cosmological, teleological, or ontological arguments. Holding theistic belief on any of these would constitute holding it by a conclusion of inference. Broader cumulative cases where one grounds their theistic belief abductively by noting various data about the world that theism best explains would also be based on a conclusion of inference. As for grounding theistic belief on a conclusion of reflection, helpful in this regard are Michael Bergmann’s (2017, pp. 36–37) comments. He notes that theistic belief might be based on theistic seemings grounded in various experiences, observations, and considerations over, perhaps, a long period of time (including considerations over relevant theistic arguments), and where the remnants of these considerations are retained in memory, enabling one’s theistic seemings to emerge upon reflection. In this case, although one’s theistic belief is based on various considerations and experiences, it is nonetheless noninferential. For instead of taking aspects of one’s reflection as evidential premises, here, one wraps up their reflection in a holistic fashion with the proposition that theistic belief is true as opposed to grounding it on specific premises.

5.2. Reflection, Disagreement, and Skepticism

To understand what relevance the above distinction has on the issue at hand, let me make three points. First, I think that the central Muslim rationalist insight is correct, that given religious disagreement, for a reflective theist’s belief to be justified, it requires an exercise of reflective responsibility. Second, that if epistemic peers S₁ and S₂’s conflicting theistic beliefs are based on an exercise of reflective responsibility, then it is wrong to think that both or either peer must completely withhold their theistic beliefs because this entails an implausible form of skepticism. However, third, an exercise of reflective responsibility required for justification—contra Muslim rationalists—does not entail the need for a theistic argument, hence satisfying IRR. It may be based on noninferential grounds thus supporting the Muslim traditionalist insight. Specifically, it is sufficient that it satisfies merely NIRR. Let us consider these points in turn.

That exercising reflective responsibility for reflective theists is (all things considered) necessary for justification can be argued by pointing out that justified theistic belief or knowledge is “upper-end” as opposed to “lower-end”. By the latter, I mean roughly justified beliefs or knowledge of the more mundane or ordinary kind (i.e., perceptual or memorial beliefs); by the former, those which are less ordinary, such as in philosophical, scientific, or religious domains (cf., Roberts and Wood 2007, pp. 109–10). The idea here is that whereas in lower-end cases, reflective responsibility is hardly necessary—which may be explained by it being more obviously involuntary in the doxastic sense—on the other hand, upper-end cases are far more in the “realm of the voluntary” and thus require an exercise of our reflective capacities (cf., Greco 1997), at least for reflective theists, all things considered. In other words, forming tradition-specific theistic beliefs, even if by religious experience, seems less like ordinary perceptual beliefs in terms of their doxastic involuntariness; belief formation here typically involves our own added reflection, interpretation, and consideration. As Duncan Pritchard points out,

… whereas this sort of [doxastic] ‘directness’ is the norm in the perceptual case, it is more naturally thought of as the exception to the norm in the religious case. Indeed, whereas perceptual beliefs seem to be, in the main, ‘forced’ upon us, religious beliefs often seem to be formed in a far less compelling fashion … the more common way of conceiving such belief is in terms of being ‘nudged’ or ‘invited’ towards a certain doxastic commitment.

(Pritchard 2005, p. 191)13

Pritchard (2005, p. 192) goes on to add that although some “agent might gain a warrant for his perceptual belief without engaging his reflective capacities at all, religious beliefs seem to directly implicate such capacities”. This does not mean that one ought to endorse the thesis of doxastic voluntarism; rather, the suggestion here is merely that religious belief—at least among reflective theists—typically stands somewhere within the realm of the voluntary (to borrow John Greco’s phrase). Furthermore, even if we do have something like a fiṭrī theistic disposition/faculty, Robert Roberts and Jay Wood (Roberts and Wood 2007, pp. 108–12) point out that the reliability of our cognitive faculties (especially in “upper-end” cases) might require reflective-like intellectual virtues, such as being truth-seeking; intellectually careful; courageous and open-minded, otherwise such faculties may not be sufficiently reliable in securing a high degree of justification. So, the Muslim rationalist insight here at least seems plausible, that an exercise of reflective responsibility is necessary. (I think this might also be what explains our intuition as to why S₁* and S₂* are unjustified, namely, that they satisfy neither IRR nor NIRR).14

Earlier, we questioned the Muslim rationalist view by wondering what the need for theistic arguments is amid peer disagreement with another theist who holds their beliefs on conflicting theistic arguments. We have at least a partial answer, that it enables S₁ and S₂ to satisfy IRR and so, prima facie, S₁ and S₂ can receive epistemic credit for their theistic beliefs. But then, does that prima facie epistemic credit not see itself wholly removed in the context of peer disagreement? For in this case, S₁ and S₂ are aware that both form contradictory theistic beliefs whilst being similarly reflective and evidentially familiar. I think the answer is no. To see why, consider that the dispute between S₁ and S₂ in important epistemic ways resembles disputes between reflective thinkers in a whole host of other domains, such as metaphysics, epistemology, ethics, politics, science, literature, and history. It resembles many disputes in those domains because epistemic peers within areas like philosophy base their beliefs in various propositions or theories on philosophical arguments. Thus, if we were to conclude that full conciliation is required when peers in philosophy dispute, given that there is widespread dispute, it seems to entail that we must embrace some implausible version of philosophical skepticism. In other words, where we would have to maintain that we and others lack justification for a great many substantive philosophical claims that are, by our lights, plausibly true.15 Intuitively, this seems false. We can put this in terms of the following argument:16

(5)

If peer disagreement in philosophy based on opposing philosophical arguments means that one ought to withhold their philosophical beliefs, then this entails philosophical skepticism.

(6)

But philosophical skepticism is false.

(7)

Therefore, peer disagreement in philosophy based on opposing philosophical arguments does not mean that one ought to withhold their philosophical beliefs.

The conditional captured in (5) seems obvious enough in that if the antecedent holds, it seems clear this would entail skepticism about many, if not most, of our philosophical views. But is (6) really true? Intuitively, it seems to be that it is. At least, it seems more plausibly true than that peer disagreement entails that epistemic peers in philosophy lose all their justification for most of their philosophical beliefs. So, the conclusion seems plausible. The result of this argument if it holds is that if S₁ and S₂ are sufficiently reflectively responsible in forming their theistic beliefs on philosophical arguments, then despite apparent peerhood, they do not thereby gain some reason to wholly give up their beliefs and thus lose their epistemic justification (at least not completely). But then, we return to the question of the need for these theistic arguments. The question this time is: can S₁ or S₂ be sufficiently reflectively responsible even if they do not form their theistic beliefs by way of theistic arguments or, in other words, by a conclusion of inference? I think they can.

5.3. IRR, NIRR, and Epistemic Significance

To see why, we need to return to the distinction between principles IRR and NIRR. Recall that as per IRR, one is reflectively responsible in an inferential sense where one bases one’s belief on some conclusion of inference and where one carefully considers the relevant evidence directly bearing on the truth of p. NIRR, on the other hand, states that one is noninferentially reflectively responsible where one acts as one does in satisfying IRR but this time bases one’s belief on some conclusion of reflection. In the previous two sections, we established that (a) exercising reflective responsibility is necessary for reflective theists, and (b) where S₁ and S₂ exercise reflective responsibility in a way that satisfies IRR, they can receive epistemic credit for that belief, which is also not wholly eradicated because of their peer disagreement (due to the problem of philosophical skepticism).

However, I think that it is sufficient for S₁ and S₂ to receive the epistemic credit of justification by exercising reflective responsibility in a way that satisfies NIRR but not IRR; that is, I think it is sufficient that they merely satisfy NIRR. Note then a shift here toward the Muslim traditionalist view, whilst also maintaining the Muslim rationalist insight that exercising reflective responsibility is necessary. For what I am saying here is that there can be a noninferential basis for justified theistic belief, whereas Muslim rationalists deny that outright. A brief remark or two on this is thus in order. It does not seem plausible to me to uphold the broadly medieval foundationalism of the Muslim rationalists, where only a few restricted beliefs may be properly basic with respect to justification or knowledge. Plausibly, if one has a basic belief about very many things—including religious things—and that belief is, say, based on a relevant seeming, the result of intellectual virtue, or properly functioning faculties (depending on one’s view of justification/warrant), then that belief can receive at least prima facie epistemic credit. To quote William James (1982, pp. 423–24), “our senses, namely, have assured us of certain states of fact; but mystical experiences are as direct perceptions of fact for those who have them as any sensations ever were for us”. In other words, if our senses or strong perceptual seemings give us prima facie justification about mundane objects, then the same applies to other areas including the religious. So, if a reflective theist bases her theistic belief on noninferential grounds, she can at least in principle receive some epistemic credit, even if not all things considered. This then tips the hat toward Muslim traditionalists, at least in regard to what counts as part of one’s evidential base.

Having cleared the ground somewhat, I now want to offer a reason for thinking that satisfying NIRR is sufficient for epistemic credit in the context of religious peer disagreement among reflective theists, and that IRR is not necessary. In essence, my argument is the following: if S₁ and S₂ deserve epistemic credit upon satisfying IRR, then because IRR and NIRR are not significantly epistemically different, they also deserve such credit when satisfying NIRR. To see why, consider that if S₁ or S₂ exercise reflective responsibility as per NIRR and thus form their theistic belief on a conclusion of reflection, then both would have nonetheless formed a kind of “inferential disposition”. An inferential disposition refers to a type of “disposition to believe” (cf., Audi 1994). Roughly, by a disposition to believe I mean a proclivity for S to believe some proposition p if S considered the thought that p. For example, although I might not occurrently believe or previously have believed that 98.184 is a larger number than 98, I have a disposition to believe that proposition given my prior beliefs (Audi 1994, p. 419). An inferential disposition is thus a disposition to believe some proposition on the basis of other propositions that one believes. For instance, if I believe that Arthur Grimshaw (d. 1913 CE) is an artist and that he famously painted many oil paintings of Leeds, that may dispose to me infer that a particular oil painting of Leeds is his, if I entertained the proposition that it is upon seeing an oil painting of Leeds.

Now, in the context of forming conclusions of reflection, Audi (2004, p. 47) points out the following: “where one arrives at a conclusion of reflection, one could figure out why and then formulate, in explicit premises, one’s basis for so concluding”. So, to put it another way, although in forming a belief by a conclusion of reflection, one is not thereby forming it inferentially. This is not to say that one lacks a disposition to form such a belief by a conclusion of inference. Indeed, in forming a belief by a conclusion of reflection, one may still thereby form an inferential disposition. That is, one may acquire an inferential disposition to believe some proposition or, in other words, a disposition to believe by a conclusion of inference. Hence, my claim here is that if a reflective theist like S₁ or S₂ forms Islamic theistic belief, say, by a conclusion of reflection and satisfies NIRR, such a theist would acquire an inferential disposition such that if and when prompted to form her believe by a conclusion of inference, she could readily do so. To see this, consider the following passage from the famous Muslim intellectual and convert Muhammad Asad (d. 1992 CE):

An integrated image of Islam was now emerging with a finality, a decisiveness that sometimes astounded me. It was taking shape by a process that could only be described as a kind of mental osmosis—that is; without any conscious effort on my part to piece together and ‘systematize’ the many fragments of knowledge that had come my way during the past four years. I saw before me something like a perfect work of architecture, with all its elements harmoniously conceived to complement and support each other, with nothing superfluous and nothing lacking—a balance and composure which gave one the feeling that everything in the outlook and postulates of Islam was in its proper place.

(Asad 1999, p. 301).

Although this passage might not be a perfect fit for a conclusion of reflection, it is good enough. Asad describes his newly found picture of Islam as taking place through a kind of mental osmosis (i.e., a gradual assimilation of ideas). In other words, perhaps, through various experiences, reflections, observations, and considerations over a long period of time (à la Bergmann).17 It is through such experiences and reflections that a conclusion as to the truth of Islam appears to emerge for Asad. For our purposes, let us imagine Asad’s reflection concerns the relevant body of evidence needed to responsibly arrive at the conclusion and that he did so with intellectual care. We can imagine that this also includes various contrasting theistic arguments. However, it is not that Asad consciously (or even tacitly) made note of particular evidential premises by which he inferred the conclusion that Islamic theism is true; rather, it just had that feeling of truth arising from his various reflections.

However, even though Asad suggests he made no effort to piece together the fragments of knowledge that relate to his conclusions, it seems to me that given sufficient prompting someone like Asad could, at least to some significant degree, stand back and reflect upon such fragments and formulate them into evidential premises from which an inference to Islamic theism could be made. In other words, although Asad might be taken to have formed Islamic theistic belief by a conclusion of reflection and satisfied NIRR, he would probably have also acquired a relevant inferential disposition. So, given this disposition to believe inferentially, he could satisfy IRR if and when he was prompted to do so. In my view, this seems to suggest that the epistemic significance of satisfying IRR instead of NIRR is not all that great, such that if a reflective theist amid peer disagreement deserves epistemic credit when satisfying IRR—as we have already suggested—so too does she deserve epistemic credit in merely satisfying NIRR.

If what I have just suggested is right, we now ought to consider how it applies to the epistemic problem of religious disagreement encapsulated in the argument outlined above (in Section 2.2) from (1) to (4). The focus here is on premise (2). Recall that this premise is that one’s RD-evidence, when considered against one’s other total evidence, speaks against one’s theistic belief. We have already argued above that if a reflective theist satisfies IRR when forming their tradition-specific theistic belief, it is implausible to think that their RD-evidence speaks against their belief to such an extent that it rids them of epistemic credit entirely, such that they ought to give up their belief. This means that such persons like S₁ or S₂ can remain steadfast in the sense that they are not epistemically obliged to withhold their theistic belief. Yet, as we have just pointed out, what holds for the one who satisfies IRR similarly holds for the one who satisfies NIRR. Therefore, if S₁ and S₂ form contrary theistic beliefs in a way that satisfies NIRR, despite an awareness of peer disagreement, they may still nonetheless be credited with epistemic justification, at least prima facie. So, premise (2) is false and the argument collapses.

5.4. A Partial Solution?

There is, however, an obvious objection here. The objection is that even if what I have just said is correct, it does not thereby show that S₁ or S₂ are not required to hold their belief less firmly and hence does not show that premise (2) is false. In other words, the argument above from (1) to (4) remains because the combination of (2) and (3) is compatible with only one of the disjuncts of (3) being true (i.e., that one ought to either reduce credence or give up the belief completely). So, the solution I have offered is only a partial one, leaving open the possibility that S₁ and S₂ should reduce their confidence in their theistic belief or hold it less firmly (i.e., that one acquires a partial but not full defeater).18

I think there are at least two ways one could respond to this objection. One is to simply bite the bullet by conceding that the solution is only partial, so that both S₁ and S₂ should hold their theistic beliefs less firmly, but then point out that a lower degree of belief-firmness is still compatible with a degree of justification/warrant sufficient for knowledge. If so, then the problem, perhaps, does not matter all that much. Indeed, I think that it may well be the case that partial defeat—some reason(s) which makes it such that one ought to hold their belief less firmly but not completely give it up—is compatible with a degree of justification/warrant sufficient for knowledge.19 Another way to respond to this objection, however, is to argue that S₁ and S₂ do not even acquire partial defeat. One way in which one might argue this point is by defending some version of epistemic permissivism. Roughly, this is the idea that a body of evidence permits more than one justified doxastic attitude based on that evidence (cf., Kelly 2014). It is, of course, beyond the scope of this paper to attempt to fully defend this view now but let me at least try to motivate it. Consider the following scenario:

Prof. Smith and Prof. Jones were reputable scholars in their field and experts in the literary work, genre, and biographical details of Russian novelist Fyodor Dostoyevsky. Over the years, Smith and Jones would often gather at their local café to both read and discuss Dostoyevsky’s novels, perhaps reminiscent of meetings between J.R.R. Tolkien and C.S. Lewis. On occasion, the two professors engaged in a dispute as to the precise meaning of certain passages in Dostoyevky’s writings. Specifically, the two disagreed over what Dostoyevsky himself intended by such passages. Although Smith and Jones argued vehemently that the passage meant either A or it meant B, their disagreement was amicable; though they both regarded the other as wrong, they did nonetheless consider the other as reasonable.

The point of this scenario is to elicit the intuition that both Prof. Smith and Prof. Jones were reasonable in holding to their interpretation as to the intended meaning of a passage from Dostoyevsky, and in regarding the other as reasonable despite also thinking that they are wrong. It seems to me that reasonable disagreement of this sort abounds in various areas of inquiry (cf., Rosen 2001, pp. 71–72; Moffett 2007). But what can account for this? I think it might be accounted for in terms of evidential ambiguity or underdetermination (cf., Jackson and LaFore 2024). That is, where there is a sort of gap between the information or evidence within our purview and the facts out there in the world (Baker-Hytch 2024, p. 9). Where there exists this evidential gap and it is neither too remote nor too proximate, we are left with a state of epistemic permissivism.

If it is right then to think that some situations are epistemically permissive, and that the dispute between S₁ and S₂ is an example of such a case, then it may well be that the satisfaction of NIRR (or IRR) ensures that their epistemic justification is not even partially defeated.20 Therefore, this would mean that premise (2) of the above argument is false in this case.

6. Conclusions

In sum, I have suggested that religious peer disagreement is a potential threat to the justification of one’s theistic belief. I referred to this concern as the epistemic problem of religious disagreement. In drawing on key insights from the Islamic Rationalist and Islamic Traditionalist religious epistemologies, I offered a response to this epistemic problem grounded in a distinction between inferential reflective responsibility (IRR) and noninferential reflective responsibility (NIRR). I argued that if successful, it shows how reflective theists may be steadfast in upholding their tradition-specific theistic belief despite peer disagreement, and even if they only do so noninferentially.21