Beliefs may be objectionable for their harmful consequences or because they have been acquired through faulty reasoning. But can we morally wrong others by the beliefs that we form about them? According to many, we can (Basu Reference Basu2018, Reference Basu2019a, Reference Basu2019b, Reference Basu2023; Basu and Schroeder Reference Basu, Schroeder, Kim and McGrath2019; Schroeder Reference Schroeder2018; Marušić and White Reference Marušić and White2018; Fabre Reference Fabre2022; Quanbeck Reference Quanbeck2023).Footnote 1 Discussion has focussed on beliefs that unfavourably stereotype their targets and beliefs about people standing in an interpersonal relationship to the believer that fall short of the way in which they ought to be regarded. Putative examples from the literature, with minor modifications, include:
Seeing a smartly dressed black man at a club dinner, forming the belief that he is an attendant rather than a fellow invitee (Gendler Reference Gendler2011; Basu and Schroeder Reference Basu, Schroeder, Kim and McGrath2019: 200).
Attending a board meeting of a FTSE100 company and forming the belief that the only woman in attendance is an administrative assistant rather than a board member (Fabre Reference Fabre2022: 47).
A father learning of his daughter’s aspiration to become an engineer believes that her ambitions are frivolous (Schroeder Reference Schroeder2018: 115).
A spouse seeing a wine stain on the clothing of their recovering alcoholic partner, coming to the belief that they have relapsed (Basu and Schroeder Reference Basu, Schroeder, Kim and McGrath2019: 182).
It seems that the people about whom these beliefs are formed have grounds to feel aggrieved and would be reasonable in seeking an apology from the holder of these beliefs for having such beliefs about them. We are apt to criticise others for having such beliefs or feel guilty about having them ourselves (Smith Reference Smith, Bagnoli and Greenspan2011; Basu Reference Basu2023).
Supporters of doxastic wronging have offered various distinct but potentially complementary explanations of how beliefs can wrong others.Footnote 2 According to one leading proposal, the targets of these beliefs can be wronged if they are ‘falsely diminished’ by the way in which the belief represents them (Schroeder Reference Schroeder2018: 116). A second account, drawing on Strawson’s (Reference Strawson1962) work on the objective and participant stances, proposes that these beliefs fail to be properly sensitive to the personhood, moral agency, or individuality of their targets (Quanbeck Reference Quanbeck2023), or to treat them as moral agents or as persons (Fabre Reference Fabre2022: 56, Basu Reference Basu2019a, Marušić and White Reference Marušić and White2018).
This paper focusses on an important upshot of research on doxastic wronging for the debate about the problem of evil, specifically, a moral and epistemic objection to beliefs about victims of evil that are advanced in (though by no means restricted to) many theistic responses to the problem. These beliefs, as we will show in Section 1, risk committing doxastic wronging. Moreover, the risk of doxastic wronging generates a further, and hitherto unexplored, epistemological problem both for these beliefs and for belief in a prominent variety of theism – what we will call, morally perfect being theism – according to which God is a morally perfect being. The problem, in short, is that given moral encroachment, the risk of doxastically wronging victims of evil increases the evidentiary standard that needs to be met in order to justifiably believe the relevant conclusions about the problem of evil and the conclusion that God is a morally perfect being. We will set out this objection in Section 2. In Section 3, we will look at attempts to escape these problems either by questioning the assumptions on which they are advanced – not least, the denial that one can wrong someone by having a belief about them – or by evading them – by arguing, for example, that religious commitments are nondoxastic. We will argue that while these strategies may address the moral and epistemic problems we raise in this paper, they can only do so by incurring other significant theoretical costs.
1. The problem of evil and risky beliefs
The problem of evil presents an epistemic challenge to theists to account for evils that appear to render belief in God unreasonable. Take, for example, unconscionable evils: evils that no morally perfect agent would permit to occur.Footnote 3 Since the occurrence of an unconscionable evil entails that God cannot be a morally perfect agent – at least insofar as God is omnipotent and omniscient – then any theist who believes that God is a morally perfect agent (‘morally perfect being theist’ hereafter) is committed to the following belief:
-
1. None of the evils in the world are unconscionable.
Following from (1) are beliefs about occurrences of actual evils:
-
2. The international slave trade was not an unconscionable evil.
-
3. The Rwandan genocide was not an unconscionable evil.
and so on for any instance of evil. These beliefs are examples of what we will call a ‘Not Unconscionable’ or ‘nu’ belief: the belief of some prima facie unconscionable evil, or of evils in general, that they are not unconscionable. To be clear at the outset, while all morally perfect being theists are committed to nu beliefs, nu beliefs are not confined to morally perfect being theists. An atheist, for example, might think that some prima facie unconscionable evils are not as they appear and, so too, might theists who are not morally perfect being ones. However, nu beliefs held by morally perfect being theists will be the more prominent target in this paper. This is because morally perfect being theists cannot – in response to reasons to reject nu beliefs – withhold or dispense with their nu beliefs without also undermining their theistic belief.
nu beliefs will extend to beliefs that the victim(s) of horrendous evils have not endured an unconscionable evil. Horrendous evils are evils so awful that they constitute ‘prima facie reason to doubt whether the participant’s life could (given their inclusion in it) be a great good to him/her on the whole’ (Adams Reference Adams1999: 26). For example,
-
4. Anyone who suffered in the Chalisa famine did not experience an unconscionable evil.
is an nu belief about an evil, horrendous for many of its participants, that is entailed by (1).
It can be seen that nu beliefs risk doxastic wronging. Consider the two accounts of doxastic wronging outlined in the introduction. If one has an nu belief then, following Schroeder’s characterisation of doxastic wronging, one is at risk of wronging victims of evil by falsely judging their situation to be less bad than it is. This is because, for any given instance of evil, it is morally worse for that evil to be unconscionable rather than conscionable, i.e., it is morally worse if an evil is incompatible with the existence of a morally perfect agent rather than compatible. While not stated in these exact terms, this conclusion is implicitly endorsed by the majority of participants in the evidential problem of evil debate. Following Rowe’s (Reference Rowe1979) landmark paper, it is customary within the debate to delineate between whether an evil is gratuitous, i.e., whether God could have prevented it without thereby losing some greater good or permitting some evil equally bad or worse, or not. Given this understanding of gratuitous evil, an evil is morally worse if it is gratuitous rather than not. For if an evil, e, is not gratuitous in this world, but is gratuitous in some nearby possible world, that means that, in the nearby possible world in question, e’s net disvalue is greater than it is in this world, thereby rendering it morally worse. But given that it is widely accepted that gratuitous evils are incompatible with the existence of a morally perfect God,Footnote 4 and gratuitous evils are, by that line of reasoning, unconscionable evils (but not vice versa), it follows that most participants in the debate are implicitly committed to the conclusion that an evil is morally worse if it is unconscionable rather than conscionable.Footnote 5
If an evil is morally worse if it is unconscionable rather than conscionable, then if there are unconscionable evils, nu beliefs present similar moral failings to other examples of doxastic wronging by falsely diminishing the moral significance of the evil in question by rejecting its impermissibility and, correspondingly, putting into question and downplaying the extremity of the experiences that the targets of the belief have endured.
Similarly, on a Strawsonian account, nu beliefs exhibit comparable risks to other examples of doxastic wronging. To adopt an objective attitude is to set aside the feelings towards others that are characteristic of interpersonal relationships and instead, as Strawson says, to ‘see him, perhaps, as an object of social policy … as something certainly to be taken account, perhaps precautionary account, of; to be managed or handled’ (Reference Strawson1962: 194). Given that morally perfect being theists are committed to the nonexistence of unconscionable evils and that God’s moral perfection seems to entail that God has a divine plan for permitting all of the evil that occurs in our world – it seems contrary to the notion of a morally perfect being that they would allow evil or suffering to occur for no reason – it seems that in holding nu beliefs, morally perfect being theists at least risk adopting an objective attitude towards suffering by viewing sufferers as the object of God’s divine policy for allowing evil. There are, as Strawson (Reference Strawson1962: 195) recognises, contexts in which taking an objective attitude can provide ‘a refuge, say, from the strains of involvement; or as an aid to policy; or simply out of intellectual curiosity’. Indeed, there may be some circumstances in which victims of evil find consolation in the belief that their own suffering fits within God’s benevolent divine policy for allowing evil.Footnote 6 Notwithstanding, the objective stance risks doxastic wronging, even if it is sometimes justified. As Basu points out: ‘Sometimes we must [take the objective stance toward another], either because our job demands it, e.g., a therapist, or because it is simply too demanding to always take the participant stance toward others. However, there is a risk in doing so’. (Reference Basu2023: 5). nu beliefs present a similar risk but lack the pressing pragmatic reasons for taking an objective stance found in medical practice and other occupations, particularly when considering that the primary debate from which nu beliefs are derived, the problem of evil, is typically seen to address a ‘theoretical’ rather than a ‘practical’ or ‘pastoral’ problem (cf. Adams Reference Adams1999: 184; Plantinga Reference Plantinga1974: 63–4; and van Inwagen Reference Van Inwagen2006: 4).
Accordingly, a modest extension of work on doxastic wronging to the sphere of religious belief yields:
-
5. (NR) nu beliefs risk wronging the victims of evil in question.
There are, of course, many other religious beliefs about which a similar worry could be raised. For example, stereotyping of non-religious groups or adherents to other religions is a commonplace part of religious belief traditions. Some of these beliefs are not only at risk of doxastic wronging but have had destructive and sometimes lethal consequences. However, nu beliefs merit special attention for two reasons. First, the people targeted or participants in the evils described by nu beliefs are among the most disadvantaged of human beings, many of whom are – as a result of being in this condition – voiceless, and number among them individuals who have endured the worst evils that any human beings have suffered. So, the moral concerns raised by the proponents of theories of doxastic wronging are particularly salient in this case. Second, many otherwise prejudicial religious beliefs can be revised or eliminated without undermining other core religious commitments. In contrast, nu beliefs, as noted above, are an inescapable consequence of belief in a morally perfect being. Accordingly, (NR), if correct, has wide ramifications.Footnote 7
2. The problem for nu beliefs and morally perfect being theism
Most accounts of doxastic wronging – and the ones that we will focus on in this paper – opt for what Enoch and Spectre (Reference Enoch and Spectreforthcoming) refer to as a ‘modest’ version of the theory according to which the epistemic and moral norms of belief do not come into conflict. According to such accounts, beliefs that wrong are always epistemically impermissible (cf. Basu and Schroeder Reference Basu, Schroeder, Kim and McGrath2019: 199).Footnote 8 Showing that nu beliefs are at risk of doxastic wronging, therefore, need not be seen as a reason for not having them, for regardless of whether a belief risks doxastically wronging, insofar as it is epistemically justified, it does not constitute a doxastic wrong. As such, morally perfect being theists and other proponents of nu beliefs could concede (NR), while also arguing that nu beliefs do not wrong anybody because they are epistemologically in good standing. If we take Schroeder’s theory of doxastic wronging, for instance, beliefs are wrong if they falsely diminish,Footnote 9 so the success of a moral objection to nu beliefs based on their doxastically wronging victims of evil depends on the falsity of nu beliefs. Accordingly, criticising nu beliefs for doxastic wronging would be putting the cart before the horse: the falsity of nu beliefs needs to be established to vindicate the moral objection to them.
Let us concede, then, that moral reasons, like (NR), will not require us to reject nu beliefs if we have good epistemic reasons for thinking that nu beliefs are true. This concession, however, does not warrant sidelining moral considerations about nu beliefs until their truth is settled. One reason for this is that doxastic wronging is widely seen as best explained by moral encroachment theories of epistemic justification. According to these accounts, while the justification of a belief depends on the evidential case in support of it, the pragmatic stakes involved in having a belief raise the evidential demands that need to be satisfied for the belief in question to be justified. These pragmatic stakes include moral ones. Accordingly, a morally risky belief – such as one that potentially harms those whom the belief is about – has a raised standard of evidential support needed to justify it:
As the moral considerations against belief increase, so does the evidence that is required in order to epistemically justify that belief (Basu and Schroeder Reference Basu, Schroeder, Kim and McGrath2019: 201–202).
The evidentiary standard … is higher given the moral risks involved. Just as practical stakes can make it harder to believe, be justified, or know, so too can the moral stakes (Basu Reference Basu2019b: 13).
Now, as we have already noted, nu beliefs risk doxastic wronging and, given this moral risk, nu beliefs have comparatively heightened moral stakes. And, in many instances, these stakes will be considerably heightened. One distinctive kind of moral stake-raising factor to which Basu (Reference Basu2019b: 13) draws attention – and that, if Basu is right, will be applicable to many beliefs about unconscionable evil – is a historical background of institutionalised oppression:
Underrepresented groups are more often mistaken for employees because of the colour of their skin and the racist institutions that make their skin colour such a determining factor of their inability to gain access to more prestigious employment opportunities. Being mistaken in this context, namely one in which you’ve historically been excluded, is a moral wrong that is absent in spaces where that historical disadvantage is lacking.
While the subject matter is different, many nu beliefs concern people with histories of historical disadvantage that can be seen as contributing to the moral risk in having them. In the case of the nu belief, (5) the Holocaust was not an unconscionable evil, for example, there is a thoroughly documented history of the ways in which military, medical, and corporate institutions were organised for systematic murder on a vast scale.Footnote 10
Given (NR) and moral encroachment, it follows that:
(NE) The risk of doxastic wronging by nu beliefs raises the evidential support required to justify them.
Moreover, since:
-
6. A morally perfect God exists.
-
7. If there is a morally perfect God, there are no unconscionable evils, i.e., evils that a morally perfect God would not permit to occur.
together entail (1) – none of the evils in the world are unconscionable – the higher evidential standard needed for justifying nu beliefs extends to belief in a morally perfect God:
-
(PBE) The risk of doxastic wronging by nu beliefs raises the evidential support required to justify belief in the existence of a morally perfect God.
The problem, therefore, is that the doxastic risk resulting from nu beliefs not only diminishes the prospects of justifying them, but also diminishes the prospects of justifying belief in morally perfect being theism.
To measure the seriousness of the challenge this problem presents, we need to consider the available support for (1) and (6). Perhaps the evidence for these beliefs is already sufficient to meet a higher evidentiary standard (i.e., one that is higher than would be needed to justify nu beliefs if they posed no doxastic risk). Though evidence for (6) is evidence for (1), evidence for (1) is not necessarily evidence for (6). Accordingly, it is worth examining the evidential cases independently, even if it turns out that the best evidential case for (1) just is the best case for (6). We will start with (1) then turn to (6). For readers already persuaded of the challenge that our arguments present for justifying belief in (1) and (6), we will turn to other responses to (NE) and (PBE) in Section 3.
A traditional method of supporting (1) is to flesh out the details of a theodicy. A theodicy is an argument or narrative that aims to demonstrate the compatibility of a morally perfect being with the existence, kinds, intensity, and/or distribution of evil and suffering in the world. Theodicies are many and varied, and the topic has generated a vast literature. A far from exhaustive list of examples include: every evil is offset by an outweighing good (Swinburne Reference Swinburne1998: 238); suffering, even if serious and undeserved, is permissible for God as part of a redemptive plan aimed at the benefit of the sufferer (Stump Reference Stump1985: 409–411); post-mortem experiences will ‘render worthwhile’ and morally justify the suffering endured in this world: ‘sufferings – which for some people are immense and for others relatively slight – will in the end lead to the enjoyment of a common good which will be unending and therefore unlimited, and which will be seen by its participants as justifying all that has been endured on the way to it’. (Hick Reference Hick1985: 340–41); all evils are given positive meaning through their being part of an ‘organic unity with a great enough good within the context of his/her [the sufferer’s] life’ (Adams Reference Adams1999: 31); and God ensures that all evils, even horrendous ones, will be ‘not only balanced off but endowed with positive meanings, meanings at least some of which will be recognised and appropriated by the participant him/herself’ (Reference Adams1999: 205). Setting aside the relative merits of these proposals, they are unavoidably speculative: the real details of God’s plan, if there is one, are unavailable to us (a consideration which often motivates sceptical theism, to which we will return presently). Even if we grant that these theories could in principle justify or make sense of prima facie unconscionable evils – by many accounts, an enormously generous concession – given their speculative nature, they do not, at least absent further evidence for thinking them not just possible, but plausible, provide reason to endorse (1).
Notably, sceptical theism, perhaps the most prominent contemporary theistic response to the problem of evil in the philosophy of religion, eschews attempts to spell out the details of a theodicy on the grounds that God’s morally sufficient reasons for permitting suffering may be ones that we are unable to grasp or, at least, ones that we should not expect to be scrutable to us. Accordingly, sceptical theist arguments do not support (1) but provide reasons to doubt whether we are epistemically well situated to either affirm or deny (1). However, insofar as sceptical theists are morally perfect being theists – as many seem to be – then sceptical theists are committed to (1) in virtue of their theological commitments, their sceptical arguments notwithstanding.
Given that neither sceptical theism nor theodicies seem to provide reason to think that there are no unconscionable evils, particularly when considering the higher evidential standard that needs to be met in order to justifiably believe such a conclusion, the case for believing that there are no unconscionable evils will seemingly rely on the evidence for the belief that there is a morally perfect God that would not allow them, i.e., (6). In that vein, a more promising and well-established supporting argument seems available: we can rationally form religious beliefs, in a similar way to our acquisition of perceptual beliefs, from religious experiences (Alston Reference Alston1991, Plantinga Reference Plantinga2000). According to Plantinga, we are disposed to form beliefs about God’s action in the world from which we can infer God’s existence. For example, someone may form the basic belief:
-
8. God is speaking to me.
From which it is inferred:
-
9. God exists.
Then, according to Plantinga, provided the acquisition of (8) meets various standards of grounding, the person who acquired (8) would be warranted in believing (8) and (9). And, if someone is warranted in believing that there is a God, then they may be well-positioned to meet the higher evidential demand needed to overcome the risk of doxastic wronging.
This response, however, does not address the epistemic problem raised by (PBE). The epistemic challenge is not to belief in God but to belief in God with a specific moral profile, i.e., one that is morally perfect. Justifying belief in (6) looks like a much heavier lift. It is not clear what religious experiences one could have that would warrant an inference to (6). Even granting that the perfect being theist could – following Plantinga – form basic beliefs like:
-
8b. God loves me.
-
8c. God cares for me.
-
8d. God is good.
None of these beliefs justify (6).
Religious experiences are, of course, only part of a cumulative evidential case that might be made for (6). The problem is that if we compare the evidential case for (9) – about which opinion is, to put it mildly, divided – with the case for (6), the evidential case for (6) prima facie seems more precarious. Among standard theistic arguments for the existence of God, few directly lend support to God’s moral perfection. As van Inwagen (Reference Van Inwagen1991: 138) has observed, ‘[w]hatever the individual merits or defects of [these] arguments, none of them but the “moral argument” (and perhaps the ontological argument) purports to prove the existence of a morally perfect being. And neither the moral argument nor the ontological argument has many defenders these days’. Even this pessimistic assessment may be too generous. Setting aside the question of whether either the moral or ontological arguments are good arguments for theism, neither, absent further argument, clearly yield (6). Moral arguments typically posit the existence of God as the best explanation for objective moral facts, rather than specifically supporting morally perfect being theism.Footnote 11 Similarly, ontological arguments are used to support the existence of an absolutely perfect being – one that is greatest possible or conceivable – but do not require that God is morally perfect. In fact, according to one recent account (Murphy Reference Murphy and Draper2019: 93), not only is absolute perfection not a reason for thinking that God is morally perfect, the ‘uncontroversial’ divine perfections – perfect power, perfect knowledge, perfect rationality, and perfect freedom – entail that ‘moral perfection is not among the absolutely perfect being’s perfections’ (emphasis ours). One could, of course, supply additional argumentation in order to bridge this gap between the conclusion that a godlike entity with some divine attributes exists to the conclusion that the godlike entity possesses all of the traditional divine attributes, including moral perfection – this gap-bridging challenge is typically referred to as ‘the gap problem’Footnote 12 – but we do not need a detailed evaluation of these argument to see that justifying (6) is a significantly more challenging prospect that justifying (9). There is less evidence to work with, and supplementary arguments would need to be vindicated. Moreover, there are arguments against (6) – but not (9) – that would need to be addressed.Footnote 13
It appears, therefore, that the challenge posed by (NE) and (PBE) is serious. To justifiably believe either (1) or (6) in light of (NE) and (PBE), the evidential support for them needs to be particularly strong. The onus is on proponents of these beliefs to show that it is.
3. Responses
nu beliefs, we have argued, risk doxastic wronging. This presents a moral and an epistemological objection both to having these beliefs and to morally perfect being theism. Having reviewed the evidential resources available to supporters of these beliefs, we have argued that there is reason to think that the problems raised by these objections are significant. In this section, we will consider three lines of response that target the assumptions on which our argument is advanced: (i) reject doxastic wronging, (ii) allow that beliefs can wrong but reject moral encroachment, (iii) deny that religious commitments to nu ‘beliefs’ or to morally perfect being theism require belief. In each case, we will argue the response either does not do away with both the epistemic and moral problem and/or it generates further significant difficulties.
3.1. Reject doxastic wronging
While this offers a straightforward way of avoiding (NR), (NE), and (PBE), there are a number of challenges to pursuing this response. The theory that beliefs can wrong is, as we saw in the introduction, supported by the evidence of the moral reactions and sensitivities that some beliefs generate: we can be wounded by the beliefs that others have about us, be apt to demand an apology from the believer, and feel guilty about having such beliefs. The immediate requirement of denying that beliefs can be wrong, therefore, will be to supply an alternative account of this evidence.Footnote 14 A further difficulty with this response is that moral constraints on propositional attitudes is a familiar theme in religious thought, particularly within the Abrahamic religious traditions (cf. Schroeder Reference Schroeder2020). For instance, the biblical injunctions against sinning not just in action but in one’s ‘heart’ (Matthew 5:28; Leviticus 19:17) suggest that our attitudes to others can be wrong in a way that is comparable to our actions. Basu and Schroeder (Reference Basu, Schroeder, Kim and McGrath2019: 181) draw a similar conclusion from the Eucharistic confession to God in the Book of Common Prayer: ‘we have sinned against you in thought, word, and deed’ (our emphasis), with some versions adding to the confession ‘and against our fellow men’; thereby extending the scope of those who can be sinned against in thought to include one’s ‘neighbour’.Footnote 15 And, within the Islamic tradition, Quran 49:12 warns believers to abstain from making negative assumptions about others as doing so can be sinful. Therefore, to the extent that morally perfect being theism is put forward as an articulation of the theistic view of a practising religious traditions – and, to the extent that adherents to these traditions do not think it viable to reject the teachings of these traditions, as we think is typically the case – or that nu beliefs are put forward as ones that practitioners of a religious tradition should have, rejecting doxastic wronging will need to avoid conflicting with one’s other religious views.
Could morally perfect being theists avoid the second difficulty by conceding that nondoxastic propositional attitudes may be wrong but maintain that beliefs do not? The problem is that the moral/religious constraints noted seem to be directed against representing the world in a certain way. Thus, if it is sinful to think a proposition – for example, to think of oneself as equal to God – it will also be sinful to believe it. For this strategy to work, one would need to explain why beliefs would be morally less risky than thoughts. One line of objection to doxastic wronging, widely discussed in the literature, focusses on the limited voluntary control that we have over our beliefs. We cannot be obliged to do something over which we have no voluntary control, and, so the argument goes, we have insufficient control over our beliefs to be held morally accountable for possessing them. However, this response is not widely seen as persuasive because we do seem to have sufficient indirect control over our beliefs to allow for their moral accountability (see Smith Reference Smith2005; Basu and Schroder Reference Basu, Schroeder, Kim and McGrath2019), a point also agreed on by some critics of doxastic wronging (Fritz and Jackson Reference Fritz and Jackson2021).Footnote 16
3.2. Reject encroachment
If moral encroachment is rejected, then the moral risk of nu beliefs and belief in a morally perfect being will not raise the evidential stakes for having these beliefs. However, this manoeuvre raises other problems. Moral encroachment provides a neat response to what Basu and Schroeder (Reference Basu, Schroeder, Kim and McGrath2019) call the coordination problem: how moral and epistemic norms of assessing belief fit together if doxastic wronging is possible. As Basu and Schroeder (Reference Basu, Schroeder, Kim and McGrath2019: 194–5) summarise the problem:
if beliefs can be wrong, then there will be non-coordinating epistemic and moral standards governing belief, because while the epistemic standards governing belief depend only on the evidence, whether a belief is wrong depends instead on moral considerations. And this could easily lead to troubling choices between being rational and being good.
Moral encroachment theories can resolve this problem by coordinating the epistemic and moral standards governing belief as follows: whether a belief is epistemically justified depends upon whether it clears a certain evidential threshold, but the evidential threshold for a justifiable belief itself depends upon the moral stakes of holding the belief such that the more morally risky the belief, the higher the evidential threshold that needs to be cleared for the belief to be epistemically justified.
If moral encroachment is rejected, morally perfect being theists will not only need to find another explanation for how doxastic wronging is possible without moral encroachment – assuming, of course, that morally perfect being theists think it is possible to doxastically wrong others – but also explain, given the moral risk of nu beliefs, how the moral reasons not to believe a proposition comport with epistemic reasons in its favour, if they do at all. If, following Gendler (Reference Gendler2011), Mugg (Reference Mugg2013, Reference Mugg2020), and Puddifoot (Reference Puddifoot2021), moral and epistemic reasons to believe a proposition do not comport – perhaps, to follow Feldman (Reference Feldman2000), they are incommensurable – and one can have moral reasons to not believe a proposition that they have epistemic reasons to believe, then one is left with what Mugg (Reference Mugg2020) calls ‘a tragic dilemma’: ‘we can either be epistemically rational, or we can be moral, but we cannot be both’ – a dilemma that looks particularly worrisome for morally perfect being theists who have a vested interest in being both moral and rational.Footnote 17 So as to guard against misunderstanding, we are not suggesting that this dilemma is irresolvable. Our point is simply that even if morally perfect being theists reject moral encroachment and also deny that moral and epistemic reasons to believe a proposition comport, they nevertheless still owe an explanation for both how doxastic wronging is possible without moral encroachment, as well as how the moral and epistemic reasons to believe in a morally perfect being – and, correspondingly, that there are no unconscionable evils – provide reason to either be moral or epistemically rational.
As recently defended by Dandelet (Reference Dandelet2023), one option for explaining how the moral reasons not to believe a proposition comport with epistemic reasons in its favour that does not appeal to moral encroachment is moral evidentialism. According to this view, moral considerations can feature in our deliberations and justification of beliefs, but what one morally ought to believe depends only on one’s evidence (Reference Dandelet2023: 83). This variety of evidentialism would be particularly promising for the morally perfect being theist if the evidence in favour of (1) and (6) was strong, for if it was, then one morally ought to believe (1) and (6), the risk of doxastic wronging notwithstanding. However, based upon our previous review of the evidence in favour of (1) and (6), it is not clear that it is. Moreover, as Dandelet (Reference Dandelet2023: 91) argues, in contexts of uncertainty, prudential considerations are unavoidable for familiar Jamesian reasons:
[W]hen we find ourselves unsure about whether p is true … in so far as we continue to deliberate about what to believe—that is, in so far as we continue to look for reasons weighing in one direction or another, with the aim of coming to a belief or a suspension of judgement—we not only can but must invoke considerations that go beyond the evidence bearing on p.
Therefore, even if one is a moral evidentialist, unless the evidence in favour of (1) and (6) is strong, the moral risks that we have already reviewed will still be reasons against forming nu beliefs.
Furthermore, moral evidentialism leaves the formation of nu beliefs open to other practical and moral objections not yet noted.Footnote 18 For example, the kind of speculative reflection on evil involved in developing nu beliefs may be a morally risky enterprise for the thinker. It invites participants to think of evils, however extreme, as conscionable and wears away the resistance to this idea. If there are, in fact, unconscionable evils, then this practice will risk encouraging a sceptical attitude towards the seriousness of the sufferings of others. A second kind of ethical risk is reputational. To have rational nu beliefs requires, on an evidentialist account, an assessment of the evidence that evils – in particular, ones that are prima facie unconscionable – are justified or, at least, morally permissible. But there is a reputational risk incurred by undertaking the enquiry. By way of comparison, suppose that a think tank conducts a review of the evidence of the economic and social benefits that could be achieved through the involuntary sterilisation of a minority group. Such a project would raise questions about the agency’s motives in prioritising this enquiry; to say that the research was aimed at finding the truth would not address concerns about the moral propriety of pursuing it.
While rejecting moral encroachment, therefore, does dispense with the immediate problem of the heightened epistemic requirements for justifying nu beliefs and morally perfect being theism, such a response requires either providing an alternative solution to the coordination problem or being left with the tragic dilemma and, additionally, does not do away with the moral and practical risks of acquiring and holding these beliefs, risks that function as reasons not to hold them. That is not to say that these risks therefore outweigh whatever moral or practical reasons there might be for acquiring and holding these beliefs, but it is to say that, if these risks do not outweigh the moral or practical reasons for holding nu beliefs – a conclusion morally perfect being theists are committed to – it remains to be seen why.
3.3. Theodicy without belief
The argument for (NE) and (PBE) relies on the assumption that morally perfect being theism and nu beliefs are believed by their proponents. But perhaps they need not be believed if they are a matter of faith? While propositional faith has traditionally been seen as a belief commitment,Footnote 19 recent accounts have focussed on nondoxastic alternatives. For example, according to William Alston (Reference Alston, Jordan and Howard-Snyder1996), faith that p is a positively valanced or pro-attitude towards p (Reference Alston, Jordan and Howard-Snyder1996: 12), but the content of propositional faith need not be believed so long as it is accepted. Following Cohen (Reference Cohen1992), acceptance that p is understood as a commitment to go along with p, that is, to use p as an assumption in the accepter’s theoretical and practical reasoning. Acceptance, unlike belief, is typically taken to be voluntary and may be prompted by the pragmatic benefits of adhering to the proposition rather than its truth. Other nondoxastic accounts have a similar structure: the agent with faith that p need not believe that p provided that they assent or ‘say yes to’ (Schellenberg Reference Schellenberg2005) assume (Swinburne Reference Swinburne2005; Howard-Snyder Reference Howard-Snyder2013), go along with, acquiesce to (Buchak Reference Buchak, Chandler and Harrison2012), or place their trust in (Audi Reference Audi1991, Reference Audi2011) the proposition in question. Moreover, agents may have faith in p while doubting its propositional content (Howard-Snyder Reference Howard-Snyder2013: 361) or be prepared to affirm their faith but not their belief in p (Buchak Reference Buchak, Chandler and Harrison2012) or even, more boldly, may have faith in p while disbelieving p (Whitaker Reference Whitaker2019). According to Alston (Reference Alston, Greco, Mele and Timmons2007: 136), ‘a sizable proportion of contemporary, sincere, devout Christians are accepters rather than believers’.
Setting aside objections to nondoxastic accounts of faith (Scott Reference Scott2020; Mugg Reference Mugg2021), suppose the morally perfect being theist does not have belief but instead has nondoxastic faith in (6) – a morally perfect God exists – and the nu propositions that follow from it, where nondoxastic faith that p is or involves having a belief-like propositional attitude that p. Clearly, in so doing, those who hold these faith commitments are not thereby at risk of doxastic wronging. However, their faith commitments still exhibit the same morally risky characteristics as nu belief set out in Section 1: they risk falsely diminishing or failing to extend participatory attitudes towards victims of evil. In fact, the moral objection to these propositional attitudes looks even more compelling if they are nondoxastic faith commitments. There are two reasons for this. First, while, as we noted earlier, there is debate on whether we have a sufficient degree of control over our beliefs to be held morally responsible for having them, there is no similar controversy about nondoxastic faith. A nondoxastic faith commitment involves a choice to commit to a proposition that is guided in part by pragmatic considerations, rather than being fixed by what the agent finds to be true. As such, the moral concerns that we have advertised provide pragmatic reasons to choose differently and avoid faith in morally perfect being theism or nu propositions. Second, a nondoxastic faith commitment involves a decision to act on a proposition, for example, to make choices on the basis that it is true, to endorse it, or to promote it to others. It is not just the attitude that is morally risky, but the potentially objectionable actions that the agent takes as a result of taking on this faith commitment.Footnote 20 To pursue such a policy for propositions that one does not have grounds for believing heightens the agent’s responsibility for the moral risk. The problem we are pressing in this second reason is not that a nondoxastic faith commitment in (6) is morally more objectionable than belief in (6) because of the potentially objectionable downstream consequences that it may result in, but that one’s moral responsibility for these potentially objectionable consequences is heightened if the commitment is nondoxastic rather than doxastic, given the degree of control exercised over the former but not the latter.
4. Conclusion
We have argued that nu beliefs – the belief of some prima facie unconscionable evil, that it is not unconscionable – risk doxastic wronging. This risk presents a moral and epistemic objection to holding nu beliefs: the moral objection being the risk of doxastically wronging victims of evil, the epistemic objection being the increased evidentiary standard required to justify the beliefs given moral encroachment. These objections are distinctively pressing for morally perfect being theists, for whom nu beliefs are a consequence of their view about God’s nature as a morally perfect being. Were the evidence for nu beliefs and morally perfect being theism strong to begin with, these objections could be met. However, as we have argued, it is not clear that the evidence for these beliefs is sufficiently strong to overcome these objections.
We have assessed three strategies for avoiding the consequences of our argument – deny doxastic wronging, deny moral encroachment, or endorse nondoxasticism – and found that they either do not resolve both the moral and epistemic objection and/or they raise further problems for the supporter of nu beliefs. If we are correct, then not only have we demonstrated a notable and hitherto unforeseen upshot of the position that we can morally wrong others by the beliefs that we form about them – namely, that nu beliefs risk doxastic wronging – but we have also demonstrated why, as a result, it remains to be seen whether nu beliefs or belief in a morally perfect God are epistemically justifiable.Footnote 21
Open access
