Introduction

Issues concerning the divine willFootnote 1 have always drawn great interest from theologians and philosophers alike–and for good reasons. Especially monotheistic religions tend to include numerous and substantial claims about the contents of the divine will, and also about the consequences of either adhering to the divine will, or alternatively violating it. In Abrahamic religions, holy books such as the Torah, the Quran, and the Bible are well known for making such claims. For instance, the First Epistle of John claims that: “And the world is passing away along with its desires, but whoever does the will of God abides forever.” (1 John 2:17). In the Lord’s Prayer, the most influential prayer in the Christian tradition, it is asserted with hope that “[…] your will be done, on earth as it is in heaven”Footnote 2. It is also very much worth noting that the concept of “sin” plays a very central normative role in various religious traditions–“sinning” being defined as a person acting against the divine will. Indeed, one of the most prolific contemporary authors on the subject, philosopher John Hare (2015, vii), is not exaggerating when he writes that in Abrahamic religions it is very often believed that “…God’s will and command settle how we ought to live.”

In this article, I investigate the possibility of “non-accidentally” adhering to the divine will. To say something quite uncontroversial about the notion of “non-accidentality”, it has played an essential role in philosophical discussions concerning the role a person herself plays in explaining her doing a given thing. When a person φ:sFootnote 3 merely accidentally, the person does not play the central role in explaining her φ:ing. Conversely, when a person φ:s non-accidentally, the opposite is the case: the person does play the central role in explaining her φ:ing. In other words: when a person φ:s non-accidentally, her φ:ing is attributable to her.

More precisely, my focus in this article is on a modal condition of non-accidentality, which is very frequently employed in the literature. Although I give some reasons to find the modal condition outlined interesting, I do not intend to argue that it is by any means the only way to look at non-accidentality. One reason for this has to do with the fact that causal explanations are plausibly dependent on various background assumptions, which can vary between contexts.Footnote 4 For this reason, saying that a person herself is a salient cause of her φ:ing is somewhat ambiguous. Another reason has to do with the fact that there seem to be fundamentally different ways of thinking about what kind of explanation we are looking for in the first place when a person non-accidentally φ:s. For instance, just what role one’s agency and dispositions need to play, and what kind of agency and dispositions are the most relevant ones, are very ambiguous matters, and I am not at all confident that there is some single unquestionably correct answer here.Footnote 5 Indeed, it seems to me to be the case that it is best to just talk about kinds of accidentality, rather than about accidentality per se.

The article is structured around a skeptical argument I formulate against the possibility of non-accidentally adhering to the divine will. My aim is to draw attention to the existence of the skeptical argument on the one hand, and to outline reasonable strategies for denying one of its premises on the other, since the premise has not received attention to date. The argument is motivated by the observation that many authors who have theorized about non-accidentality have given reliable reasoningFootnote 6 a central role. These accounts often associate non-accidentality with a person’s actions being guided by normative factsFootnote 7. Guidance is commonly characterized in terms of a person having her normative reasons to do something as her motivating reasons to do it; that is, the person’s normative reasons to act in a certain way are the motivational basis on which she acts. This is important, because many authors have argued–quite credibly, in my eyes—that one can only be guided by a normative reason, if one has a robust epistemic access to the fact that is one’s normative reason to act.Footnote 8 Further, some authors (most prominently Sliwa, 2016, 3–4) have explicitly presented accounts of non-accidentality, according to which one has to reliably reason about what a given norm requires one to do in order to adhere to it non-accidentally. This role given to reliable reasoning seems at first glance to give rise to problems in the context of adhering to the divine will, because many philosophers of religion have argued that it is very hard for us humans to reliably reason about the requirements of the divine will – or about religious matters in general.

We can formulate the skeptical argument emerging from these observations in the following way:

  • 1. At least in many cases it is not possible to reliably reason about what the divine will requires one to do.

  • 2. Without being able to reliably reason about what the divine will requires one to do, it is not possible to adhere to it in a non-accidental manner.

  • 3. Thus, it is not possible, at least in many cases, to adhere to the divine will in a non-accidental manner.

In the next section, I briefly give reasons to accept the first premise, although it is not my intention to defend it in detail here; rather, I aim to illustrate that since the first premise has at least some credibility and is something that various authors take to be true, it is important to assess the second premise. In the third section I introduce the modal condition of non-accidentality I operate with here. In the sections four and five I set out to assess the second premise, and in doing so point out promising strategies for denying it.

Before moving on, one final remark is in order: my arguments here are conditional on a certain conception of the nature of the divine will. First, I of course take it for granted here that the divine will is actually something that exists, which is something that many people would deny – atheists are the most obvious case in point. Second, I conceive here of the divine will as something that one actually can adhere to. This might be denied by, for example, someone who thinks that there is an evil divine being who wills that we act in ways that are not simply possible (maybe the evil being wills that we stop being affected by the laws of nature), and punishes us for this failure. And third, I conceive of the divine will as something that one can actually fail to adhere to. This is something those thinkers would have a problem with who conceive of the divine will as something that is always fulfilled regardless of what we do (for instance, one could regard this as a great-making feature of God). Although this conception of the divine will I am concerned with here is not universally shared, it merits attention because of its widespread acceptance in religious thought.Footnote 9

A defense of the first premise

P1: At least in many cases it is not possible to reliably reason about what the divine will requires one to do.

How could one be aware of the requirements of the divine will? The proposed ways include reading about the divine will from Holy Texts, being directly informed about the divine will during a prayer or a religious experience, using one’s God given specialized faculties, and being taught by one’s parents or religious authorities about the contents of the divine will. (see Mouw, 1990, 8; Evans, 2013, 39–41) However, in contemporary philosophy, many are highly skeptical about the prospects of reliable reasoning about religious matters, or at least about a significant portion of religious matters. In what follows, I outline some of the most important reasons for the skepticism.

Recently, perhaps the most discussed skeptical arguments have arisen from the epistemology of disagreements literature, which has investigated the consequences disagreements among one’s epistemic peersFootnote 10 have for one’s epistemic situation. Now, there are religious disagreements about the contents of the divine will as much as there are disagreements about anything else. Of course, certain principles are something that pretty much every practioner of religion agrees on; for instance, pretty much everyone agrees that it is in accordance with the divine will to not arbitrarily hurt others, and that it is in accordance with the divine will to be grateful for the good things gods do for us. However, it is equally easy to find very wide ranging and persistent disagreements about the divine will in which each side seems to include equally intelligent and competent people. For instance, to take a contemporary issue that is fiercely debated, does the divine will allow abortions, or does it forbid them? For a more historical example, Christian theologians have debated for a very long time whether Christians should ever take part in wars. Some follow St. Augustine in thinking that Christians can, and even should, take part in certain kind of “just wars”, but others, like the Quakers, have advocated uncompromising pacifism as the ideal of Christian life. Even if equally intellectually competent practioners of religion could agree pretty much universally about certain principles concerning the contents of the divine will, these kinds of debates will likely not end any time soon.

Because some parties of a disagreement have to be in the wrong, it has to be the case, then, that many people have false beliefs about the contents of the divine will at least when it comes to more specific questions, such as the one concerning whether Christians are allowed to take part in wars. Further, in the epistemology of disagreement literature, many have adopted a conciliatory position, according to which disagreements among one’s epistemic peers call for a significant reduction in one’s confidence in the contested propositions, and make consequently it harder for one’s contested beliefs to have positive epistemic statuses, such as knowledge, even if they were true. (e.g. Christensen, 2007; Elga, 2007) The conciliatory positions are motivated by the thought that retaining one’s beliefs that are contested by one’s epistemic peers would be viciously arbitrary; why, after all, should a person think that it is she who happens to be right in a situation of peer disagreement, if the disagreeing parties are indeed one’s epistemic peers? In order for a person to reason reliably about a given matter, the thought goes, the person needs to be able to refute the higher-order evidence against her beliefs being true presented by the disagreement, but if the disagreeing parties are the person’s epistemic peers, it is not clear how something like this could be reasonably done.

Following the conciliatory arguments presented in the epistemology of disagreements literature, many have recently argued that we are not capable of reasoning reliably about religious matters, because it is not possible for us to reasonably refute the higher-order evidence presented by the peer disagreement. (e.g. Feldman, 2007) Moreover, even if religious disagreements did not make religious knowledge outright impossible, they could still make obtaining it considerably harder. For instance, although John Pittard (2019) has recently argued at length against religious knowledge being impossible, he accepts that religious disagreements make it so that one has to have access to strong epistemic reasons in order to refute the higher-order evidence presented by a religious disagreement, and Pittard further concedes that many people do not meet the requirement. Religious disagreements seem to, then, pose a problem even for those who are actually right about the requirements of the divine will.

Besides the problems stemming from religious disagreements, there are other obstacles facing a person hoping to reason reliably about the contents of the divine will. For instance, some philosophers and theologians alike have highlighted the transcendent character of the divine, and argued that this makes obtaining knowledge of it very hard. This position is perhaps most commonly associated with Immanuel Kant who maintained that limited human beings have no hope of obtaining religious knowledge with their theoretical reasoning. (see Pasternack & Fugate, 2020 for an overview of Kant’s position) More recently, J. L. Schellenberg (2013) has influentially argued that human beings are unable to reliably reason about religious matters at this stage of our evolutionary history. Schellenberg maintains that one day the cognitive makeup of humans might be sufficient for grasping religious truths, but that this day is yet to come.

Finally, evidentialism has been a very influential position in the epistemology of religion–even if its heydays of the twentieth century are over. In the epistemology of religion, evidentialism is a position, according to which knowledge or justified beliefs about religious matters require one to base one’s beliefs on good evidence–the good evidence most often taken to consist of natural theological arguments. One characteristic feature of evidentialism is that having knowledge and justified beliefs about religious matters is quite demanding, because natural theological arguments for the existence of God are often very complex in nature, drawing on things like modern modal logic and cosmology. For this reason, natural theological arguments tend to require intellectual sophistication that people not dedicating their lives to the philosophy of religion just cannot possess.

Of course, evidentialism is most often formulated as a thesis about the conditions of having knowledge or justified beliefs about the existence of God, not about having such epistemic relations to what the divine will requires one to do. That said, knowledge of the existence of God is relevant for having knowledge about the requirements of the divine will, because beliefs about the divine will presuppose the existence of a divine who’s will the beliefs are about. For instance, think of someone asserting that she knows what a certain person has said at some time, but that she does not know whether the person exists. Situations such as this would seem to be most peculiar, to say the least.

I have not defended these positions to any great degree here, but it should now be clear that it is a relevant possibility in the contemporary philosophical discourse that reasoning reliably about the requirements of the divine will is at least often very hard, if not even impossible for us. This would mean that the first premise of the skeptical argument is true. Before moving on to assess the second premise of the skeptical argument, I first need to introduce the kind of non-accidentality I operate with here more clearly, and this is what I do in the next section.

Non-accidentality and modal robustness

Many philosophers theorizing about non-accidentality give a modal condition an important role (e.g. Sliwa, 2016; Lasonen-Aarnio, 2019; forthcoming; Sosa, 2007; Lord, 2015; Greco, 2010; Whiting, 2020). According to the modal condition in question, in order for a person to φ non-accidentally, she has to φ in a way that would also result in her φ:ing in a range of normal counterfactual situations–not just in the actual situation. That is, the person has to φ by manifesting a disposition that robustly manifests itself as φ:ing across a range of normal situationsFootnote 11.

Kant’s famous example of a shopkeeper is very often used as the starting point for the discussions on non-accidentality, and it nicely illustrates the role the modal condition is intended to play here. In the example, a shopkeeper gives his customer the right amount of change. Although in so doing the shopkeeper does the morally right thing, he does it only out of self-interest, because it helps his business. Kant (1998, 3–4) claims that while the action is morally right, the moral rightness is not to be attributed to the shopkeeper, because the action’s connection to the moral law is “only very contingent and precarious”, and the way the shopkeeper acted “will indeed now and then produce actions in conformity with the law, but it will also often produce actions contrary to law.” Although I do not wish to enter into exegetical debates, a very natural reading of the above remarks suggests that Kant himself embraced the modal condition outlined; the disposition the shopkeeper manifests results in a morally right action only in a very limited range of counterfactual situations, in which maximizing one’s profits and moral rightness happen to come together, and is for this reason from the moral point of view a very unreliable way of acting. For this reason, according to the modal condition, the shopkeeper does the morally right thing merely accidentally.

It is important to stress that the dispositions that are relevant in this context are required to be sufficiently rooted in the agent herself. This requirement is intended to exclude dispositions that owe their reliability to a great extent to some external influences. For a radical example, Lord (2015, 36) conceives in this context of a demon who guarantees that whenever a person is about to fail, the person actually succeedsFootnote 12. The disposition the person manifests might be reliable in a way, but since it owes its reliability to an external influence, it is not relevant in the present context – after all, it is the activity of the person herself we are ultimately interested inFootnote 13.

Although I won’t here attempt to sell you the modal condition outlined at any great length, I will say that I find the modal condition to be a very fruitful way to think about non-accidentality; it highlights in an interesting way the role a person herself plays in her φ:ing, and excludes certain kind of external influences from playing a role. Some, such as Johnson King (2020) and Paterson (2020), have denied that the modal condition is necessary for the kind of non-accidentality they are interested in. However, as this allows for mere chance – contrasted to a person’s reliable dispositions manifested—to play a role in a person’s φ:ing being attributable to her, I find a concept of non-accidentality that does not respect the modal condition at least insufficient for all of our purposes. For instance, although I think that Paterson (2020, 314) is right in claiming that people often attribute to athletes their spectacular successes even if the athletes do not manifest a reliable disposition–think of a successful hole-in-one in a hard golf course—humble athletes tend to admit that, in an important sense, their successes in tough situations are often merely accidental, as chance, rather than the athletes themselves, plays a significant explanatory role. By requiring that a person manifests a disposition that results in the outcome in question in a modally robust way, the modal condition provides an interesting way of making sense of why a humble athlete might say something like this. Similarly, assuming that a person adheres to the divine will, the modal condition provides an interesting way–although, again, not the only way—of analyzing the role the person herself played in her doing so.

The modal condition outlined here can also be defended by contrasting it to another kind of modal condition discussed in the literature. For instance, Markovits (2010, 211) distinguishes the kind of modal condition outlined here from a modal condition of a different kind, which is not ultimately concerned with whether we manifest reliable dispositions, but rather with the more general likelihood of a person successfully φ:ing across a range of counterfactual situationsFootnote 14. This way of looking at non-accidentality is not only interested in the reliability of the dispositions one manifests, but also in what dispositions one manifests across a range of counterfactual situations. I am in agreement with Markovits in that this kind of modal condition can easily have undesirable consequences. For instance, according to this kind of modal condition, even a skilled sharpshooter manifesting a reliable disposition would succeed in shooting a target merely accidentally, if she had a condition which made her unable to shoot at all most of the time. I, for one, find it much more interesting whether a skilled sharpshooter actually manifests a reliable disposition in shooting a target than whether she could or could not have been unable to shoot at all some other times, for instance. Hence, my focus in this article is on whether people manifest dispositions that result in φ:ing in a modally robust way.

We can say, then, that when a person does not manifest a disposition that results in her φ:ing in a modally robust way, her φ:ing is what we might call a modal accident. In what follows, I investigate whether one could adhere to the divine will in this kind of non-accidental way, if one cannot reliably reason about the requirements of the divine will. I first give a brief initial defense of the second premise of the skeptical arguments I formulated earlier. In the two sections after this, I outline strategies a practioner of religion could reasonably pursue in challenging the second premise. It should be noted that the strategies I outline attempt to be relevant in the sense that they do not merely describe such dispositions that robustly manifest themselves as adhering to the divine will, but which no-one pretty much ever manifests. Although such dispositions are interesting to some degree, I set out to describe dispositions which are reliable, and which also are something that some people might actually quite consistently manifest.

A defense of the second premise

P2: Without being able to reliably reason about what the divine will requires one to do, it is not possible to adhere to it in a non-accidental manner.

I now set out to give an initial defense of the second premise of the skeptical argument, although I challenge it later on. Let us begin by asking why someone might want to give reliable reasoning an important role in non-accidental acting in the first place. To understand the attraction of such position, it is useful to look at Franck Jackson’s (1991, 462–463) classic thought experiment of Dr. JillFootnote 15:

Dr. Jill

Dr. Jill needs to prescribe one of three possible drugs to a patient. Of the three candidates, one drug will kill the patient, another drug will completely cure the patient, and that the third drug will cure almost all of the patient’s symptoms so as to be almost completely healthy. But here comes the twist: Dr. Jill’s perspective is limited in such a way that she is only aware of which drug will cure most of the patient’s symptoms, and is consequently unable to reliably tell which of the two remaining drugs is the best one and which is the worst one.

Now, Jackson asks us, what should Dr. Jill do? There is a lot that could be unpacked here, of course. That said, one way of looking at the case is to note that Dr. Jill seems to be in a position to prescribe the drug with the best outcome only in a way that is not modally robust. Since Dr. Jill cannot reliably identify the best drug, even if Dr. Jill happened to prescribe it by something like guessing, Dr. Jill’s way of acting could be characterized as merely arbitrarily picking one of the two drugs. Such a way of acting only results in a success in a limited range of counterfactual situations – in half of them, to be exact, as there are two candidate drugs to choose from in the situation. By contrast, it is illustrative to notice that since she is aware of which drug is the one that is almost as good as the best one, Dr. Jill seems to be in a position to prescribe a drug with the highest expected value in a way that is modally robust, and this is what Jackson himself suggested as the recommended course of action. Maximizing expected value would not seem to involve on Dr. Jill’s part anything arbitrary, because Dr. Jill can reliably identify the drug that needs to be prescribed in order for the expected value to be maximized. Based on cases such as this one, it indeed is plausible that reliable reasoning about how one can φ often plays an important role in making modally robust φ:ing possible.

It is important to highlight here why mere true beliefs that do not reliably track the relevant facts are not sufficient for modally robust φ:ing. In practical reasoning, our relevant beliefs often function as kind of a mental map by which we navigate the world. If Dr. Jill had a true belief about which drug is the best one based on a mere guess, her mental map of the situation would be accurate, and it could guide her to prescribe the best drug. But in order for Dr. Jill to prescribe the best drug in a modally robust way, her mental map would need to adapt to the relevant counterfactual situations in such a way that it also correctly guides Dr. Jill to prescribe the best drug in these situations. And if Dr. Jill formed the true belief unreliably based on a mere guess, she would form a false belief about the best drug in many counter-factual situations, and proceed to prescribe the worst drug rather than the best one. Hence, in order for a person to φ in a modally robust way based on the mental map of the world consisting of her beliefs, the beliefs have to track the relevant facts in the relevant counterfactual scenarios in such a way that the person also φ:s in them.

Now, to return to the issue of adhering to the divine will, consider the case of a friend of Dr. Jill’s, Rev. Jillian:

Rev. Jillian

While Dr. Jill is concerned with prescribing a drug for her patient, Rev. Jillian is concerned with making a decision that is in accordance with the divine will. Despite this difference, they happen to be in somewhat similar situations. Just like in the case of Dr. Jill, Rev. Jillian has three courses of action to choose from, A, B, and C. Rev Jillian has reliably formed a true belief that the option A would be something that God is pretty happy with, and that of the options B and C, one is something God would be very happy with, and one is something that God would be very angry with. However, sharing the faith of Dr. Jill, Rev. Jillian has no idea which of the two options is which!

Now, just like Dr. Jill, Rev Jillian seems to be able to choose the pretty good option in a modally robust way by relying on her beliefs. However, is Rev. Jillian in a position to choose the best or the worst option in a modally robust way? It seems to be initially plausible that she is not, because her situations seems to be very analogical to that of Dr. Jill’s, and it is very plausible that Dr. Jill is not in a position to choose the best or the worst drug in a modally robust way. And from this example it might be tempting to generalize that modally robust adherence to the divine will always requires that one has reliably reasoned about the requirements of the divine will. Perhaps, then, we should conclude that the second premise of the skeptical argument is to be accepted, and that someone wishing to secure the possibility of non-accidental adherence to the divine will should rather aim to deny the first premise concerning the possibility of reliable reasoning about the requirements of the divine will.

Depending on the scope of the suggested skepticism concerning reliable reasoning about the divine will, one might see the seriousness of the situation differently. If one only thinks that reliable reasoning is often not possible about certain specifics, the case of Rev. Jillian would be quite illustrative of the situations practioners of religion often find themselves in; in these situations the divine will could not perhaps be adhered to perfectly in a modally robust way, but one could at least robustly choose an option that is quite good. But if one advocates a more broad skepticism, this would pose a problem for pretty much ever adhering to the divine will to any degree non-accidentally.

Having said that, in what follows, I set out to argue that there are various strategies religious thinkers can reasonably pursue in challenging the second premise.

Denying the second premise

P2: Without being able to reliably reason about what the divine will requires one to do, it is not possible to adhere to it in a non-accidental manner.

Modal robustness by reasoning reliably about something else

In this section I argue that even when we focus on the picture of persons acting based on mental maps consisting of their beliefs, it is possible for religious thinkers to reasonably challenge the second premise. My argumentation draws here on the fact that even if it is not possible to reliably reason about the requirements of a given norm A, it can still be possible to reliably reason about the requirements of a different norm B that has a robust relationship to the former norm. The relationship can sometimes be so robust that adhering to norm A always leads one to adhere to norm B across the relevant counterfactual situations. For instance, let us assume for a moment that classical utilitarianism, according to which morally right actions are those that maximize utility, is the correct moral theory. Now, since classical utilitarianism is the correct moral theory, if a person’s action maximizes utility, it is necessarily morally right. Yet, one does not need to believe that classical utilitarianism is the correct moral theory in order to reliably reason that a given action maximizes utility, and to act on this basis.

I now set out to investigate whether we could in a similar fashion reliably reason about the requirements of norms that are related to the divine will in a sufficiently robust way. I argue that there are various possible strategies to argue to this effect, which religious thinkers could reasonably pursue. I outline two such strategies here; the first relies on general revelation, and the other relies on special revelation. Should a strategy like this be successful, it would be a counterexample to the premise 2 of the skeptical argument, because it would show that modally robust adherence to the divine will is possible without the capability to reliably reason about what it requires one to do.

The first strategy arises from the observations that there exists a strong tradition of arguing that the divine will has such a close relationship with certain moral facts, like moral obligations, that the requirements of the divine will are coextensive with them. There are various different ways the nature of such relationship could be seen. One way of seeing it is that although morality is completely ontologically independent from God, God nevertheless necessarily wills us to act in a morally right way, because God is an infinitely good being. This seems to have been Kant’s (2001, 68) position when, in answering his own question “How, then, do we know the divine will?” he wrote that “God wills everything that is morally good and appropriate, and that is why His will is holy and most perfect. What is in fact morally good, is shown to us by ethics.” However, the proponents of the so-called divine command theories, which have gained a lot of new momentum during the last decades, have argued that the divine will has an even closer relationship to certain moral facts; namely, according to these theories, the divine will plays an essential role in somehow generating moral obligations. How exactly the divine will is supposed to generate moral obligations has been explained in several different ways. For a brief sample of the major positions, Adams (1987) sees “it being obligatory for one to φ” as meaning exactly the same thing as “God having commanded for one to φ” does,Footnote 16 Hare (2001) argues that one’s moral obligations to φ supervene on God having commanded for one to φ, and according to Quinn, one’s moral obligations to φ are caused by God’s commands for one to φ. (Quinn, 2006 contains several articles defending this position) However, all these positions agree that moral obligations are coextensive with the divine will.

The relationship of the divine will to certain moral facts is very important for our purposes here, because human beings seem to be naturally wired towards having certain moral beliefs, and various philosophers think that these beliefs usually reliably track what morality requires one to do. (see Campbell, 2019 for an overview of moral epistemology) Thus, many philosophers take the majority of people to be capable of reliable moral reasoning. Moreover, it is important to note that even those sympathetic to divine command theories (e.g. Evans, 2013, 21) often argue that even the majority of atheists tend to reason reliably about moral requirements, because an atheist does not have to conceive of morality as being grounded in the divine will, even if it actually is. I cannot here assess the relationship between morality and the divine will, or issues having to do with moral epistemology, but given that many philosophers take there to be a close relationship between them, religious thinkers should find it a relevant possibility, then, that acting in a moral way also always results in adhering to the divine will, and that people often reliably reason about what morality requires of them.

Some might here complain that the reason we cannot reliably reason about religious matters also makes it so that we cannot reliably reason about morality. In particular, some authors (e.g. Feldman, 2007) have argued that when it comes to skepticism arising from peer disagreements, religious and moral beliefs are in the same boat, because both domains contain seemingly insolvable and wide ranging disagreements among intellectually competent people. For this reason, one might argue that we can only reliable reason about moral matters, if we can also reliably reason about religious matters, and that this makes this strategy insignificant in the context of the skeptical argument formulated in this article. This is an interesting point, and I think that it is a defensible position to think that religious and moral beliefs live together or fall together to a great extent when it comes to peer disagreements, but it should be noted that I also outlined earlier other reasons one might have for skepticism in a religious context than religious disagreements, and these might have no effect on whether one can reason reliably about morality. Additionally, it should be remembered that it holds true especially with respect to morality that there are certain principles that are very widely agreed upon (e.g. those having to do with the wrongness of causing suffering for fun), and these principles might consequently be out of the scope of disagreement skepticism.

Moving on to the second strategy, it is also possible, quite trivially, to reliably reason about religious rules and guidelines acquired by the means of special revelation, even if reliable reasoning about the divine will was not possible. For just one possible example, it is possible to reliably reason about the Ten Commandments given in the Books of Moses, and, should an Abrahamic religion be true, the Ten Commandments would be the manifestation of the will of God, and thus robustly related to it. This strategy might strike one initially as somewhat suspicious, but it is important to notice that reliable reasoning about, say, the Ten Commandments, would be possible even if reliable reasoning about what the divine will requires one to do was not, because one does not have to conceive of the Ten Commandments as the manifestation of the divine will in order to form beliefs about what the Ten Commandments require one to do – after all, even an atheist can pick up the Bible and reliably reason about the rules and guidelines found in it.

To be sure, many who find the Ten Commandments important in deliberating what to do indeed do believe them to be the manifestation of the divine will, and base their actions on such beliefs, but there are exceptions, which makes this strategy relevant. For instance, consider the so called “cultural Christians”, who do not believe in the truth of Christianity, but nevertheless appreciate the general ethos of Christianity as a way of life, and might consequently base their actions on the rules and guidelines found in the Bible, like the Ten Commandments, to a great degree. Alternatively, perhaps even some devout believers might sometimes get so obsessed with e.g. the Ten Commandments that they are motivated to act in certain ways only because they believe that the Ten Commandments require it–not because they believe it is in accordance with the divine will. Here utmost dedication to the Ten Commandments themselves could, quite surprisingly, prove superior to dedication to the divine will itself in this respect–there is perhaps some deep wisdom to be found here, but I leave further assessment of this issue to another occasion.

We can conclude that even if it was impossible to reliably reason about the requirements of the divine will, it might nevertheless be possible to reliably reason about norms that are robustly related to these requirements, and that this would make it possible to adhere to the divine will in a modally robust way. Arguing along these lines is possible because of the fact that the divine will is often conceived of as being very closely related to various things, such as to moral obligations or certain Holy Texts, which we might well be able to reliably reason about. Therefore, we have found a family of ways in which the second premise of the skeptical argument can be reasonably challenged. Although the family surely includes various other reasonable strategies to the two outlined hereFootnote 17, I now proceed to assess another kind of ways in which the second premise can be reasonably challenged.

Modal robustness independently of reliable reasoning

In the last section, we established that reliable reasoning does indeed often play an important role in making modally robust φ:ing possible. However, it would be unjustified to conclude at this point that reliable reasoning plays a necessary role when it comes to modally robust adherence to the divine will, because often the modal status of one’s actions is in an important way independent from one’s reliable reasoning. I now set to argue that this observation provides further resources to challenge the second premise of the skeptical argument formulated earlier.

For starters, especially in the virtue-theoretic tradition many philosophers (e.g. Greco, 2010; Sosa, 2007) have justifiably claimed that we often manifest all kinds of dispositions that are not intimately connected to our propositional attitudes, but which might nevertheless manifest themselves as certain kind of outcomes in a modally robust way. For instance, when an experienced tennis player quickly reacts to a fast approaching ball, and manages to successfully hit it back to the other side of the net, the player does not necessarily form anything like a belief that she can successfully hit the ball back by acting in a certain way, and proceed to do so; rather, such actions are commonly based only on something like the player’s muscle memory that functions automatically outside of the player’s direct awareness. Moreover, an experienced tennis player might in so acting actually manifest a disposition that results in a success across a very wide range of relevant counterfactual situations.

One might argue that although this much is true, it is not very relevant when it comes to things that are more abstract than athletic performances such as playing tennis. For instance, in the example above, whether Dr. Jill is able to prescribe the best drug in a modally robust way indeed seems to be completely dependent on whether or not she can reliably identify which of the three drugs is the best one. Hence, perhaps it is the case that when it comes to more abstract norms such as “act in a morally right way” or “adhere to the divine will”, we need to reason reliably about the requirements of these norms–or about norms that are robustly related to them—in order to adhere to them in a modally robust way.

However, this is not necessarily the case. With respect to morally right actions, various authors have justifiably argued that something like a person’s well-toned sympathy can be capable of functioning as the basis of modally robust morally right actions, even if one does not reason reliably about morality, or about something robustly related to it.Footnote 18 For one thing, sometimes a person’s sympathy can plausibly function in a way not unlike the muscle memory of an experienced tennis player in the example above; one’s strong feelings of sympathy can by themselves make one act in a quite unreflective manner without anything like explicit practical reasoning taking place. Alternatively, a person’s sympathy could function in such a way that she across a range of relevant counterfactual situations adopts only such goals in her practical reasoning that lead to a morally right action in the particular situation she is in. For instance, one’s well-working sympathy could cause a person to adopt the goal of relieving the pain of another person only in situations in which this is the morally right thing to do, but not in situations in which it is not, such as in situations in which relieving the pain of one person would cause a lot of pain to various other persons in the near future. Here one’s practical reasoning would function in such a way that results in a morally right action in a modally robust way, but the modal robustness would be based on the fact that one reliably adopts goals that are in accordance with moral requirements, not on the fact that one reasons reliably about how to reach a certain goal robustly related to these requirements.

Hence, we should not underestimate in the present context the importance of things like muscle memory and sympathy; although they mostly function on a more unreflective level, they can nevertheless play an important role in enabling one to act non-accidentally. In the context of adhering to the divine will, could there be similarly, then, dispositions that do not draw their reliability from one’s reliable reasoning about certain requirements, but which nevertheless result in one adhering to the divine will in a modally robust way?

For one thing, should morality and the divine will be coextensive–which various authors have uggested to be the case, as we saw earlier—basing one’s actions on something like one’s well-working sympathy could also be a way of adhering to the divine will in addition to moral norms in a modally robust way. However, even if the divine will and morality came apart, or sympathy falled short of modal robustness as the basis of actions, it would still be possible to argue that there are modally robust ways of adhering to the divine will that do not rely on one’s reliable reasoning. To give an example that many religious persons should find convincing, the faculty of conscience, as it is often depicted, could plausibly play such roleFootnote 19. Conscience has, of course, been understood in various different ways throughout the history, but it has frequently been conceived of as giving rise to strong emotions in accordance with the divine will—the emotion most commonly associated with conscience being guilt. For instance, John Henry Newman (1979, 100) characterized the functioning of conscience in the following way: “Conscience has an intimate bearing on our affection and emotion, leading us to reverence and awe, hope, and fear, especially fear. Wrongdoing generates a lively sense of responsibility and guilt.” Moreover, in theological traditions, it has been common to think that this is a way God directs humans to act in accordance with His will, and even that the feelings of guilt associated with conscience are ultimately directed at GodFootnote 20. Now, should something like conscience indeed exist, and should it function in a reliable way—consistently giving rise to strong emotional reactions in accordance with the divine will–basing one’s actions on these emotions would plausibly make modally robust adherence to the divine will possible, just like basing one’s actions on well working sympathy would make it possible to do morally right things in a modally robust way.

We can conclude that there are also reasonable ways religious thinkers can challenge the second premise of the skeptical argument by focusing on dispositions that draw their reliability from other sources than one’s reliable reasoning about the divine will, or about something robustly related to it. I take this conclusion to be very important, because in our modern times it is very common to ignore or even look down upon our more unreflective capabilities in philosophical theorizing. Yet, especially because of the fact that practioners of religion are often associated with very lively emotional lives, paying attention to the role our more unreflective capabilities such as emotions play in our actions is a task of utmost importance.

Conclusions

I began the article by formulating a skeptical argument against the possibility of adhering to the divine will in a non-accidental way, which arises quite naturally from recent discussions on non-accidentality on the one hand, and religious epistemology on the other. According to the first premise, at least in many cases it is not possible to reliably reason about what the divine will requires one to do, and according to the second premise, this means that at least in many cases it is not possible to adhere to the divine will in a non-accidental way. In the second section I offered a defense of the first premise, and in the fourth section I offered a defense of the second premise. However, I spent the two following sections outlining various strategies religious thinkers can reasonably pursue in challenging the second premise of the skeptical argument, which is why we concluded that it is not necessary to challenge the first premise in order to secure the possibility of modally non-accidental adherence to the divine will.

Before concluding the article, I say a little bit about why the issue of non-accidental adherence to the divine will is something anyone should care about. There are plausibly various reasons for this, and I now mention a few of them.

For starters, non-accidentality might be deemed philosophically important by itself; given that a person φ:s, it is a very natural thing to do to ponder what role the person herself played in doing so. This also holds true in the context of adhering to the divine will.

A further reason philosophers have been interested in non-accidentality has to do with its normative relevance. Whether a person’s φ:ing is accidental or not is often deemed essential when assessing whether she is praiseworthy or blameworthy for doing so. Thus, when assessing whether a person is to be praised or blamed for adhering to the divine will, or for failing to do so, it is common to think that it is important whether one did so non-accidentallyFootnote 21. Very closely related to this point, non-accidental adherence to the divine will might also play an important role in one’s φ:ing having religious worth or theological worth. Following Kant, philosophers have most often been interested in morally worthy actions, but recently Whiting (2020) has theorized about epistemically worthy beliefs in the same fashionFootnote 22. While morally worthy actions are morally right actions that are done in a certain way, epistemically worthy beliefs are epistemically right beliefs that are formed in a certain way – non-accidentality of the success being often offered as the most important, or even sufficient, criteria. Following this trend, religious worth or theological worth can be associated with a person φ:ing in a way that is religiously or theologically right in a non-accidental wayFootnote 23, and I think that an account along these lines could be further developed.

Finally, because the notion of “accidentality” is closely related to the notion of “luck”, the conclusions are of course relevant for the philosophy of religious luck literature (e.g. Axtell, 2019; Zagzebski, 1994), which is concerned with analyzing the kinds of luck that might be involved in theological doctrines and religious phenomena, and assessing them on this basis. For instance, whether a person adhering to the divine will is a modal accident can plausibly be relevant for analyzing and assessing the doctrines concerning justification and the obtaining of salvation.