Pragmatic encroachment

Some recent epistemology has been concerned with the idea of pragmatic encroachment. In brief, pragmatic encroachment is the idea that practical considerations can alter the epistemic status of beliefs – they can make the belief more or less justified. Most famously, Jeremy Fantl and Matthew McGrath have argued that: ‘whether you can know something can vary with what's at stake for you, even if your evidence and general intellectual position with respect to what you know remains constant’ (Fantl & McGrath (Reference Fantl and Mcgrath2009b), 4). In other words, practical considerations such as what is at stake are epistemically significant: they can raise or lessen the extent to which your belief is justified or the amount of knowledge you have, even if the alethic considerations remain the same.

Fantl and McGrath argue that knowledge and justification rely on our rationality to act accordingly, expressed in the following principles:

Notably it is JA rather than KA that is most relevant for this article, since what is at stake is the rationality of holding different religious beliefs rather than whether they are actually true or not. Most of the pragmatic encroachment literature focuses more on knowledge than on justification, but if knowledge (broadly speaking) entails justified belief, then if KA holds then JA holds too. Therefore I will follow or adapt the pragmatic encroachment arguments and examples, applying them to justification for belief rather than knowledge per se.

Why should we accept KA and JA? One reason, Fantl and McGrath argue, is that we tend to cite our knowledge or our justified belief as a justification for acting in a certain way. To give one example:

It would be irrelevant to assert my justified belief in having Brussels sprouts already, if this did not justify going straight home, rather than stopping at the supermarket.

We also often blame others for what they have done on the basis that they did not have enough justified belief to justify acting in a certain way, or else they had epistemic justification for a belief that should have prevented them from acting in a certain way. Imagine that I told my mother than I was cancelling my travel medical insurance to save money. She would berate me on the grounds that I do not know (and do not have justification for believing) that I will not get ill or have an accident (adapted from Hawthorne & Stanley (Reference Hawthorne and Stanley2008), 572). Imagine that a student became overly anxious that I had not replied to a request for an essay extension sent by e-mail on a Sunday evening after they failed to receive a reply after half an hour, and sent panicked emails to all the other lecturers in the Department. I would probably reply: ‘but you should know by now (or have sufficient justification for believing) that I never reply to emails on Sunday evenings, and always reply to them on Monday mornings. That is what I have always done, and that is what my email autoreply states’ (see Fantl & McGrath (Reference Fantl and Mcgrath2007), 562). An attraction of pragmatic encroachment, as these examples show, is that it returns epistemology to its practical roots: we often want to know whether a belief is true or justified because we want to know what we or someone else should do in a given situation, or to work out whether someone was culpable for doing otherwise.

Significantly for pragmatic encroachment, it seems to be knowledge or justification rather than insufficiently justified belief (on the one hand) or else the truth of the matter (on the other) that are considered morally significant. A doctor would be considered negligent for using a needle she did not know or was not sufficiently justified in believing to be clean, and a captain would be considered negligent for setting sail in a ship he did not know or was not sufficiently justified in believing to be seaworthy (see Rizzieri (Reference Rizzieri2011), 219; see Clifford (Reference Clifford1877)). Neither the insufficiently justified belief that the needle was clean or the ship safe, nor the fact that the needle was clean or the ship safe, would establish that the doctor and captain were not negligent – what is relevant is knowledge-level justification. These examples also highlight that second-order knowledge is also relevant: if the doctor does not know that she does not know that the needle is clean, or the captain does not know that he doesn't know that the ship is seaworthy, we are less likely to consider them culpable (Hawthorne & Stanley (Reference Hawthorne and Stanley2008); see Williamson (Reference Williamson2005)).

These examples only seem to work one way with respect to KA and JA (justifying actions on the basis of justified-belief) but not the other way (justifying beliefs on the basis of justified-action). However, an additional consideration for pragmatic encroachment gives us reason to think that it works the other way too. Comparing high stakes with low stakes cases reveals that we tend to think that people have knowledge or justification for belief in some situations but not others, even if there are the same alethic epistemic credentials in both situations. Consider the following questions:

If your answer is different in the high stakes cases from your answer in the low stakes ones, then you have entered into the conundrum that pragmatic encroachment seeks to address.Footnote ¹ In the low stakes cases, most people are intuitively more likely to regard Mary and Berlioz as having knowledge or justified belief that the bank will be open or that the heating is switched on. This supports KA and JA, pragmatic encroachment epistemologists argue, because we are more likely to think that Mary and Berlioz are rational to act (waiting till Saturday to pay in money; making a bet with a stranger) – and it is precisely the rationality of the action that determines whether the belief is justified. For the low stakes cases, a lower level of epistemic warrant is required to constitute knowledge or justified belief. KA and JA account for why the epistemic standards differ in different cases, in a way that purism or epistemological orthodoxy, which holds that epistemic considerations are purely alethic, cannot do.

Notably, a further implication of Fantl and McGrath's argument is that two people with identical alethic epistemic credentials might in fact end up in different situations regarding the epistemic status of their beliefs, because one might have more at stake than the other. This can be illustrated by the fact that we might give Mary and Berlioz in the two cases above different names for either the high or low stakes cases, without the point changing significantly (see also Rizzieri (Reference Rizzieri2011)). This will become important when we discuss the implications of this argument for the way we do philosophy of religion.

There are lots of possible objections to the high/low stakes argument for pragmatic encroachment, which have been explored and responded to elsewhere. I will briefly note two objections, and responses to them, because these seem to me the most obvious and pressing. First, in the cases above, both high stakes and low stakes cases involve justified belief or knowledge. Second, in the cases above, neither high stakes nor low stakes cases involve justified belief or knowledge.

In response to the first objection, the claim is that not only the low stakes but also the high stakes cases involve knowledge or justified belief: it is enough for Mary to know that when she transferred money two weeks previously, it arrived the following day, to justify her saying she knows she could wait till Saturday (even though her mum will have nothing to eat if it does not come through). The significance of this view for pragmatic encroachment is that we would probably not agree that it is rational for Mary actually to wait till Saturday. If this is correct, then knowledge or justification for belief is separated from rationality to act (and KA and JA invalidated). Why should we not allow that this is the case?

One response to this view is that it reduces Mary to saying something like: ‘I know that the money will come through in time, but nevertheless I would be negligent or careless to act accordingly’. As Aaron Rizzieri points out, this kind of statement is infelicitous for the same reason that the Moorean sentence ‘I know, but I might be wrong’ is infelicitous: in both cases, despite the statement of knowledge, there is an admission of the presence of a non-negligible possibility of error – and the non-negligible possibility of error precludes knowledge (Rizzieri (Reference Rizzieri2011), 222; see also Rizzieri (Reference Rizzieri2013); Fantl & McGrath (Reference Fantl and Mcgrath2009a), (Reference Fantl and Mcgrath2009b); Dougherty & Rysiew (Reference Dougherty and Rysiew2011)). This is highlighted by the fact that if someone were to say ‘I know, but I am not sure, but I indeed know’ then we would take her as wavering, rather than expressing a constant and consistent mental state (Fantl & McGrath (Reference Fantl and Mcgrath2007), 573). Therefore, we should reject the idea that the high stakes as well as the low stakes cases involve justified belief or knowledge.

How about the possibility that Mary and Berlioz should not be said to have knowledge even in the low stakes cases? The problem with this approach is that it sets the bar far higher with respect to what constitutes ‘knowledge’ and ‘justification for belief’ than is suggested by our intuitions and ordinary language use. If we went down this path, we would probably have to reject fallibilism (the idea that something can still count as knowledge even if there is a non-zero chance of it being wrong) in favour of a counter-intuitive and paralysing inductive scepticism (Rizzieri (Reference Rizzieri2011)).

Why not believe in an evil God?

Returning to the evil-God challenge, what reasons might the idea of pragmatic encroachment give us for preferring one over the other? Relatedly, about what kinds of things might the high stakes pragmatic encroachment speaks of be? Were we to live our lives as though God were evil rather than as though God were loving, what practical differences would this make? I suggest that there are some reasons to suppose that it would have negative consequences for at least some people; in particular, as I will argue, that it is likely to be detrimental to a person's well-being.

One of the contexts in which the relationship between religious belief and well-being is most discussed is health, since health professionals are interested in the ways in which religious beliefs can cause or provide buffers against, exacerbate, or enhance recovery from, various forms of illness (Koenig, King, and Carson (Reference Koenig, King and Carson2012)). Evil-God hypothesis is not a common religious belief in real life, and the health literature tends to focus on Christian and Jewish beliefs in the US and UK, which involve belief in a good God; therefore evil-God hypothesis has not received much attention. However, belief in demons and evil spirits is common in a wide range of religious traditions, including various forms of Christianity and Judaism, and has been discussed in health literature. These beliefs tend to be evaluated negatively in terms of their relationship with health and also other forms of well-being.

For example, responding to the problem that religious activities such as prayer are sometimes treated as homogeneous in health literature, Kenneth Pargament has developed a research questionnaire in order to provide more nuanced information about the ways in which different kinds of religious attitudes and activities relate to the way in which people cope with stressful situations, and whether these reflect better or worse health outcomes. Pargament distinguishes between positive and negative religious coping, where positive coping includes the perception of a good relationship with God, belief that greater meaning is to be found, and a sense of spiritual connectedness. Negative coping includes the perception of a less secure relationship with God, an ominous view of the world, the belief that the event may be a punishment by God, and the idea that the devil or demons might be involved (Pargament, Koenig, and Perez (Reference Pargament, Koenig and Perez2000), 539). Clinical studies suggest that when dealing with stressful situations such as bereavement or divorce, the idea of demonic involvement (along with other negative forms of religious coping) causes distress and has a negative impact on mental and other forms of health (ibid., 520).Footnote ²

Kate Loewenthal argues that, while religious beliefs generally are not pathogenic, belief in possession by evil spirits can exacerbate and even part-cause psychiatric illnesses such as schizophrenia and culturally bound syndromes. She stresses that belief in evil spirits is ‘clearly terrifying for those who have been affected by a horrible affliction and who believe it is the result of a curse or evil spirits’ (Loewenthal (Reference Loewenthal2007), 24). This is borne out by people's personal accounts. One woman diagnosed with bipolar disorder relates the negative effect that being told by a counsellor that she was possessed by demons had on her mental health. She reports ‘never being able to get over that experience’ and that she later had psychotic symptoms, during which the counsellor's words came back to her and she saw ‘demons all around’ (Wonder (2006)). Belief in powerful evil spirits, then, seems to be detrimental to well-being.

This, I suggest, may be relevant for the prospect of believing evil-God hypothesis, since belief in demons shares significant features with belief in evil-God. According to both, there is a malign and powerful entity who is in control of facets of one's existence and experience, and this gives rise to increased anxiety and distress. In fact, belief in evil-God may in fact be more severe in this respect than belief in demons. Since ‘God’ (in this context) entails that the being is both creator and omnipotent, rather than merely quite powerful, belief in evil-God may colour one's perceptions of the nature of the world as a whole, and also of one's final destiny. While people who believe in demons but a good God can hope ultimately for the triumph of good over evil and thus liberation from the demonic, people who believe in an evil omnipotent God would have no such hope.

In contrast, a very large body of evidence suggests that belief in a good God seems to be conducive to well-being (see Koenig, King, and Carson (Reference Koenig, King and Carson2012)). The results of studies using Pargament's questionnaire indicate that this is particularly so when the person adopts a ‘benevolent religious reappraisal’: that is, when she redefines the stressor through religion as benevolent and potentially beneficial. This could include seeing the situation as part of God's plan, trying to find a lesson from God in the event, or thinking the event might bring her closer to God (Pargament, Koenig, & Perez (Reference Pargament, Koenig and Perez2000)). People who adopt these positive religious coping methods show better adjustment in relation to a range of facets of existence including stress-related growth, physical health, mental health, emotional distress, and religious outcome. Belief in other benevolent and involved entities also seems to have a positive relationship with well-being. For example, Lana Jackson, Mark Hayward, and Anne Cooke undertook a study of people who hear voices but, far from finding it distressing, experience it positively, as a source of strength and comfort. Significantly, eleven of the twelve participants have spiritual beliefs about the voices, identifying variously as Shamanic, Spiritualist, and born-again Christian. A number of voices heard by these people are perceived as guides (whether ancestors or dead relatives, the voice of nature, or of God) (Jackson, Hayward, & Cooke (Reference Jackson, Hayward and Cooke2010)). That ‘hallucinatory’ voices can be positively rather than negatively experienced seems to be something that emerges through the interpretation of them, rather than an essential and immutable characteristic that distinguishes positive hearing from an inherently negative, pathological kind (see Heriot-Maitland, Knight, and Peters (Reference Heriot-Maitland, Knight and Peters2012); Fulford & Jackson (Reference Fulford and Jackson1997); Scrutton (forthcoming)). In other words, positive religious beliefs have a causal and not merely correlative relationship to positive experiences of voice hearing. Therefore, belief in benevolent powerful entities, including God, seems (ceteris paribus) to be beneficial for well-being.

So far, I have suggested that we might have a practical reason to prefer good-God to evil-God hypothesis: the literature discussed is suggestive of the idea that good-God hypothesis is likely to be more conducive to human well-being than evil-God hypothesis. If JA is correct, this also gives us an epistemic reason to prefer classical theism to evil-God hypothesis. It is more reasonable to live as though classical theism is true because classical theism is just as likely to be true as evil-God hypothesis but is more conducive to human well-being. Therefore, we are more justified in believing classical theism than believing evil-God hypothesis. Or, to put it the other way round, we are not rational to act as if evil-God hypothesis is true given that evil-God hypothesis is likely to result in a diminishment of our well-being, and given that good-God hypothesis is equally likely to be true and more likely to be good for us. We are therefore not rational to believe in an evil God rather than a good one.

Why should we allow that well-being, in particular, is one of the practical considerations that can justify our beliefs? That we do take well-being considerations as a reason for action and as a justification for belief is highlighted by an example given in the context of an objection to my argument that I think in fact supports my case. In the course of corresponding about this article, Roger White objected to the idea that well-being might be an epistemically significant factor, but suggested that another practical factor – self-fulfilling beliefs (or at least beliefs the holding of which increases the probability that they are true) – may have epistemic weight (Roger White, email communications 13 and 25 February 2015). White argues that there is a contrast between beliefs that are believed because they are self-fulfilling and beliefs that are believed because they are conducive to well-being, the former being epistemically viable and the latter epistemically dubious. Self-fulfilling considerations are significantly disanalogous to well-being considerations, since self-fulfilling beliefs change the truth of the proposition in question, whereas belief in a good God (for example) does nothing to bring about the existence of a good God. Furthermore, White argues, well-being considerations could be used to justify wishful thinking. Self-fulfilling beliefs are less open to the charge of justifying wishful thinking, because they are self-fulfilling (or probability-enhancing) and so can actually make the beliefs true (or more likely).

That we might sometimes choose to believe certain things in part because they are self-fulfilling seems correct to me, and can be illustrated with the following examples:

I agree with White that in these cases we are likely to think that belief can be justified on the grounds of being self-fulfilling. However, it seems to me that far from forming a contrast to well-being cases, the point about self-fulfilling beliefs in fact supports well-being as a factor justifying beliefs. This is because the rationality to hold the self-fulfilling belief is predicated on it being a belief that is good for one's well-being. If Jemma's doctor is right, then Jemma believing that she will recover is a self-fulfilling belief, but believing that she will not recover is also a self-fulfilling belief. It is primarily the fact that recovering is conducive to well-being that makes the belief that she will recover, rather than the belief that she will not recover, rational. Likewise, Fred could decide to believe that everyone at the party will hate him, and this would seem to be just as self-fulfilling as the belief that everyone will like him. It is the fact that he wants people to like him, that liking him is conducive to his well-being, that makes it the more rational. Boudicca could sabotage her army's chances of success by persuading herself and the army that they will lose, and this would constitute a self-fulfilling belief just as much as the reverse, but the fact that winning is conducive to her well-being means that we are more likely to think her rational for holding the self-fulfilling belief that she will win rather than the belief that she will not. Self-fulfilling beliefs are, therefore, dependent on considerations such as well-being, as the example of self-fulfilling beliefs highlights, and the proposed objection in fact supports my argument. It therefore seems to me reasonable to regard well-being as one of the practical considerations that might influence our justification to do, and thus to believe, particular things.

This leaves White's objection that allowing well-being as a justification for belief might legitimate wishful thinking. I think that including well-being as a justification for belief might promote wishful thinking if we suggested that well-being ought to take priority over alethic considerations or exclude them, but that is not the case here. According to the evil-God challenge, evil-God and good-God hypothesis seem to have equal alethic parity: what I am arguing is that well-being considerations tip the balance in favour of one option rather than the other.

This, of course, does not preclude options other than belief in evil-God or good-God. We might decide that belief in no God is more likely than either of these options – Law's ultimate objective. Or we might consider arguments for both good-God and evil-God extremely strong, and believe both by adopting a dialetheic framework. All I have done is to answer Law's challenge about why we should prefer good-God to evil-God hypothesis: we should prefer it because there is some evidence to suggest that doing so is better for our well-being. Since being rational to act as though p constitutes rationality to believe p, if we have to choose between these two options, all other things being equal we are (contra Law) more rational to prefer good-God to evil-God hypothesis.

In this argument, I have focused on well-being as a reason for preferring one doxastic option over another. I think there might be other reasons too. One is the moral consequences of our beliefs. For example, there might be equal evidence to suggest that all human beings are of equal value, and that some (say, clever, talented, or beautiful people) are more valuable than others. Severely mentally disabled people may seem less valuable on account of having fewer of the functions we often associate with human value: for example, reason, ability to communicate, relations with other humans on an equal footing. They may not be, at least by conventional standards, clever or talented or beautiful. Yet people such as parents and siblings, who have a close relationship with people with severe mental disabilities, often testify to their equal value, and to the way in which their presence contributes something of value and perhaps unique value to the lives of people around them. Hearing such people speak from experience seems to provide some evidence that severely mentally disabled people are equally valuable after all. Therefore the evidence seems to be roughly equal on both sides. But believing that people with severe mental disabilities are of lesser value can lead to callous and insensitive behaviour towards those people and to those around them. Believing that people with mental disabilities are of equal value can lead to kindness and patience, and perhaps also greater humility about our own humanity and (almost certainly temporary) capacities and characteristics. In the absence of definitive evidence for or against the equal value of mentally disabled people, then, it seems there are practical reasons in the shape of the moral consequences of our beliefs for preferring one option over the other. Moral consequences seem to me to be a separate kind of justification to well-being, though if we are virtue ethicists we might think there is significant overlap.

What are the implications of this for philosophy of religion?

So far, I have offered pragmatic encroachment as possible grounds for surmounting the evil-God challenge, suggesting well-being and moral consequences as two practical reasons for choosing between beliefs. What are the implications of pragmatic encroachment for philosophy of religion more broadly? Philosophy of religion has traditionally been concerned with arguments for the truth or otherwise of religious beliefs, such as arguments for the existence of God, or else the rationality of religious belief, based on criteria such as religious experience (Hick (Reference Hick, Geivett and Sweetman1992); Alston (Reference Alston1993)), or because they constitute properly basic beliefs (Plantinga (Reference Plantinga1967)). Practical elements of and reasons for religious beliefs have often been seen as more properly the concern of practical or pastoral theology, counselling, or therapy that involves people with religious beliefs (as with the health literature), or perhaps sociological and anthropological approaches to religious studies. But if the practical elements of religious beliefs, such as the ways in which they enhance or detract from well-being and moral behaviour, are part of what constitutes justification for beliefs, then we have philosophical reasons to regard these as a central part of philosophy of religion alongside alethic considerations.

Furthermore, as noted in relation to high and low stakes cases, the idea of pragmatic encroachment suggests that some beliefs might be more rational for some people to hold than others, because of the different ways in which they interact with practical considerations. This gives us reason to pay attention to the roles religious beliefs play in individual people's lives. What is helpful to one person may not be helpful to another, and so what is rational to believe for one person may not be rational for another. This is the case with respect to individual people within a single culture but, even more, with respect to different cultures, where people can relate very differently to different beliefs.

The following example may serve to illustrate this point. The Ju/'hoansi, a nomadic San or Bushman people in the Kalahari, practised one of the oldest African religions up until the 1970s. Ju/'hoansi believed in two deities: ≠Gao N!a, the creator god, and //Gauwa, the lower god. Neither ≠Gao N!a nor //Gauwa seem to have been especially nice. Stories were told of ≠Gao N!a's incest and cannibalism. The creation of the current world was ≠Gao N!a's second attempt, and the ‘proper order’ was achieved only as an unintentional side effect of //Gauwa's pranks (Platvoet (Reference Platvoet, Harvey and MacLeod2001), 126). Humanity was split into superior and inferior groups, with the San, including the Ju/'hoansi, at the bottom. At death, ≠Gao N!a transformed humans into //gauwasi, spiteful dead people ≠Gao N!a used to spread dissent, disease, and death among humans by shooting them with tiny arrows – a pastime the //gauwasi shared with //Gauwa (ibid., 127). At the same time, ≠Gao N!a gave to some people n/um, or healing power, and appeared to the owners of n/um in dreams to teach them the songs that should be sung during the ‘curing dance’ in order to heal. While n/um itself is a gift, it was also regarded with a degree of ambivalence, since it was also only won through great pain and an experience of death, the realm of //Gauwa and the deceased. #Gao N!a was not prayed or sacrificed to by the Ju/'hoansi, and in fact they had only one ritual – the curing dance – which, far from being an act of worship of #Gao N!a, was instead perceived as a war against him (ibid.). The curing dance included not only healing and protection against #Gao N!a and //Gauwa, the sources of evil, but also chasing the gods away with firebrands and yelling abuse at them. It could also involve pleading with //Gauwa to cure the sick by shooting them with evil arrows, though in the knowledge that //Gauwa might capriciously decide to kill them instead (ibid., 129).

Ju'hoansi religious beliefs have some commonalities with beliefs in evil spirits and evil-God hypothesis, and so we might suppose that, like belief in evil spirits, Ju'hoansi religion induced anxiety and distress among its adherents. Yet anthropologists’ reports suggest that this was not the case. They report that stories of #Gao N!a's incest and cannibalism were met with raucous laughter (ibid., 126). Night-long curing dances were participated in, not only for therapeutic purposes, but also because they involved singing and dancing, and were regarded as fun (ibid., 125, 128). In general, while the studies did not focus specifically on individual health as such, it is reported that the Ju'hoansi had a high quality of life.

Nor does the belief in non-benevolent deities seem to have made the Ju'hoansi especially immoral. In fact, while violence sometimes broke out within Ju'hoansi communities, they did not go to war (ibid., 126). They are also believed to be among the most egalitarian societies that ever existed in terms of equality between women and men and between older and younger generations (ibid.). They tended to share possessions, and decisions were reached by mutual consensus (ibid., 124–125) – characteristics and practices that are often held up as morally praiseworthy.

How are we to reconcile this with Pargament and Loewenthal's conclusions that beliefs in the involvement of evil spirits are detrimental to health? One limitation of the religion and health literature, well recognized by the researchers themselves, is that it is focused overwhelmingly on religions in the United States and western Europe which are either Christian or else have been strongly influenced by Christianity. So, for example, in their recent Handbook of Religion and Health, Harold Koenig, Dana King, and Verna B. Carson write that:

Hisham Abu-Raiya and Kenneth Pargament agree, reporting that one of the main limitations of the current literature on religion and well-being is that ‘the studies that have been conducted and the measures that have been developed have focused almost exclusively on Christian samples, and have been geared largely to members of Judeo-Christian traditions’ and that ‘Further empirical studies are needed to reveal a clearer picture of the relationship between religion and health among different religious groups, especially non-Western religious traditions’ (Abu-Raiya & Pargament (Reference Abu-Raiya, Pargament, Cobb, Puchalski and Rumbold2012), 337; see also Dein (Reference Dein2006); Loewenthal (Reference Loewenthal2007), 60; Pargament, Feuille, & Burdzy (Reference Pargament, Feuille and Burdzy2011), 68, 72).

The example of the Ju'hoansi raises an interesting point: while we have some evidence to suggest that beliefs in evil supernatural entities are often unhelpful and detrimental to health in western cultures, it may be quite a different case in another culture. The way in which the belief interacts with the culture's narratives, community identity, and ritual may give it a different relationship with well-being. We do not have sufficient evidence to say that this certainly was the case with the Ju'hoansi – but there are sufficient indications to mean that in this kind of case it is worth trying to find out. This, then, is a further implication of the high and low stakes cases for philosophy of religion: not only are practical as well as alethic considerations at the heart of religious epistemology; we also need to look closely at specific concrete cases in our own and other cultures, since these cannot be generalized or deduced a priori.