The evidential challenge for petitionary prayer

Abstract

In the past decade, philosophers have devoted a great deal of attention to the practice of petitionary prayer. Philosophical inquiries have posed a priori problems – issues that arise from an analytical investigation of the concept of God, the concept of petitionary prayer, and the relationship between the two. Taking a different direction, this article shifts the focus from possibility to actuality. Accordingly, this article does not deal with the question ‘Can God answer petitionary prayers?’ but rather with the question ‘Does God answer petitionary prayers?’ and, mainly, its implications. More accurately, I will present the tension between the religious belief that petitionary prayers can be effective and the fact that this does not seem to be so in reality, a claim that has been the conclusion of several empirical studies. Then I will present and examine several solutions to this tension. Although I will try to promote my preferred solution, my main aim is to clarify the problem and discuss the advantages and disadvantages of the solutions offered to solve the problem under discussion.

Introduction

In the past decade, scholarship has devoted a great deal of attention to the practice of petitionary prayer. Philosophical inquiries have posed questions such as ‘Given the nature of God, does petitionary prayer ever make a difference to whether something happens?’¹ and ‘Can God answer prayers concerning the past?’² Notably, these investigations deal with a priori challenges – issues that arise from an analytical investigation of the concept of God, the concept of petitionary prayer, and the relationship between the two.

Taking a different direction, this article shifts the focus from possibility to actuality. Accordingly, this article does not deal with the question ‘Can God answer petitionary prayers?’ but rather with the question ‘Does God answer petitionary prayers?’ and, mainly, its implications. This seems to be an empirical question and, according to the empirical data that we have, prayers do not seem to be efficacious. This is the evidential challenge for petitionary prayer in a very quick sketch.

Yet, this challenge is ambiguous and there are many ways to understand it. Each way will pose a different challenge. In what follows, I will try to clarify the various issues that arise. In its different formulations, this challenge is broadly relevant to theists as there is a deeply held idea in the Abrahamic religions (and in many other religions as well) that God responds to prayer, at least sometimes. In the philosophical literature, there are a handful of brief references to the challenge (Lewis 1958; Miles 1959, 180–188; Brümmer 1984, 1–10; Swinburne 2006; Dawkins 2006, 61–66; Dennett 2006, 274–277; Davison 2017a, 78–80; 2022, 25–26). However, there is no detailed presentation of the challenge or discussion of the different solutions. Thus, I will start by sharpening the question, after which I will present a variety of possible answers. Each of the proposed solutions will be analysed in turn. Ultimately, I aim to motivate scholarly discussion of this intriguing topic.

My aim in this article is not to reject the solutions suggested. And although I will promote my preferred solution, my main aim is to clarify the problem and discuss the advantages and disadvantages of the solutions offered to cope with the challenge under discussion. I hope this article will motivate other thinkers to continue investigating the topic.

What is an answer to petitionary prayer?

This article asks the question, ‘Does God answer petitionary prayers?’, or more accurately ‘Given our empirical data – what should we believe regarding the efficacy of petitionary prayer, and how does this affect the rationality of engaging in such a prayer?’. However, in order to explore whether there are petitionary prayers that have, in fact, been answered by God or what we should believe regarding the subject – we will first need to characterize the phenomenon of ‘answered prayer’.

A simple characterization such as ‘S's petitionary prayer for a result E is answered by God if and only if S prayed for E to happen and then E actually happened’ will not suffice. If I pray that the sun will shine tomorrow, the fact of the sun shining tomorrow does not prove, or even indicate, that God has answered my prayer. That is because the sun might have shone whether or not I prayed for it. It seems probable, or at least possible, that my prayer does not lead God to make the sunshine. The fact that the sun shines is better explained by appealing to natural principles and causes than by appealing to my prayer and God taking it into account. Moreover, another, and more acute, problem with such a proposal is that there are possible cases (whether probable or not) in which the right side of the account is satisfied and the left side is not. Therefore, it cannot serve as an account at all. A more sophisticated characterization of ‘answered petitionary prayer’ is thus required, one that includes a counterfactual condition.

Scott A. Davison offered precisely such a characterization:

S's petitionary prayer for an object E is answered by God if and only if God's desire to bring about E just because S requested it plays an essential role in a true contrastive explanation of God's bringing about E rather than not. (Davison 2017a, 38)

As one can see, this characterization includes the counterfactual element but also much more. For the purpose of this article, the additional elements will not play a significant role. Therefore, I will not delve deeper into the considerations that led Davison to formulate and embrace this specific characterization. In general, however, I will say that Davison suggests that to say God answered prayer is to say that the full causal explanation of the occurrence of the state of affairs requested by the petitioner will include, among other things, God responding to the petition itself.

The evidential challenge for petitionary prayer

In 1872, Francis Galton (1822–1911), the Victorian polymath, published an essay titled ‘Statistical Inquiries into the Efficacy of Prayer’. In this work, Galton tried to prove that petitionary prayers do not work. He noted that the life expectancy of the British royal family and clergy was no longer (and, in fact, was even shorter) than that of any other group in the English society of his day – despite the fact that those two groups were the focus of more prayers than any other group in English society. On the basis of this reasoning, Galton concluded that prayers are not efficacious.³

Galton was not alone in using scientific tools to examine the efficacy of prayer. Since the Victorian era, others have followed in his footsteps. Studies on this topic have mostly found prayer inefficacious (Astin et al. 2000; Powell et al. 2003; Benson et al. 2006; Masters et al. 2006; Hodge, 2007). Research reports indicating the efficacy of prayer tend to be marked by significant methodological problems (Powell et al. 2003; Dennett 2006, 274–277; Masters et al. 2006; Sloan 2006, 157–179; Schlitz et al. 2012).

Moreover, some may think that we need not take scientific recourse as our only evidence because we can simply see with our own eyes that prayers do not seem to work, at least not often (Overall 2009). As Galton himself wrote:

If prayerful habits had influence on temporal success, it is very probable, as we must again repeat, that insurance offices, of at least some descriptions, would long ago have discovered and made allowance for it. It would be most unwise, from a business point of view, to allow the devout, supposing their greater longevity even probable, to obtain annuities at the same low rates as the profane. Before insurance offices accept a life, they make confidential inquiries into the antecedents of the applicant. But such a question has never been heard of as, ‘Does he habitually use family prayers and private devotions?’ Insurance offices, so wakeful to sanatory influences, absolutely ignore prayer as one of them. The same is true for insurances of all descriptions, as those connected with fire, ships, lightning, hail, accidental death and cattle sickness. (Galton 1872, 134)

However, in contrast, there are many reports of religious people who claim that their prayers, or the prayers of others, were answered. So the whole issue of when, and under which circumstances, we can justify relying on ‘our own eyes’ is complicated and I will not address it in this article. For me, the empirical studies will suffice and our (or my own) everyday experience is only an addition.

As I noted earlier, in order to fully and precisely state and understand the challenge, we need to provide some clarifications. First, it is important to distinguish between the challenge to the rationality of religious belief regarding the efficacy of petitionary prayer, and the challenge to the rationality of the religious practice of petitionary prayer. These challenges are strongly connected but not identical. Most of my article will deal with the rationality of religious belief, but it will also relate to the question of practical rationality. For the sake of clarity, in each section I will note the kind of rationality I am considering.

Let us begin with a discussion of religious belief. The formulation of the question being asked here, ‘Does God answer petitionary prayers?’, is ambiguous. Are we asking whether God ‘always’ or ‘typically’ or ‘sometimes’ or ‘ever’ answers prayers? The phrasing will make a difference in how one would go about trying to answer the question.⁴ The answer to the question ‘Does God ever answer petitionary prayers?’ will be affirmative even if there was or will be only one prayer in all human history (including the future) that has been answered. The burden of proof to justify a negative answer to such a question would thus be extremely heavy, so heavy that it could never be satisfied.

Therefore, in this article, I will not investigate whether God has ever answered, or will answer, a petitionary prayer. Rather, I am to discuss the epistemic rationality regarding the belief in the efficacy of petitionary prayer and the practical rationality of engaging with such practice in order to bring about a particular future among various possibilities.

Regarding epistemic rationality, I will ask what, given our empirical data, we would be rational to believe regarding the efficacy of prayer. In this regard, it seems, prima facie, that the lack of empirical data that support the belief in the efficacy of prayer, and the positive empirical data that indicate that prayers do not work, need to make us not believe in the efficacy of prayer and even to believe in its ineffectiveness. In what follows, I will discuss several approaches that try to undermine the credibility of such an empirical approach in regard to the efficacy of prayers.

Regarding practical rationality, it is important to say that one may have many motivations to practise petitionary prayers, including, for example, the desire to please a religious parent, the psychological effect of such prayers on the petitioner or even compliance with God's command to do so. The evidential challenge does not pose any problem for those kinds of motivations and justifications. Therefore, I will put them aside (for a moment) and focus on the motivation to practice petitionary prayer that stems from the goal of bringing about a particular future among various possibilities.

The question regarding the rationality of the religious practice of petitionary prayer in order to bring about a particular future among various possibilities is strongly connected to the question of the rationality of belief in the effectiveness of such prayers. Prima facie, in order to be justified in offering petitionary prayer for the purpose of bringing about a particular future among various possibilities, one should believe that petitionary prayers are effective. However, one does not need to believe that God always, or even often, answers petitionary prayers in order to justify offering such a prayer. Even if one is almost certain that petitionary prayers do not work at all, he may be in a situation in which it may be rational to engage in such a practice in order to bring about a particular future among various possibilities. After all, if one is drowning and has nothing else to do, maybe asking God for help is not a bad idea.⁵ So, we can see that the connections between the rationality of the belief in the efficacy of prayers and the rationality of practising such a prayer (for the purpose of bringing about a particular future among various possibilities) are complicated.

Therefore, in order to make things less complicated and more acceptable, I propose to state the evidential challenge to the practice as a comparison between homeopathy and petitionary prayer. In both cases, we lack empirical evidence that these practices are effective and we even have some empirical evidence that such practices are not effective.⁶ Additionally, as far as we know, there are no risks in engaging in these practices.

However, it is important to be careful with comparisons. As we well know, everything can be compared to everything, and nothing is identical to anything besides itself. Therefore, we need to be careful with the comparison between petitionary prayer and homeopathy. There are many clear differences. Furthermore, many religious people will claim, and rightly so, that they do not relate to petitionary prayer in the same way they relate to medicines. Indeed, one of the aims of this article is to identify the differences between the uses of the two.

Following these qualifications, the only comparison that I want to focus on is the comparison between the rationality of participating in a petition-prayer for the purpose of bringing about a particular future among various possibilities and taking homeopathic treatment for the purpose of bringing about a particular future among various possibilities.⁷ According to this framing, the evidential challenge to the practical rationality of engaging in petitionary prayer, in order to bring about a particular future among various possibilities can be stated thus:

Based on the empirical evidence we have, belief in the efficacy of petitionary prayer can be compared to belief in the efficacy of homeopathy. Accordingly, the justification and rationale for practising such prayer, for the purpose of bringing about a particular future among various possibilities, should be the same as the justification and rationale for the use of homeopathic drugs and methods.

As previously stated, even if the above claim is true, it is important to note that this does not imply that it will never be rational to engage in such practices for the purpose of bringing about a particular future among various possibilities. There may be special circumstances that justify the use of such practices. Even so, I believe that this challenge will trouble many religious believers. They want it to be rational to offer petitionary prayer regularly, not only under unusual circumstances. It is not for nothing that many philosophers, theologians, and religious thinkers have tried to undermine the empirical evidence regarding the efficacy of prayer on theological and philosophical bases.

The comparison between the two practices – homeopathy and petitionary prayer – can be presented by these two propositions:

1. 1. One (central) aim of the use of homeopathy/petitionary-prayer is to bring about a particular future among various possibilities.

2. 2. We are in the same situation regarding the empirical data about the efficacy of both these practices – that is, there is no empirical evidence that supports the claim that homeopathy/petitionary-prayer is effective, and there is even some evidence that it is not effective.

This article can be seen as a discussion of the adequacy of this comparison. If it is not true that one of the (central) aims of the use of petitionary prayer is to bring about a particular future among various possibilities, then the comparison in (1) is false. And if the empirical data should function differently in the case of homeopathy and in the case of petitionary prayer, so that it will not undermine the belief in the rationality of the belief in the efficacy of prayer, then the comparison of (2) is false.

Accordingly, in what follows I will discuss two kinds of attempts to cope with the challenge. The first is to reject the first proposition according to which one aim of petitionary prayer is to bring about a particular future among various possibilities. If this is so, the challenge is just misguided. This kind of response sharply separates the epistemic rationality regarding the belief in the efficacy of prayer from the practical rationality of participating in such a practice. As we will see, according to this view, the rationality of offering petitionary prayer does not rely at all on the efficacy of such prayer to bring about a particular future among various possibilities.

The second response that I will discuss is the rejection of the second proposition by claiming that, unlike the case of homeopathy, in the case of petitionary prayer the empirical evidence should not undermine our belief in the effectiveness of such prayers. According to this line of thought, there is a fundamental difference between empirical evidence regarding homeopathy (or any other empirical hypothesis) and empirical evidence regarding God's way of acting in the world. While we must embrace the conclusion that follows from the evidence of the former, there are good philosophical and theological reasons to reject or to be suspicious or doubtful of the evidence in the latter case. This is due to the fact that the mechanism (if it can be called a mechanism at all) behind the efficacy of petitionary prayer prevents the possibility of testing its efficacy via any empirical observation or study. If we turn back to homeopathy, we can say that homeopathic drugs are supposed to work via a biochemical mechanism while petitionary prayer is supposed to work through an agent, and not just any agent but God himself – with all of his attributes. This, according to the second response, changes the whole story and undermines the empirical evidence regarding the efficacy of petitionary prayers. By that, the advocates of this strategy of dealing with this challenge are rejecting the claim that empirical data pose any challenge to the epistemic rationality for the belief in the efficacy of prayer, and therefore, they also do not pose any challenge to the practical rationality of offering such a prayer.

In general, it is interesting to note that religious believers are generally not moved by the evidential challenge for petitionary prayer. The fact that ‘prayers do not seem to work’ does not make religious people abandon the practice of prayer and does not even lead them to abandon their belief in the efficacy of prayer. Furthermore, no small number of theists (including theist philosophers) consider inquiries into the efficacy of prayers as preposterous – and perhaps even blasphemous. I myself am sympathetic to this view. Part of my aim in this article is to better understand such responses – what, exactly, is ridiculous about the idea of examining the efficacy of prayers? After all, if petitionary prayers aim to change something in the world, it seems reasonable, at least prima facie, that we should have the ability to check the efficacy of these prayers (Dennett 2006, 274–277; Rey 2009, 23–25).

I will begin by addressing the second approach that rejects the second claim and then move on to discuss the first approach that rejects the first claim. So, let's dive in.

Rejecting the second claim: the empirical data should not undermine our belief in the efficacy of prayer

There are several ways to reject the second claim that can be found, or extracted, from the philosophical literature. I will deal with each one separately.

Petitionary prayers are effective, and the fact that we do not have evidence for that should not surprise us

One family of approaches claims that the way God acts in the world in general, and particularly the way he answers petitionary prayers, can explain why, in spite of the fact that He answers prayers, we cannot notice that. In this subsection, I will discuss this kind of approaches.

We do not know what is good for us. Therefore, in response to our prayer for E, God can make E*, which is better for us, happen In Alcibiades II, a Platonic dialogue attributed to Plato (Cooper and Hutchinson 1997, 596–597), Socrates discusses the rationality of petitionary prayer. He remarks:

Do not you agree that there is a great need for caution, for fear you might, all unawares, be praying for great evils when you think you are asking for great goods? Suppose the gods were in a mood to give whatever was asked; it might be just like the case of Oedipus who blurted out the prayer that his sons might take arms to settle their inheritance. He could have prayed for relief from the ills which beset him without begging for others in addition! But in fact, what he asked for came to pass, with many terrible consequences which there is no need to enumerate. (Cooper and Hutchinson 1997, 596–597)

Socrates’ view, as portrayed in the above passage, makes sense in the framework of Greek mythology. However, in the Abrahamic traditions, where God's essential character is perfect goodness, it is impossible that God will answer prayer when, all things considered, it is a bad thing to do (Mawson 2013). In these traditions, then, it is reasonable to think that God will not play along if one prays for a thing that is overall a bad outcome. Additionally, if S prays for E, and E* is what S needs and not E, God will make E* happen and not E. Richard Swinburne used this idea to explain why empirical tests show that God does not answer prayers. In Swinburne's words: ‘When we pray for another person, God knows far better than we do whether it will be best for that person and others affected by him, that he should recover immediately or later or not at all’ (Swinburne 2006, 7). In a similar vein, C. S. Lewis wrote:

I'm not asking why our petitions are so often refused. Anyone can see in general that this must be so. In our ignorance we ask what is not good for us or for others, or not even intrinsically possible. Or again, to grant one man's prayer involves refusing another's. There is much here which is hard for our will to accept but nothing that is hard for our intellect to understand. (Lewis 2020, 79)

If this is so, it is not surprising that we get a negative answer when we test whether prayers for E are effective. In other words, when we want to test whether God answers prayers, we do not know what we should test because God does not give what one asks for, but what one truly needs. It is not that God is hiding, he is just doing what is best. But, as Swinburne and Lewis claim, that makes God's actions much more hidden.

This is an interesting response, and may explain why there is a difference between the blind mechanism of homeopathy and petitionary prayers to God and why we do not see a strong correlation between one's prayer for E and E happening (as opposed to when one does not pray for E). Nonetheless, it also has its implications.

First, suppose we accept this line of thought. In that case, we must conclude that according to our definition of ‘answered prayer’,⁸ whenever God answers E* (what one needs), instead of E (what God is asked for), the petition for E is an unanswered prayer (Finley 2019; Davison 2022, 38). However, as all that is required is a slight modification of our definition of ‘answered prayer’, this may not be a high price to pay.

Second, if God acts in this way, it seems that there is no reason for us to pray for a specific event (E). All we need to do is pray for a random event, and then God will think about what is good for us, and act accordingly.⁹ However, this also does not seem to be a high cost to pay because there are several ways to avoid this conclusion. For example, as has been suggested in the literature (Davison 2017a, 96–146), it is possible that the fact that you prayed for E gives God a reason to actualize E, a reason that God would not have had if you had not prayed for E. Therefore, although God will eventually do what He has more reasons to do, when you pray for E, you increase the likelihood that E will occur. Alternatively, perhaps God only accepts sincere prayers. Thus, if you pray for a random event you do not really want, God will not answer such a prayer. To ensure that your prayer is sincere, you must pray for things that you really want, regardless of what God will do with that prayer. In any case, this kind of solution needs to assume that your prayer for E does not increase the likelihood of the occurrence of E. Otherwise, we would have evidence and arrive back at the evidential challenge.

Third, and most important, the evidential challenge does not disappear even if God actually acts this way. It is true that in that case, we would not see as strong a correlation between one's prayer for E and E's happening as we would expect in the ‘regular’ view. Nevertheless, we would still expect to see a difference in the lives of the people and the communities that practise petitionary prayer. If we think that we know, by and large, what a good life is, religious people and religious communities are supposed to have better lives. As we do not see such differences,¹⁰ this, once again, raises the evidential challenge. Therefore, this solution will not work on its own; in order to do the job, we will need to combine this solution with another one. Let us then move on to the other solutions.

Prayer is a request: God sometimes answers ‘no’ C. S. Lewis wrote extensively on the philosophy and theology of prayer. In one of his reflections on the subject under discussion, published under the title ‘The Efficacy of Prayer’, he writes:

Prayer is request. The essence of request, as distinct from compulsion, is that it may or may not be granted. And if an infinitely wise Being listens to the request of finite and foolish creatures, of course he will sometimes grant and sometimes refuse them. Invariable ‘success’ in prayer would not prove the Christian doctrine at all. It would prove something more like magic – a power in certain human beings to control, or compel, the course of nature. (Lewis 1958, 4–5).¹¹

There are a number of ways to understand this passage. Among them is the widespread idea that the efficacy of prayer cannot be put to the test because prayers are neither medicine nor magic. Rather, petitionary prayer is a request to God, and, like all requests, it may be granted or denied. Louis Jacobs puts it this way: ‘It is said that a little girl prayed repeatedly for a bicycle, without success. “You see,” taunted her unbelieving friend, “God does not answer prayer.” “Oh yes he does,” answered the girl, “His answer was no!”’ (Jacobs 1962, 10).¹² Even if Jacobs’ characterization of petitionary prayer is correct, however, I do not see why there is a problem checking the efficacy of such prayers with statistical methods. Suppose I have a friend, and I want to know how effective it might be to request that he help me. Is there any epistemological problem in testing that? It may be difficult, it would definitely be rude (more on that later), and our findings may not be definitive, but none of that makes the effort impossible or unreasonable.

By the same token, I see no epistemological problem in testing the efficacy of petitionary prayer. Recall the aforementioned scientific studies on the topic. If we see that the members of a certain group (such as the royal family or clergy) who were the subject of prayers for long life died just like the people in other groups, we can take it as a good indication that God did not answer the prayers which were put to the test.

One may argue, however, that even if we discover that the same number of people died in both groups, it is not an indication that God did not answer the prayers – because God's answer can be ‘no’. Therefore, it is possible to claim that God answered the prayers but that the answer differed from what we wished for.

This approach has one main flaw, which concerns the definition of ‘answered prayer’. If we discover that the two groups – one that is prayed for and the other that is not prayed for – have the same death rate, it seems safe to say that God did not answer the prayers of our subjects. In other words, characterizing a prayer to which God answered ‘no’ as an answered prayer fundamentally alters the meaning of the term ‘answered prayer’. In the taxonomy that we have used until now, a prayer that God answered ‘no’ to is an unanswered prayer. In any case, the prayer was ineffective in getting God to comply with our request.

An alternative understanding of Lewis's claim is as follows: there is no epistemological problem in testing the efficacy of prayers, but there is a religious and/or ethical problem. It is inappropriate and even blasphemous to put God on trial. This leads me to the next solution to the evidential challenge for petitionary prayer.

God is hiding!

One other family of approaches to reject the second claim is to see the evidential challenge for petitionary prayer as a token of the problem of divine hiddenness, as these problems share a similar logical structure. In this subsection (‘God is hiding!’) I will deal with these approaches.

The problem of divine hiddenness can be presented as the tension that arises between the idea that a perfect and loving God exists and the fact that, for many of us (who do not experience the presence of God or are not convinced by the arguments that He exists), God is hidden (Van Inwagen 2002; Schellenberg 2006; 2015; Howard-Snyder and Green 2022). Many philosophers and theologians have argued, however, that there are good reasons for God to hide. This explains why many of us do not ‘see’ God regularly. Similarly, the evidential challenge for petitionary prayer will be solved if we have good reasons to believe that God does answer prayers but, for some reason, He hides this fact. Nevertheless, as we will see, even if we embrace this line of thought, it is interesting to ask, ‘How exactly is God hiding?’ – a question that raises metaphysical, ethical, and religious challenges. In this section, I present solutions that take this path, according to which God hides or, at least, acts in a way that makes it harder (or even impossible) to reveal his actions.

Leaving God out of the courtroomIn Deuteronomy 6:16, we read: ‘Do not put the Lord your God to the test.’ This is precisely the verse that Jesus quotes to the Devil when the latter urges him to test God by jumping from the Temple rooftop (Matthew 4:7). What relevance does this have to our topic?

It appears that there is something wrong with putting God in the dock (as Lewis puts it).¹³ However, the error seems to be moral or religious rather than epistemological. Testing the efficacy of prayers, then, may be epistemologically neutral. Nonetheless, a convergence between the epistemological and moral/religious aspects is possible. After all, God is perfectly good, and He will prevent us from doing something wrong or, at the very least, He will not let us get what we want from that trial.

Richard Swinburne has suggested something similar. He writes that if there is a God, He clearly does not want it to be easy for us to know that He exists because if He wanted that, then that would be the case, but it is not the case. Regarding a trial that concluded that there is no evidence that God answers prayers, Swinburne wrote:

But if there is a God, he does not need to answer such prayers in order to do this [let us know of his existence] – if he wanted to do so, he could fill the world with super-miracles. But there is quite a lot of evidence anyway of God's existence, and too much might not be good for us. (2006, 7)

Elsewhere, Swinburne offers an explanation for God acting in a similarly mysterious way. As he writes: ‘The more uncertainty there is about the existence of God, the more it is possible for us to be naturally good people who still have a free choice between right and wrong’ (Swinburne 1998, 2006).¹⁴ Robert M. Adams has proposed a similar idea. According to Adams, God does not reveal how He acts and responds to our prayer because that will destroy the possibility of a real personal and intimate relationship with Him (Adams 1984, 12–15).

Let us suppose that the Swinburne–Adams account is correct. We are left with the question of how God hides the fact that He answers prayers. My claim is that regardless of how we answer that question, we will find ourselves struggling with epistemological, metaphysical, or ethical problems.

I can think of three possible ways to answer the question, ‘How is God so successful at hiding the fact that He answers prayer?’:

1. 1. When we test God, He abstains from any intervention and lets nature take its course.

2. 2. When we test God, He intervenes in our test results in such a way that we will not notice that He actually answered the prayers that we tested.

3. 3. God answers prayers in a way that is not predictable or testable.

We will now examine each of the above possibilities.

When we test God, He abstains from any intervention The first possibility is that God does answer prayers in general but does not cooperate when we put Him on trial. That is why we do not see the efficacy of petitionary prayers whenever we test Him; it looks as if God does not intervene in the world. God ‘pulls his hand from the world’ and lets nature take its course when we put Him on trial.

I think that there are several problems with this line of thought:

1. i. An ethical problem: If God abstains from any intervention when we test him, then we can ‘neutralize’ God whenever we want. If I know that a friend of mine is praying for an outcome I wish would not occur, I can check whether his prayer is being answered (or maybe sign him up for a scientific study that concerns the efficacy of prayer) and God will not intervene. That will raise (but not secure) the chance that my friend's desired outcome will not occur. This option is not flattering to God, to say the least.

2. ii. An epistemological problem: The suggested line of thought assumes that to conduct an experiment, one must wear a white coat, record data, and so on. But we do experiments all the time. Suppose you walk on the street and kick some rocks that lie in your path. Suppose that suddenly these rocks move oddly (not according to the laws of nature as we know them). No doubt this will catch your attention. Moreover, if it continues happening, you will, at some point, start to doubt what you thought until now about the laws of nature. Did you conduct an experiment? You certainly did – although you did not intend to do so. In a similar vein, we ‘test’ the efficacy of prayers all the time, even if we do not intend to. Therefore, if prayers often work, we will notice that. The opposite is also true: if prayers do not work often, we will notice that too – and I would argue that this is actually the case. So, if God abstains from any intervention when we test Him, he must almost always abstain from any intervention, at least as much as is needed to avoid making us suspicious (and if that is really the strategy, He is doing a good job). This means that God is very limited in His actions.

3. iii. A metaphysical problem: If God abstains from any intervention when we test Him, what will He do when we check whether God answered past petitionary prayers? Suppose I want to know what Jews in Egypt prayed for in the eleventh century and, incidentally, I check whether these prayers were answered. Will God change the past? Or maybe God knew all along about my experiment, and therefore He ‘abstains from any intervention’ in advance? Perhaps, but each of these lines of thought raises a problem: the first raises the problem of ‘changing the past’ but maybe most theists will have to deal with that challenge nonetheless. The second raises the problem of ‘divine knowledge and free will’ (in addition to the ethical problem just mentioned). Therefore, if one thinks God cannot or does not change the past, the former solution is not available. If he thinks that God does not know the actions that people will take in the future, the latter solution is not available to him.

For these reasons, I think the solution suggested in this section is unconvincing. In the next section, I will discuss a more promising version of the above solution.

When we test God, He intervenes in our test results in such a way that we will not notice that He actually answered the prayers that we tested The second approach I suggested is that when we test God, God does not abstain from any intervention, but He actually intervenes in our test result, in some way or the other, so that we will not notice that He has answered the prayers we tested.

This line of thought seems much more promising than the previous one. Nevertheless, it also has difficulties. First, embracing this line of thought comes with a certain epistemological scepticism regarding our ability to get to know reality. If God intervenes in how we see reality, then we cannot be sure that what we see is there. This is not an overly serious problem because, as every philosophy student knows, we cannot always rely blindly on our senses. However, in order to embrace this line of thought, we should be more convinced that God exists and that God answers prayers than we are convinced that our senses are reliable when they tell us that prayers are ineffective. This is because we tend to see scientific studies as a reliable source of knowledge. Therefore, to reject, based on a theological belief, the result of a well-conducted scientific study, one should be pretty confident in that theological belief. I am not sure that that is the case for most of us, at least not for me. In other words, this line of thought is available only to those convinced that God exists and that He answers prayers. For the rest of us, it will not work.¹⁵

God answers prayers in a non-predictable way As I said at the beginning of this article, in order to make the belief that ‘prayer is sometimes effective’ a true belief, there must be at least one prayer in human history (including the future) that has been answered. No single observation can ever refute such a claim. Therefore, our everyday observations and the empirical tests mentioned above may be irrelevant to our belief in the efficacy of prayers. It thus seems that the classic theist can keep his belief in the efficacy of prayer regardless of any observation.

But this is, of course, too easy. As I said earlier, this line of thought is compatible with the claim that one is rational in offering petitionary prayer (for the purpose of bringing about a change in a particular future among various possibilities) exactly as one is rational in using homeopathy.

However, one may think that we have some reason to believe that God answers prayers more often than homeopathic medicines work well, yet even so, we cannot see this in our observations. But then we must ask – why is it the case that we do not see this in our empirical tests regarding the subject? In addition to the explanations that I outlined earlier in this section, one more presents itself: God rarely answers prayers, so rarely that we cannot detect His answers with our everyday observations or even with empirical tests.

However, we must continue to ask – why is it that God so rarely answers prayers? I can think of two explanations. First, God randomly answers prayers, and He does so infrequently. This explanation does not strike me as complete. Regardless of how we fill in the details, however, it will not be attractive to the classic theist to think that God acts randomly because this raises many ethical and metaphysical problems (e.g. why does God act in a random fashion?).

The second way to explain the infrequency with which God answers prayers is to claim that few prayers deserve an answer, because of either the request's nature or the requester's nature. This line of thought is common in many religious traditions. For example, one can think that only the prayers of true saints deserve to be answered. And as we all know, there are not many true saints, which is why it is so rare that God answers prayers. Of course, this is only one way of filling in the details, but I think that however we complete the picture, we will need to ask whether it is irrational for us to offer a petitionary prayer.¹⁶ For example, if God answers prayers only for a true saint, and I know that I am not even close to being a true saint, I must conclude that God will not answer my prayer. This implication will not be welcomed by many theists.

However, perhaps there is a different way of filling in the details that has fewer unattractive implications. One can take the ‘sceptic theist’ course and say that he does not know when and why God answers prayer. Anyway, it seems that if we embrace this line of thought – for most people and most of the time – it would be justified for each person to believe that his next prayer will not be answered. He will be, more or less, in the same situation that he is in regarding homeopathy. However, it may still be justified for him to offer such a prayer because of the slight chance it will work or for reasons unrelated to the efficacy of prayer (religious duty, desire to pray, etc.).

Is petitionary prayer really a petition?

In this section, I will discuss the family of solutions according to which the evidential challenge for petitionary prayer is not really a challenge because (1) is simply false. Petitionary prayers do not work, and this is not a challenge because such prayers are not meant to work. As I said at the beginning, this family of approaches separates the epistemic and the practical questions. In other words, the advocates of such an approach claim that even if we are irrational in believing in the efficacy of prayer, this does not undermine the practical rationality of participating in such an activity. This is because the petitioner does not, or should not, aim to work. I will present three possible ways to reject (1) found in the literature, and I will discuss the three suggestions as a unit after presenting each independently.

Petitionary prayer is a duty, not a tool. It is not supposed to workThe renowned Israeli philosopher Yeshayahu Leibowitz (1903–1994) thought that petitionary prayers are ineffective (1987, 111–112). Moreover, he considered such prayers (praying for things to happen) tantamount to blasphemy. In his view, the only legitimate reason to pray is to comply with the religious obligation to pray. For Leibowitz, if one prays for one's own benefit, one is not worshipping God but rather using God and worshipping oneself. In Leibowitz's words:

The sole meaning of prayer as a religious institution is the service of God by the man who accepts the yoke of the Kingdom of Heaven. His acceptance becomes real through his assumption of the burden of Torah and Mitzvoth. Only the prayer which one prays as the observance of a Mitzvah [obligation] is religiously significant. The spontaneous prayer a man prays of his own accord is, of course, halakhically [according to the Jewish law] permissible, but, like the performance of any act which has not been prescribed, its religious value is limited. As a religious act it is even faulty, since he who prays to satisfy his needs sets himself up as an end, as though God were a means for promotion of his welfare. As in the case of any Mitzvah, prayer – especially prayer – is religiously significant only if it is performed because it is a Mitzvah. (Leibowitz 1992, 31. See also Leibowitz 2011, 117)¹⁷

It is worth mentioning that to conclude from this position that petitionary prayers do not aim to work is to conflate one's (legitimate) reasons and motivations to pray with what one is doing in prayer. However, if we embrace this view, alongside the view that petitionary prayers are ineffective, then there is no religious or theological price to rejecting the claim that petitionary prayers aim to work because the efficacy of prayer is of little or no importance.

Petitionary prayers affect the person who prays, not the subject of the prayerIn reflecting on prayer, Søren Kierkegaard wrote: ‘The prayer does not change God, but it changes the one who offers it’ (Kierkegaard 1956, 51). According to one reading of this view (Brümmer 1984, 16–18; Cockayne 2017), it is no surprise that petitionary prayer does not heal people: it does not seek to do so. The role of petitionary prayer is to shape the psychological and spiritual character of the petitioner. Following Brümmer, I will call this the ‘therapeutic’ approach.¹⁸

The therapeutic approach is widespread among modern religious people and thinkers in both Christian (Brümmer 1984, 16–28; Woznicki 2017) and Jewish traditions (Hirsch 2012, 359–360; Kook, 1963, 26; Rosenberg 1978; Achituv 2013, 17–28; Barth 1972, 23–26; Lebens 2022, 183). It successfully evades the evidential challenge for petitionary prayer because it denies its foundational assumption (or at least makes it superfluous): that petitionary prayers aim to work, or at least they aim to work in the same way we thought they were supposed to when we spoke of an ‘answered prayer’. According to this view, prayers do not aim to cause God to act at all. All they aim to do is to affect the petitioner's psychological character. The right way to pray, according to this view, is not to ask God for something in order that it will happen but rather to use prayer as a tool for religious and spiritual upbuilding. If so, according to those who hold what we might call the Kierkegaard-inspired view, the evidential challenge for petitionary prayer is a non-starter.¹⁹

What do religious people do when they pray?I shall conclude with a final approach to our challenge, one that I see as the more promising and convincing one, which is based on George Rey's notion of ‘meta-atheism’. Rey made the following claim: ‘Despite appearances, many Western adults who've been exposed to standard science and sincerely claim to believe in God are self-deceived; at some level they believe the claim is false’ (Rey 2009, 3). In a similar vein, and regarding the topic of prayer specifically, Yeshayahu Leibowitz wrote:

You do not really think that your morning prayer somehow influences what is happening to you. A religious person might say that he believes in the efficacy of prayer, but this is not true, and I can prove it. In the [Jewish] morning prayer, one can pray with full intent and sincerely say ‘He who heals the sick of Israel’; but if that person, or one of his children, falls ill, he will go to the doctor and behave exactly like the atheist who does not pray at all. There are no differences between them! In the morning, he says ‘The one who provides food for each creature's needs’; but if he needs money, he goes to work . . . exactly as the atheist does . . . This means that it is not important that this person says he is praying because he thinks that it may help him be healthy, rich, or safe . . . he does not really think so. (Leibowitz 1987, 111–112, my translation)

If Leibowitz and Rey are right, we can understand why the evidential challenge does not move religious people. The idea that petitionary prayer does not seem to work does not bother them because they do not believe in the efficacy of prayer. It also addresses the challenge itself. According to Rey and Leibowitz, there is a sense in which many religious people, taking part in the practice of petitionary prayers, are deceiving themselves and behaving in an irrational way.²⁰

A similar approach, albeit one that is perhaps more respectful towards religious individuals, is endorsed by D. Z. Phillips. As an introduction to Phillips's philosophy on this matter, I will share a story. A few years ago, I received a call from someone conducting a survey on religious diversity. A friend who was with me took the phone from my hands and announced that he wished to participate. The caller then asked, ‘Do you self-identify as a religious person?’ ‘Not anymore’, my friend answered. ‘I prayed for something last month, and God didn't answer my prayer – so I think that I do not believe in Him anymore.’

My friend was joking, of course, and indeed – hearing this answer, both the caller and I started to laugh. But what was so funny? If endorsement of the efficacy of prayer is an essential part of religious belief and practice, my friend's answer (or at least a similar answer) makes perfect sense. It seems reasonable that if one sees that God does not answer prayers, one will stop believing in a God who answers prayers and, furthermore, will stop offering petitionary prayers to that entity. But a look at how petitionary prayer functions in the lives of religious people tells us differently. As Rey and Leibowitz rightly noted, people generally do not stop praying because they see that prayers often appear to go unanswered. Perhaps, then, philosophers are missing the boat in terms of petitionary prayer. This leads me to Phillips's work on the nature of prayer.

In his seminal work ‘The Concept of a Prayer’, Phillips considers prayer as it functions in the life of religious people, and not as philosophers and theologians describe it. He distinguishes between petitionary prayer and superstition. According to Phillips, petitionary prayer is not an actual ‘request’, at least not the kind of request we find between human beings. As he puts it:

prayer of petition is best understood, not as an attempt at influencing the way things go, but as an expression of, and a request for, devotion to God through the way things go . . . When deep religious believers pray for something, they are not so much asking God to bring this about, but in a way telling Him of the strength of their desires. They realize that things may not go as they wish . . . (Phillips 1965, 120–121)

In order to better understand this approach it will help us to compare it to Kierkegaard's approach. Both approaches agree that petitionary prayer is not really a petition. However, there is a crucial difference between the two approaches. Kierkegaard's approach comes from the inside. He is making a religious claim, not a claim about religious people. Kierkegaard does not try to claim how prayer is functioning in the life of religious people, but rather how it ought to function. Therefore, he is not interested in observations and reflections of religious people. In contrast, Rey and Phillips²¹ are trying to understand and present the nature of petitionary prayer as it is actually taking place in the life of religious people. Their claim is purely descriptive and lacks any normative dimension.

The Rey–Phillips approach also deflects the evidential challenge for petitionary prayer because, like the first two approaches I discussed (those of Leibowitz and Kierkegaard), it rejects the assumption that petitionary prayer aims to ‘influence the way things go’. Additionally, this approach also explains why many theists (as well as some non-theists) feel that there is something misguided or even ridiculous about inquiring into the efficacy of petitionary prayer. For Phillips, the philosopher or scientist who investigates the efficacy of prayer does not understand the nature of petitionary prayer, and confuses superstition and magic with petitionary prayer (Phillips 1965, 112–130). If this is so, the evidential challenge is in fact a dead end.

Discussion: what is the problem with this approach? The main problem with this approach is that it will not satisfy the traditional theist, regardless of how we fill in the details. This is because the traditional theist sees her petitionary prayers as petitions. When she prays for God to protect her from danger, she at least thinks that she is requesting, via this prayer, that God protect her. Therefore, although this approach will let her off the hook regarding this particular challenge, she will probably not accept it.

While this problem may be a bug for Kierkegaard who tries to defend the traditional theist, to the Phillips–Rey approach, it is more likely to be a mere feature or, at least, not a significant bullet to bite. This is because an essential part of their theory involves a lack of self-awareness on the part of religious people. In other words, in his account, Kierkegaard tries to satisfy the traditional theist while Phillips–Rey do not. Therefore, the rejection of their account of petitionary prayer by a traditional theist would pose no problem for them, while it does pose a problem for Kierkegaard. Nevertheless, it is important to note that according to the Phillips–Rey approach, although the challenge as stated is misguided, there may be a challenge to the rationality of petitionary prayer because many times it involves some kind of self-deception.²²

Another objection that can be raised against the Phillips–Rey approach is that it may not fit with all cases of petitionary prayer. There certainly seems to be a difference between asking God for general help with coping with the difficulties of life, and asking God for help when one is in a boat that is about to sink in shark-infested waters. In the former, the Phillips–Leibowitz–Rey approach may fit well, whereas in the latter, things seem less promising with such an approach.

I deliberate between two ways that the Phillips–Rey advocate should respond to this challenge. The first is to bite the bullet by accepting that there are times when the Phillips–Rey approach is not convincing. In such cases, the evidential challenge is still a challenge that should be addressed with the classic philosophical and theological tools that I presented and investigated earlier.

The second approach or response is to reject the claim that the Phillips–Rey approach is indeed unconvincing regarding the latter cases such as an urgent prayer for help when one is in danger. It is hard to deny that in such cases one is expressing his deep desires to be saved – but does he really ask for anything? I am not sure. If when writing that ‘There are no atheists on turbulent airplanes’ (Jong 2003, 234), Erica Jong meant that in such situations everybody prays for help, then this may indicate that people's prayer in these situations is not really a request from God but rather a mere expression of their strong desire to survive. If all people, regardless of their religious belief, tend to pray in these circumstances, perhaps the prayer is not addressed to God but rather represents a mere psychological phenomenon according to which people in extreme circumstances tend to express (to themselves or to others) their fears.

Conclusion

In this article, I presented the evidential challenge for petitionary prayer and discussed various ways to approach it. In light of the relatively limited attention that the challenge has attracted in the philosophical and academic literature, I hope to bring this exciting question back into the contemporary discussion in the field of philosophy of religion.