For over fifteen years by now, debunkers and defenders of theistic belief have debated whether or not cognitive science of religion (CSR) undercuts belief in God. Debunkers argue that CSR theories show supernatural beliefs to be typically produced by belief-forming and belief-sustaining processes that are epistemically unreliable (e.g., Braddock, 2016; Davis, 2020; Kvandal, 2022; Nola, 2013). From this they infer that most believers’ god-beliefs are unjustified. A popular response is that CSR cannot undermine belief if the believer has independent evidence or reasons to believe that God exists (e.g. Jong & Visala, 2014; Launonen & Visala, 2023; Thurow, 2014a, 2014b, 2023; Van Eyghen, 2020). Thurow (2023) has recently called this response the Religious Reasons Reply. In a paper, he squeezes the reply into a single sentence: “[M]ost believers in a god have fairly specific beliefs about that god and those specific beliefs are supported by various arguments that believers typically find plausible—e.g. cosmological arguments, design arguments, arguments from miracles and religious experience, and the like and reliance on testimony from those who are familiar with these arguments” (Thurow, 2018, 300).

Many debunkers agree that having such reasons can get the believer off the hook (e.g., Davis, 2020, 205; Kvandal, 2022, 192; Nola, 2013, 185). However, they usually deny that the Religious Reasons Reply helps salvage most believers’ belief. Four distinct worries can be raised regarding the reply. First, it is unclear whether most believers in God are aware of good, independent reasons such as theistic arguments. Second, while many reasons like cosmological arguments and design arguments may support thin belief in a transcendent creator, they do not support thick belief in a relational God. Many Christians believe that God is sometimes present to them, God guides them, and God speaks to them. Such beliefs have great existential and practical implications for the everyday lives of many people. While this worry is rarely raised, the debunking of such beliefs would be bad news for numerous believers. Third, most believers who do invoke independent reasons to believe may be rationalizing (Leben, 2014). That is, they could be citing evidence such as theistic arguments without the arguments having any causal influence on their belief in God. Fourth, the reasons could themselves be undermined by cognitive science (de Cruz & de Smedt, 2015; Teehan, 2014).

This paper focuses on everyday religious experiences (EREs) as reasons to believe, especially on the experience of hearing God speak. The evidential nature of such experiences is compared with another popular reason to believe: the appearance of design and beauty in the natural world. The paper is divided into four main sections based on four questions:

  1. 1.

    What reasons do Christians have for believing in God?

  2. 2.

    What kinds of beliefs about God can the reasons support?

  3. 3.

    Are the reasons rationalizations?

  4. 4.

    Can cognitive science undercut the reasons?

What reasons do Christians have for believing in God?

Debunkers often argue that even if “a few philosophers and theologians are epistemically justified in holding religious beliefs … few religious believers actually do hold their beliefs for reasons that are independent” of the beliefs’ underlying, epistemically unreliable evolutionary causes (Davis, 2020, 205). In other words, the god beliefs of most Christians and other religionists can be explained without any reference to theistic arguments or other epistemically reliable influences. People believe in God or gods, say, because human brains are hard-wired for making sense of the world in terms of intentional agency (Barrett & Lanman, 2008) or because belief in moralizing gods produces prosocial behavior that benefits the group (Norenzayan, 2013). Whichever exact theories debunkers opt for, such cognitive and cultural belief selection processes are typically not truth-tracking regarding supernatural agents.

One reason to think most Christians do not hold their beliefs for respectable reasons is that they may simply be unaware of reasons such as theistic arguments. While nowadays anyone with an internet connection can find discussions on theistic arguments, it is still doubtful whether most Christians in the world are familiar with them. So to know what kinds of reasons Christians actually have for believing in God, we should ask them. Unfortunately, surveys looking into this question are hard to find. A few scholars engaged in the CSR debunking debate have referred to a study in Michael Shermer’s pop science book How We Believe (2003). Sociologist Frank Sulloway led a study where around one thousand randomly chosen Americans were asked two questions.

“Do you believe there is a God (a purposeful higher intelligence that created the universe)?”

“In your own words, why do you believe in God, or why don’t you believe in God?”

The most popular reasons to believe in God were these:

  1. 1.

    Good design/natural beauty/perfection/complexity of the world or universe (28.6%).

  2. 2.

    Experience of God in everyday life/God is in us (20.6%).

  3. 3.

    It is comforting, relieving, consoling, gives meaning and purpose to life (10.3%).

  4. 4.

    The Bible says so (9.8%).

  5. 5.

    Just because/faith/need to believe in something (8.2%).

  6. 6.

    Raised to believe in God (7.2%).

  7. 7.

    God answers prayers (6.4%).

  8. 8.

    Without God there would be no morality (4.0%).

  9. 9.

    God has a plan for the world, history, destiny, and us (3.8%).

  10. 10.

    To account for good and avenge evil in the world (1%).

    (Shermer, 2003: 272–273).

In this paper I shall focus on the two most popular reasons to believe: the “good design/natural beauty/perfection/complexity of the world or universe” and the “experience of God in everyday life/God is in us”. The first reason is, of course, the stuff of design arguments. Perhaps several respondents were aware of claims about intelligent design or of the fine-tuning argument for God. But more probably most people were not trying to point to a formal argument. They were simply pointing out how the wonders of nature provide a reason to think there is a creator.

“Experience of God in everyday life/God is in us”—or what I shall call EREs—are presumably low-level experiences of divine presence, guidance, and communication. Many Christians speak of the importance of having a personal relationship with God. EREs foster a sense of a reciprocal relationship with the divine. Since all types of EREs cannot be considered here, I shall focus on experiences of divine communication or “hearing God speak” (which, however, often coincide with experiences of divine presence and guidance). “Many might be surprised to discover”, Christian thinker and author Dallas Willard writes, “what a high percentage of serious Christians—and even non-Christians—can tell of specific experiences in which they are sure God spoke to them” (Willard, 2012, 21). While experiences of hearing an external, audible voice are less common, many believers report often having low-level experiences of divine communication. Anthropologist Tanya Luhrmann (2012, 2020) has documented and analyzed a variety of such experiences as part of her field work around the world, most famously at the evangelical-charismatic Vineyard churches.

At the Vineyard, God was understood to speak back in several ways. To begin with, he spoke through the Bible. When congregants remembered scripture or felt powerfully moved or affected by a particular passage as they read, they might infer that God spoke to them through that passage: that He led them to it in order to have them read it and respond to it. […] God was also understood to speak through people and circumstances. Congregants would describe events that might seem to be coincidences, but they would say that God was speaking to them through these circumstances in order to communicate something to them: that He loved them or wanted them to make this decision or that one. […] Congregants at the Vineyard church also expected God to speak back to them by placing mental images or thoughts (sometimes called “impressions”) in their minds or by making their body feel a certain way [e.g., by giving them goosebumps]. (Luhrmann, 2020, 49.)


Three remarks about such experiences are of importance here. First, experiences of hearing God speak (as well as of other EREs) seem globally widespread. Luhrmann observes that, “the basic ideas about who God was and how God would interact were very much the same in Chennai and in the United States. God was a person. He was your best friend. You should talk to Him about everything, and He would give you what you needed because He loved you.” (Luhrmann, 2020, 91). Such experiences are associated with Pentecostal and charismatic Christianity—the fastest growing form of Christianity in the world (and only a bit behind Islam; Zurlo et al., 2021). However, reports of divine communication are not restricted to charismatic movements. The idea that God speaks directly to individual believers is normal in lived Christianity. Moreover, stories of God speaking to people can be found throughout the Bible and the writings of mystics and theologians (Cook, 2019).

Second, such experiences have great evidential value to the believers who have them. Luhrmann argues that even the most devout Christians have trouble believing in God’s reality, and they struggle to feel his presence in their everyday lives. However, the “moments when someone hears a god, sees a spirit, or feels the presence of the dead are important because they become evidence, for the person of faith, that does not rely on the testimony of others” (Luhrmann, 2020, xiv.)

Third, EREs have great existential and practical implications for the daily lives of many Christians. Luhrmann’s congregants reported that experiences of hearing God’s voice influenced decisions about what to study and where, whether to accept a job or not, whom to befriend or marry, whether to move to a different city, and even what color shirt to wear on a particular day (Luhrmann, 2012, 39–100). The debunking of voice-hearing experiences would be bad news to millions of Christians. It would suggest that their reciprocal relationship with God is largely imagined.

What kinds of beliefs about God can the reasons support?

This second question has two aspects: First, what kind a of god concept (which set of divine attributes) does the reason cited by believers point to? Second, how well does the reason justify belief in the reality of such a being? Consider first the appearance of design/natural beauty/perfection/complexity. This data points to the existence of a creator/designer deity. Presumably this being is also a powerful one (since creating is no small feat) and transcendent (since an agent cannot be captive in se to what it brings into being). Design and beauty also imply that creation was not an accident. The world has a purpose. This again suggests that the creator/designer has personhood. And since design, beauty, and purpose are obviously good things, we may also infer that the creator/designer is benevolent. Therefore, the wonders of nature seem to support belief in a transcendent, powerful, personal, and benevolent creator/designer (a few other theistic attributes may be added to the list). The second aspect of the question pertains to the evidential force of the data of nature. Here it suffices to note that there are respectable arguments that the appearance of design and beauty do offer a good prima facie reason for believing in a theistic creator/designer (e.g., Evans, 2010).

But to most religious believers God is much more than a transcendent creator who got the show rolling. Most Christians believe in a relational God who is actively involved in their lives. In his Religious Reasons Reply, Thurow points out that “most believers in a god have fairly specific beliefs about that god and those specific beliefs are supported by various arguments…” (Thurow, 2018, 300). Well, which specific beliefs? Which arguments? In fact, there is a disconnection between popular theistic arguments and the actual theistic beliefs of Christians. Most arguments, such as cosmological arguments or design arguments, merely support thin belief in a creator God, not thick belief in a loving Heavenly Father who knows me, answers my prayers, guides my path, and has a plan for my life. Likewise, the appearance of design and beauty in nature do not seem to give good grounds for believing in God’s relational attributes and activity. In fact, it is at least as good evidence for deism as it is for Abrahamic theism. After all, deists deny that God intervenes in the universe and in the lives of people partly because such beliefs cannot be deduced from universally available data (Rowe, 1998).

However, experiences of divine presence, guidance, and communication may serve as reasons to believe in a relational Heavenly Father. But are EREs, such as experiences of hearing God speak, good reasons to believe? While there is not enough space in this article to defend the evidential force of religious experience, many philosophers of religion defend an epistemic attitude that has been called the principle of credulity (Swinburne, 2004) or the critical trust approach (Kwan, 2009). According to this approach, experiential beliefs are prima facie justified or “innocent until proven guilty”. For example, if it seems to me that God is speaking to me, I am justified in believing that God is speaking to me for the same reason I am justified in believing that another human person is speaking to me when it seems to me so. Of course, my experience could be easily defeated.

One may question, however, whether beliefs grounded on voice-hearing experiences enjoy even prima facie justification. Luhrmann’s work, as well as Christian books on how to hear God’s voice (e.g., Willard, 2012), indicate that in a typical experience there is no perception of an external, audible voice. The “prayer experts” Luhrmann got to know at the Vineyard “spoke as if what they were learning to do was to take their inner sensory world more seriously, to treat their thoughts and images and sensations as more meaningful, and to blur the line deliberately between what they might once have attributed to an internal cause and what they might now wish to attribute to an external one (Luhrmann et al., 2010, 72). God was taken to speak through one’s own flow of thoughts or “inner speech”. But as the congregants themselves noted, such divine messages get easily mixed up with one’s own desires, emotions, and beliefs. When Luhrmann asked a pastor about the risks of mishearing God, he pointed out that faith “is about taking risks” (Luhrmann, 2012, 84). Some congregants also gossiped about others who thought they had heard from God but who probably just heard what they wanted. Importantly, there seemed to be no clear phenomenological difference between ordinary mental and bodily experiences and purported divine messages. As one man said, “Sometimes when we think it’s the spirit moving, it’s just our burrito from lunch” (Luhrmann, 2012, 70). So while Christian philosophers have made a good case that we are justified in trusting our experience if it seems to us like God is speaking, does it really seem to one that God is communicating if the experience does not differ from having poor digestion? While probably sometimes there is a difference, God is generally taken to speak with a “still small voice”. He is rarely loud and clear. Therefore, most experiences of hearing God speak are weak experiences. And weak experiences are weak evidence.

But if so, how do believers know when God has spoken? The skill for separating divine messages from “thoughts and images and sensations” with no divine origin is called discernment. At Vineyard, people were encouraged to apply four tests (similar tests are typical in books about hearing God):

  1. 1.

    Was it something the person would think on their own?

  2. 2.

    Was it something God might say—did it reflect the character of God?

  3. 3.

    Could it be confirmed through circumstances or through other people’s prayers?

  4. 4.

    Did it bring a feeling of peace and comfort? (Luhrmann, 2012, 63–65.)

Epistemically speaking, such rules are a double-edged sword. On the one hand, believers are subjecting their experiences to evaluation. On the other hand, such evaluation is based on theological presuppositions. The rules of discernment are based on beliefs about a relational God who talks to people in this way, speaks these kinds of things, and makes people feel that way. In fact, the rules themselves may be a partial cause of experiences of hearing God speak. They may lead people to interpret random mental events, bodily sensations, and life situations as containing a divine message. So not only are experiences of hearing God’s voice weak ones, they are also heavily driven by theological presuppositions. This makes such experiences weak reasons to believe.

Nonetheless, perhaps many voice-hearing experiences could still count as what Harold Netland calls ‘soft religious experiences’. Such experiences have evidential value when supported by appropriate background beliefs.

Soft religious experiences are experiences that could plausibly be interpreted in strictly naturalistic terms but that, given broader considerations, might also plausibly be understood religiously. Whether one views the experience in naturalistic or religious terms depends on the background beliefs one brings to the experience. An example of a soft religious experience might be someone who, having spent considerable time in prayer and fasting over an important decision she faces, claims to have heard God’s “voice” directing her to a particular course of action. In the appropriate circumstances and with appropriate background beliefs, it can be reasonable for her to believe that she has experienced God communicating with her. At the same time, it could also be reasonable for someone with a naturalistic worldview to regard this as yet another example of wish-fulfillment, which has a perfectly reasonable naturalistic explanation. (Netland, 2022, 60.)


Now, do most Christians have justified background beliefs that allow them to attribute some of their thoughts, images, and sensations to God? Above I suggested that the data of nature may ground belief in a powerful, personal, and good creator/designer deity. Perhaps this justified background belief can, in turn, justify the belief that certain soft religious experiences are caused by this very creator/designer. After all, EREs are obviously experiences of the presence, guidance, and communication of a good and personal being. So if experiences of hearing God speak enjoy prima facie justification, they can support beliefs about a relational God. But prima facie justification is cheap. Before considering defeaters, we need to address another worry.

Are the reasons rationalizations?

So far I have argued that EREs such as experiences of hearing God speak might offer many (perhaps most) Christians sufficient prima facie reasons to believe in a relational God. Some philosophers think that the reasons cited by many Christians are merely rationalizations. A rationalizing person is someone who gives a false account of why she or he believes that p. Thurow offers a modified version of Robert Audi’s definition of rationalization:

(Rat) A first-person rationalization, by S, of S’s belief that p (call it B) at t, is a purported account of B, given by S, which (a) offers one or more reasons, R, for B, (b) represents the belief as prima facie rational given R; but where (c) R does not explain why S believes p at t. (Thurow, 2023, 138.)


The last bit is important. A false reason cited by a rationalizing person does not explain why she has the belief that p while a real reason does explain it. A real reason to believe that p is therefore also a cause of the belief that p.Footnote 1 Derek Leben thinks believers citing theistic arguments are rationalizing because such arguments, unlike CSR mechanisms, do not serve as good explanations or predictors of belief.

[I]f agent detection mechanisms are a better predictor of intuitions underlying these beliefs, then sources such as the cosmological argument or historical arguments from divine revelation might be dismissed as rationalizations as well. Is it the case that psychological mechanisms are better predictors of deontological and religious intuitions? I suggest that they are. None of the standard arguments for the existence of God are convincing or influential enough to be the source of religious beliefs for more than a handful of people. The fact that theists tend to reject the same type of historical argument in other religions suggests that this, too, is a poor predictor of religious beliefs. (Leben, 2014, 18–19.)


There are several problems with Leben’s claims (cf. Thurow, 2014b). Importantly, even if psychological mechanisms posited by CSR were better predictors of intuitions underlying theistic beliefs, this would not mean that theistic arguments have no causal influence on theistic beliefs.Footnote 2 In order to serve as a justifier, a reason must have some causal influence on belief, but it need not be the sole cause of belief. Beliefs rarely have a single cause (cf. McBrayer, 2018). Even if Leben’s point has weight, say, in case where a nonreflective lay believer cites a counterintuitive reason such as the ontological argument, it seems that the reasons mentioned in Sulloway’s study do play a causal role in originating and/or sustaining belief in God.

Cognitive science provides evidence that intuitions of design in nature and EREs cause god-beliefs. For example, Deborah Kelemen and her colleagues have found that school children (Kelemen, 2004), adults (Kelemen & Rossett, 2009), and even professional physical scientists (Kelemen et al., 2013) display teleological and purpose-based reasoning. Kelemen has suggested that this disposition makes humans “intuitive theists”, since we intuitively infer that the purpose-driven universe must have a designer.

Likewise, EREs can explain why many Christians have beliefs about a relational God. Such experiences are likely to be underpinned by (in)famous CSR mechanisms such as HADD (hypersensitive agency detection device) and the Theory of Mind. A great deal of debate and confusion surrounds the role of these mechanisms in explaining religious beliefs (cf. Atkinson, 2023). In a programmatic paper, Van Leeuwen and van Elk (2019) survey several attempts to test HADD. To their knowledge, “no one has obtained positive evidence for the effects of agency detection on supernatural beliefs” (p. 227). However, they don’t think agency detection explains nothing. They differentiate between general religious beliefs and personal religious beliefs. General religious beliefs, like the belief that God exists, are “culturally widespread representations of supernatural agents (among other things) that do not indexically refer to the believers themselves”, while personal religious beliefs such as that God visited me in the hospital “are directly and indexically about the very person who has the belief” (Van Leeuwen & van Elk, 2019, 221). This distinction partly overlaps with my own. The belief that there exists a God with such and such relational attributes counts as a general religious belief, but personal religious beliefs are always beliefs about a relational God. EREs give rise to personal religious beliefs.

Van Leeuwen and van Elk argue that general religious beliefs are not explained by individual differences in HADD or Theory of Mind—or what they call hypersensitive agency detection capacities (HADC)—but by social learning fostered by credibility-enhancing displays (Lanman & Buhrmester, 2017 Maij et al., 2017). If you happen to grow up surrounded by believers who practice what they preach, you are also likely to become one yourself. However, HADC are crucial for explaining religious experiences that undergird personal religious beliefs. The authors summarize their Interactive Religious Experience Model (IREM) as follows:

IREM: Religious believers arrive at general beliefs about supernatural agents mostly by way of cultural learning from others in their society. But given their general background religious beliefs, believers may further seek experiences that allow them to form personal religious beliefs as well. Hence, many religious believers seek out situations that trigger HADC and other socio-cognitive biases, like teleological thinking and meaningful coincidence. Such situations include rituals, prayer, enactments, various forms of sensory deprivation, and even playing make-believe that a supernatural agent is present. The low-level intuitions triggered by such situations then allow religious believers to form personal religious beliefs, since those experiences are their experiences as if an agent is present, as if something happened for a reason, as if a sign has been sent, etc. A HADC-based agency-intuition allows a believer to transition from merely having the general belief that God exists to having the personal belief that God visited me (and so on, mutatis mutandis, for other triads of general belief, low-level intuition, and personal belief). Intuition and experience, interpreted in light of general belief, ground personal belief. (Van Leeuwen & van Elk, 2019, 233–234.)


The IREM converges with Luhrmann’s findings. At Vineyard people actively sought to experience God not only by focusing on their inner “thoughts, images, and sensations”, but also by engaging in pretend-play. For example, some set an extra dinner plate for God or went on “date nights” with God (Luhrmann, 2012, 39–100). Van Leeuwen and van Elk comment that, “Vineyard members, who antecedently generally believe that God exists, deliberately create situations that cause their HADC to fire. The resulting agency-intuitions allow them to link their abstract, general representation of God to a time and place that is experienced from their own ego-centric point of view. One then forms the personal belief that he or she was with God at that time and place.” (Van Leeuwen & van Elk, 2019, 240.)

EREs, therefore, seem to play a causal role on belief. Such experiences may not serve as the origin of the general belief that there exists a relational God. After all, the Bible clearly depicts God as a relational being, and presumably children first learn about such a God from their parents and pastors. However, EREs clearly serve to strengthen and sustain the belief that God is truly relational. There is also evidence that people who tend to lack such experiences, like people on the autistic spectrum, also lack beliefs about a relational God (Norenzayan et al., 2012). This also means that believers citing EREs as reasons to believe are not rationalizing. Moreover, implicit cognitive processes and explicit reasons cited are not always competing explanations of belief. Of course, the naturalistic explanations of EREs would seem to cast doubt on the beliefs’ veracity. We now turn to this issue.

Can cognitive science undercut the reasons?

Philosophers defending the critical trust approach to religious experience usually agree that experiences may lose their justification in the face of defeaters (Kwan, 2009). A successful undercutting defeater shows that what once seemed like a good reason to believe that p, is, in fact, not a good reason to believe that p (Pollock, 1986). Cognitive science can undercut belief in God by showing that the intuitive appeal of some reason—such as a theistic argument or a religious experience—is not appropriately independent but relies on epistemically unreliable cognitive processes (cf. de Cruz & de Smedt, 2015; Teehan, 2014; Thurow, 2014a).

Consider why psychologist Deborah Kelemen has described our teleological intuitions as ‘promiscuous’. When put under time pressure, even professional physical scientists endorse claims such as “trees produce oxygen so that animals can breathe”, “germs mutate in order to become drug resistant”, and “the sun makes light so that plants can photosynthesize” (Kelemen et al., 2013). Such evidence could be used to argue that humans tend to attribute goals and purposes to non-teleological natural objects and events. Studies like this, however, do not automatically make intuitions of design unreliable. Firstly, the participants had less than three seconds to judge a claim as true or false. The purpose was to give enough time for their intuitive System 1 to react, but not enough for their reflective System 2 to step in. Given a few more seconds to respond, the participants would have probably answered correctly. So while implicit intuitions do inform and constrain people’s explicit beliefs, in many cases reflective thinking debugs explicit beliefs from misleading intuitions. Evidence for promiscuous teleological thinking, therefore, does not necessarily confer unreliability upon teleological beliefs that survive critical scrutiny. We can study cells under a microscope or the rotation of celestial bodies through a telescope and reflect on our experience without the intuition of design wearing off. Secondly, claims like “trees produce oxygen so that animals can breathe” are not obviously unwarranted. The role of teleological notions in biology is debated, and they are “largely considered ineliminable from modern biological sciences” (Allen, 2020). Thirdly, theists have good reasons to believe that there are in fact goals and purposes in nature. Unless we presuppose that there are none, evidence of ‘promiscuous teleology’ is not a reason to mistrust intuitions of design.

Does cognitive science undercut EREs and the belief in a relational God? We saw that, according to Van Leeuwen and van Elk at least, such experiences are underpinned by HADC. There is a well-known debunking argument based on the hypersensitivity of HADD/HADC (Nola, 2013). The HADD/HADC is said to be a trigger-happy, better-safe-than-sorry mechanism. It is prone to produce false positives (‘detections’ of a hidden agent when there is none) because they are far more harmless than false negatives (failures to detect a dangerous animal or an enemy). This makes HADD/HADC epistemically unreliable. We may apply the debunking argument to the reliability of EREs as follows: if HADD/HADC causes us to feel God’s presence, makes us view a random coincidence as an occasion of special divine providence, or leads us to interpret a part of our inner speech as God’s message, we should not trust such experiences. They are proven guilty, and the resulting beliefs are debunked. Now, perhaps this claim is too sweeping. Not only is it doubtful whether agency detection tendencies are in fact hypersensitive (e.g., Maij et al., 2019), but their causal role behind various EREs is unclear.

Let us, therefore, focus again on the experience of hearing God’s voice. Above I have argued that in most cases their evidential value is weak. Most voice-hearing experiences do not involve an audible, external voice. Instead, they are heavily theology-laden interpretations of one’s own “thoughts and images and sensations”. While such experiences may enjoy prima facie justification, their evidential force is so weak that it is easily defeated by scientific knowledge about our proneness for agency. But as the work of Luhrmann and others (e.g. Cook, 2019; Dein & Cook, 2015) shows, many Christians also report unexpected experiences of an audible voice. Such strong experiences may not count as everyday experiences in that they do not happen regularly, but they are not highly uncommon either. For instance, out of forty participants from an English Pentecostal group who completed a questionnaire on prayer, 25 people reported hearing God’s voice and 15 hearing Him aloud (Dein & Littlewood, 2007).

McCauley and Graham (2020) provide a naturalistic account of religious voice-hearing experiences as auditory verbal hallucinations (AVHs) or as ‘thought insertions’. AVHs happen when one hears “words or speech as if from another agent or person but in the absence of an actual external speaker” (McCauley & Graham, 2020, 49), although AVHs don’t always include auditory or acoustic phenomenology. In their view, AVHs are caused by a failure in source monitoring. In everyday life most people have no trouble distinguishing their own inner speech from the voices of others. But an error may occur “when multiple, simultaneous inputs overwhelm a person's source-monitoring capacity or when speech perception is degraded by external circumstances (e.g., a noisy crowd), by a person's stress and anxiety, or by distraction, inattention, or interference (including difficulties arising from focusing on the content of an experience or an utterance)” (McCauley & Graham, 2020, 61).

McCauley and Graham discuss also other cognitive systems that presumably underpin AVHs: linguistic processing, agency detection, and Theory of Mind. A failure in source monitoring may trigger agency detection and Theory of Mind to attribute linguistic content to an external source. Cultural inputs are also relevant. Citing Luhrmann’s work, the authors point out that “the prevailing cultural norms may encourage or bias people to mistake their private inner speech acts for voices and to misattribute them to an origin or source other than themselves” (McCauley & Graham, 2020, 62). That is, an AVH is easily attributed to a supernatural agent prevalent in one’s environment, such as an ancestor spirit or a relational God. Self-narration is another important building block of religious AVHs. Humans weave their memories into a coherent story, and believers do so with their relationship to God. If one prays to God for guidance in a particular matter, for instance, one may later interpret an inner voice as God’s answer (McCauley & Graham, 2020, 65–68).

Christian philosophers defending the veridicality of religious experiences are unlikely to be deterred by such accounts. One reason is that a naturalistic theory cannot take the possibility of supernatural communication into account. While some scholars use the word hallucination in a neutral sense without the implication that AVHs are nonveridical and thus never caused by God (e.g., Cook, 2019), McCauley and Graham do not intend to leave room for supernatural causes (personal communication with George Graham). Talk about ‘errors’ or ‘failures’ in source monitoring implies that no one is speaking. Since theists do not share that starting point, for them such a scientific explanation is not necessarily superior to a theological one. Unless we presume that there is no relational God who speaks to people, why should believers think that all experiences of hearing God’s voice are merely the result of cognitive malfunction?

There is a worry, however. In my personal experience, when charismatic Christians suspect that a congregant has a psychotic disorder like schizophrenia, they treat reports of divine messages from such a person with healthy skepticism. While few believers would deny that God may sometimes speak to such persons, their proneness to hallucinate makes their experiences unreliable. Thus believers agree that some religious voice-hearing experiences are due to cognitive malfunction. However, such skepticism should cast doubt also on the voice-hearing experiences of mentally healthy believers. There is good evidence that their experiences are undergirded by a similar cognitive process. Obviously, I am not saying that everyone who hears voices is psychotic. There are important differences between most experiences of hearing God and the voices heard by schizophrenia patients (Luhrmann (2012, 227–266). For example, people with schizophrenia often hear harsh and commanding, disturbing voices, but God’s voice is nothing like that. However, this does not mean that two kinds of experiences are undergirded by a wholly different cognitive process.

McCauley and Graham argue that religious experiences often overlap with mental disorders. There are continuities between normal cognition, religious cognition, and disordered cognition. While CSR highlights continuities between normal cognition and religious cognition, psychopathology recognizes continuities between normal cognition and disordered cognition. For instance, while psychotic experiences like hallucinations are the key feature of schizophrenia, they are not uncommon among mentally healthy populations (e.g. Unterrassner et al., 2017). While schizophrenia affects approximately 0.32% of people worldwide (World Health Organization, 2022), most estimates of the prevalence of voice-hearing experiences among general population are much higher, falling somewhere between 3% and 20% (see Cook, 2019, 7).

Importantly, there are continuities between religious cognition and disordered cognition. According to the psychologist Caroline Brett, “The relationship between spiritual and psychotic states seems to be very close, because they exhibit similar organizational features and may be caused by the same kinds of processes” (Brett, 2002, 355). Psychotic experiences and experiences of hearing God’s voice seem to be linked. Luhrmann notes that not everyone at Vineyard could hear God speak. Some had trouble experiencing God in a personal and intimate way (in our terms, they lacked EREs). Luhrmann found out that those who often had spiritual experiences were high on a psychological trait of absorption (Luhrmann et al., 2010). Absorption is a tendency to become totally immersed in experience and thought. People with proclivity for absorption agree with claims such as “I can be greatly moved by eloquent or poetic language.” and “At times I somehow feel the presence of someone who is not physically there”. There is good evidence that the proclivity for absorption drives spiritual experiences like hearing God speak (e.g., Coleman et al., 2020; Granqvist et al., 2005; Lifshitz et al., 2019). Crucially, one study also found that absorption “was positively correlated with all types of hallucinations and multiple types of delusions” and it “is a potentially important but under-researched component of psychosis” (Rosen et al., 2017).

While I am not aware of other studies linking absorption to hallucinations, there is also evidence that mentally healthy people’s experiences of hearing God’s voice and the voice-hearing experiences of psychotic individuals are underpinned by similar cognitive processes. Having strong mentalizing skills or Theory of Mind has been linked to religious belief in a personal/relational God in several studies (e.g., Schjoedt et al., 2009; Norenzayan et al., 2012; Maij et al., 2017; Willard & Norenzayan, 2013). There is also a link between overactive mentalizing and psychosis. For example, schizophrenia patients tend to overattribute intentionality to nonintentional phenomena such as random movements of shapes (e.g., Bliksted et al., 2019). Crespi and Badcock (2008) argue that autism and psychosis are two extremes of mentalizing, one of underactive and another overactive, respectively (see Langdon & Brock, 2008 for criticism). Autistic individuals with impaired mentalizing skills tend to lack personal religious beliefs such as ‘‘I have a personally meaningful relationship with God” or “I believe that God is concerned about my problems.” (Norenzayan et al., 2012). Believers who experience having a personal relationship with God are thus more likely to be closer to the psychotic end of the spectrum.

While the science continues to be debated, experiences of divine communication by healthy individuals and psychotic experiences like AVHs seem to be underpinned by similar psychological profiles and cognitive tendencies. Since most believers regard reports of divine communication by people with a psychotic disorder unreliable, we can formulate the following argument against experiences of hearing God speak.

  1. 1.

    Experiences of hearing God speak by people with a psychotic disorder are unreliable because they are probably underpinned by unreliable cognitive processes.

  2. 2.

    Experiences of hearing God speak by mentally healthy individuals and by people with a psychotic disorder are probably underpinned by similar cognitive processes.

  3. 3.

    Therefore, most experiences of hearing God speak are unreliable. (from 1. to 2.)

  4. 4.

    Unreliable experiences of God are bad reasons to believe in God.

  5. 5.

    Therefore, most experiences of hearing God speak are bad reasons to believe in God. (from 3. to 4.)

I am not arguing that all experiences of hearing God’s voice are unreliable or have no evidential force. The actual epistemic outcome of my argument for the person having the experience depends on the person’s background knowledge, on one’s model of justification, on the force of the experience, the content of the divine message, and so on. However, most experiences of hearing God speak do not seem to serve as good, independent reasons to believe in God.

Conclusion

According to the Religious Reasons Reply, most believers have good independent reasons to believe in God. Thus naturalistic theories pointing to the epistemically unreliable causes of religious belief cannot debunk their belief. I have discussed four questions regarding the reply: whether most Christian believers have good independent reasons, whether their reasons can support thick beliefs about a relational God, whether belief in God is partially caused by such reasons (or whether they are rationalizations), and whether cognitive science can undercut them.

The appearance of design and beauty in nature and EREs (experiences of divine presence, guidance, and communication) constitute popular reasons to believe in God in the US and presumably elsewhere. Design and beauty arguably serve as a good prima facie reason for believing in a transcendent and personal creator/designer. While most voice-hearing experiences are weak or ‘soft’ experiences, a justified background belief in a personal creator/designer help justify a critical trust in such experiences and the belief in a relational God. The appearance of design and EREs causally influence belief, and thus believers citing such reasons are not rationalizing.

While cognitive science does not undercut the evidential value of intuitions of design, it does seem to undercut most experiences of hearing God speak, weak or strong. Persons prone to spiritual experiences like hearing God speak have a proclivity for absorption and overactive mentalizing. Both have been associated with psychosis and hallucinating. Christians are typically wary of reports about divine communication by people who seem to suffer from a psychotic disorder. Such skepticism should be extended to most voice-hearing experiences, including one's own. Therefore, while most believers seem to have good independent reasons for believing in a creator/designer, and these reasons are not easily debunked, it seems that popular reasons for believing in a relational God can be debunked.