The foxhole argument and what it can show
The saying "there are no atheists in foxholes" is usually meant to do more than make a claim about wartime psychology. In apologetic use, it suggests that mortal danger strips away intellectual postures and reveals what human beings already know at a deeper level: that they are dependent on God. The cry for God in a crashing aircraft, a hospital corridor, or under shellfire is taken to be a disclosure of latent belief. Fear, on this reading, does not create faith; it uncovers it.
The atheist reply considered here is that this argument fails both empirically and epistemically. Empirically, the claim is too simple. People under threat do not all move toward religion. Often they intensify the commitments they already have. Believers may become more explicitly religious; non-believers may become more explicitly non-religious. Epistemically, even if some atheists do call out to God in extremis, a desperate utterance is not the same thing as evidence that God exists. Panic can explain the behavior without vindicating the belief-content expressed in the panic.
That reply is substantially correct. But it is important to state it carefully. It does not show that theism is false. It does not show that religious belief is always a fear-response. It does not show that death anxiety explains away all religion. It shows something narrower: the foxhole argument is a weak argument for theism, and the psychology of crisis cannot carry the evidential weight sometimes placed upon it.
The argument in its strongest form
The strongest version of the foxhole argument is not the crude claim that every atheist literally prays before death. That is plainly false, and serious defenders need not rely on it. The better version says that human beings have a deep practical orientation toward transcendence, dependence, and accountability. Under ordinary conditions, people can suppress, reinterpret, or intellectually deny that orientation. Under mortal threat, however, self-protective abstractions fall away. What remains is not an argument but a disclosure: the person reaches beyond the finite because he knows, at some level, that the finite is not enough.
This line of thought has affinities with several serious positions, though it should not be credited to them as if they endorsed the foxhole slogan. William James treated religious experience as philosophically important data, while also denying that another person's experience functions as a coercive public proof. Alvin Plantinga's Reformed epistemology argues that belief in God may be properly basic under appropriate conditions rather than inferred from an argument; it is not a claim that panic-prayer proves God to outsiders. In cognitive science of religion, scholars such as Justin Barrett, Pascal Boyer, and Deborah Kelemen have examined ordinary human tendencies toward agency-detection, memorable supernatural-agent concepts, and intuitive teleology. On such accounts, crisis might activate latent cognitive and existential structures. But that claim is compatible with both theistic and naturalistic explanations, so it needs careful handling.
Put charitably, then, the apologist is not saying that fear is a syllogism. He is saying that fear can reveal what a person is already disposed to acknowledge when ordinary distractions collapse. The atheist response must address that stronger claim, not merely mock the slogan.
The empirical problem
The empirical problem is that the evidence does not support a universal or near-universal conversion-under-terror pattern. Terror management theory, associated especially with Jeff Greenberg, Sheldon Solomon, and Tom Pyszczynski, is usually understood as claiming that mortality salience can lead people to defend the cultural worldviews and sources of self-worth that give their lives meaning. The literature is large and influential, but also contested: effect sizes, boundary conditions, and some replications have been debated within social psychology. So the cautious conclusion is not "death makes everyone religious." It is that reminders of death sometimes intensify already significant meaning-systems, and those systems may be religious, secular, national, moral, political, familial, or philosophical.
Jonathan Jong and colleagues have examined "foxhole atheism" with both explicit and implicit measures. Their findings are more complicated than the slogan. In some studies, mortality salience was associated with stronger explicit religious belief among religious participants and stronger explicit disbelief among non-religious participants, while particular implicit tasks suggested that some non-believers retained or displayed intuitive supernatural-agent associations under death-related prompts. That is interesting, but it is not the same as showing that atheists "really believe in God" in the same sense that theists do. Implicit associations can reflect culture, language, salience, emotional learning, or cognitive readiness rather than settled assent. Human cognition is layered: a person can have culturally acquired associations, emotional reflexes, implicit imagery, and explicit beliefs that do not all line up neatly.
The same caution applies to real crises. Studies of religious coping in illness, bereavement, and parental distress show that many religious people pray more intensely in crisis. That is unsurprising. Prayer is already part of their coping repertoire. But research on non-religion and secularity, including work by scholars such as Phil Zuckerman and Lois Lee, has repeatedly shown that non-religious people also have stable identities, moral communities, rituals, and coping strategies. They may turn to family, medicine, stoic acceptance, political meaning, human solidarity, or simply endurance. Crisis does not automatically produce prayer.
The point is not that no atheist has ever prayed under terror. Some have. The point is that this behavior is not uniform enough, clean enough, or psychologically unambiguous enough to function as evidence for theism. If the claim is empirical, the empirical world is messier than the slogan allows.
The epistemic problem
Even if the empirical claim were stronger, the epistemic problem would remain. A cry under mortal terror is not normally a considered belief-forming process. It is a behavior produced under conditions of extreme stress, uncertainty, and narrowed attention. Desperation changes the threshold for action. A person who does not believe a door is unlocked may still shake the handle when trapped in a burning room. That does not show he believes it is unlocked; it shows that the cost of trying has fallen to nearly zero compared with the cost of not trying.
This is why "trying prayer" in extremis is not equivalent to affirming theism. In ordinary life, people sometimes say, "I know this probably will not work, but I will try anything." That sentence is not irrational. It is a practical response to catastrophic stakes. If prayer is psychologically available because of upbringing, language, films, funerals, public rituals, or family memory, it may emerge as one more desperate act. The behavior is intelligible without treating it as evidence that the proposition "God exists" has become more probable.
There is also a problem of religious underdetermination. In crises, people call upon the gods, saints, ancestors, spirits, or sacred powers available within their cultures. A Christian may call upon Jesus, a Hindu upon Krishna, a Muslim upon God using the name Allah, a traditional practitioner upon ancestral powers, and a secular person may cry out using inherited religious language without intending a doctrinal confession. These appeals are not all doctrinally equivalent. If crisis-invocation were treated as direct evidence for the object invoked, it would support too many incompatible conclusions. The fact that human beings reach for culturally available ultimate help cannot by itself identify which ultimate reality, if any, is real.
A natural psychological explanation is therefore available and often powerful: under threat, people reach for the deepest symbols of rescue, agency, protection, and meaning available to them. Those symbols are usually learned socially and emotionally before they are assessed philosophically. This helps explain why many religious people pray, why some non-religious people may reflexively use religious language, why others do not, and why the content of the appeal often tracks culture. This explanation is not a disproof of God; a theist could still argue that natural dispositions are part of how humans are oriented toward the divine. The narrower point is evidential: the behavior by itself does not significantly discriminate between theism and non-theistic psychological accounts.
The strongest apologetic reply
The best reply from the apologist is that this critique may assume too narrow a model of evidence. Not all evidence is detached, calm, and laboratory-like. Some truths are recognized most vividly under pressure. Danger can clarify. A soldier may discover courage he did not know he had; a parent may discover the depth of love for a child; a dying person may finally face moral facts previously avoided. Why could a cry to God not be like that: not a panic error, but a moment in which self-deception fails?
The apologist can also object that a natural psychological explanation does not automatically debunk the belief. There are natural explanations for hunger, perception, memory, and moral emotion. We do not infer that food, trees, the past, or injustice are unreal merely because we can describe the cognitive mechanisms by which we become aware of them. If humans have a natural disposition to seek God under conditions of dependence, that could be exactly what one would expect if God exists.
These replies have force. It is too quick to say, "This belief has a psychological cause, therefore it is false." That is the genetic fallacy. It is also too quick to assume that emotionally intense circumstances are always epistemically worthless. Some emotions disclose reality; fear can alert us to danger, guilt to wrongdoing, grief to loss.
But the reply does not rescue the foxhole argument as public evidence for theism. The question is not whether crisis could ever disclose truth. The question is whether crisis-invocation reliably distinguishes truth from conditioning, fear, hope, and cultural reflex. It does not. The same psychological pattern supports incompatible religious contents. It appears unevenly across people. It is strongest where prior cultural materials are already present. It often produces action without settled belief. And it lacks an independent success condition: the plane may still crash, the shell may still fall, the child may still die. The cry may be humanly profound, but profundity is not the same as proof.
What the critique establishes
The atheist argument succeeds against the popular apologetic slogan. There are atheists in foxholes. More importantly, even where there are prayers in foxholes, the fact of prayer does not establish the truth of theism. Mortality salience tends to intensify existing worldviews rather than erase them. Desperation can generate attempts that do not express considered belief. Cultural conditioning can explain why certain words and images surface under terror. And incompatible crisis-prayers cannot all be straightforward confirmations of the specific gods invoked.
But the critique should not overstate itself. It does not prove that God is unreal. It does not prove that all religion is reducible to fear. It does not show that human beings lack any deep religious disposition. It shows that one proposed piece of evidence is weak. The foxhole cry may be psychologically revealing, existentially moving, and worthy of compassion. As an argument for theism, however, it carries little weight.