Why do you believe?

When I was about five years old, my grandmother learned I did not believe in God. I remember being ushered into her room, where she lay weeping, urging me to promise I would become a believer. When my deep anger, in light of this brutal attack on my self, subsided, I vividly recall a sense of wonder. A wonder of how anyone could believe something so ridiculous; my brain simply rejected the notion of a God, any God.

I don’t believe in an interventionist God
But I know, darling, that you do

Nick Cave, Into My Arms

Many years later, I am still baffled by the fact that so many accept religion and the belief in an unseen interventionist God (or Gods) as a fact regardless of the lack of evidence, not to mention the inherent contradictions between different competing religions, each claiming to have captured the one true God. How can so many conform to sets of rules given by an invisible, all-seeing, and interventionist entity without as much as a second thought?

Answers differ.

When setting out to research this, I promised myself I would remain free of prior bias. To do so, I tried to examine the validity of religious belief from many different perspectives, hoping that the amalgamation of the latter would yield an objective answer. Whether my quest was successful is for the reader to decide.

Philosophers, Psychologists, and Sociologists

Philosophers, as philosophers do, have debated the origins of religion and faith for years. From Xenophanes (~500 BCE), who claimed we humans simply make gods in our own image (and if horses could draw, their gods would be horses), to Hume (The Natural History of Religion, 1757), arguing the origin of religious beleifs are rooted in fear and hope in the face of the unknown (disease, weather, death, fortune) combined with a universal tendency in mankind to conceive all beings as being like themselves, going on Feuerbach (The Essence of Christianity, 1841) describing God as our capacities for love, wisdom, will, idealized to infinity, then externalized as an entity standing over us. Then came Freud (The Future of an Illusion, 1927), for whom religion is wish-fulfillment. Confronted with the terrors of nature and our own helplessness, we resurrect the protective father of childhood as a cosmic one, an exalted father-figure.

It would seem philosophers and early psychology see us all as frightened children, looking for a father figure.

A second school of thought claims religion originates not in individual projection but in what it does for the group. The Sisyphus fragment attributed to Critias (5th century BCE) contains the first social-control theory, namely, a clever man invented the gods so that people would behave when no one was watching. Religion is nothing but a surveillance mechanism, an internal policeman of sorts. Durkheim gave this its own take (The Elementary Forms of Religious Life, 1912), claiming that when a group gathers in ritual and feels a power larger than any one individual (what he called collective effervescence), that force is real, but it’s the social group experiencing its own power; misattributed to a totem or a God.

Nietzsche begs to differ. In On the Genealogy of Morality (1887), he does not ask what religion is; rather traces the hidden power-relations from which its values stemmed. Christian morality, on his account, originates in ressentiment: the “slave revolt in morality,” in which the powerless invert the aristocratic value scheme, recoding their weakness as virtue and their masters’ strength as sin. The origin of religion, ala Nietzsche, is a concealed psychological victory of the weak over the strong.

But those are all debunking theories, claims against the mystical, or better put, spiritual, nature of religion.

Not everybody agrees.

William James (The Varieties of Religious Experience, 1902) points out that you’d have to apply the same acid to science and philosophy. James locates religion’s root in first-hand religious experience, the individual’s sense of contact with something “More”, while Calvin called it the sensus divinitatis, an innate sense of the divine; Alvin Plantinga has revived this in contemporary epistemology, arguing that belief in God can be “properly basic”, produced by a reliable cognitive faculty for perceiving God the way perception produces belief in kitchen tables (more on that when we delve into the neurological subsystem underlying belief).

Preplexing? Sure. It seems the same agency-detection system that Hume’s heirs cite as debunking, a theist can cite as the God-given means by which we come to know God, so nobody wins.

Economists

Pre-modern populations faced ruinous idiosyncratic risk with no insurance market, no courts, and no safe way to trade. Iannaccone’s foundational survey (Journal of Economic Literature, 1998) frames religion as a production good; Chen (JPE, 2010) showed Indonesian households turning to Islamic institutions as ex-post consumption insurance precisely where credit collapsed after 1998, and Bentzen (Economic Journal, 2019) found religiosity rises after earthquakes in districts lacking formal protection. Greif’s work on the eleventh-century Maghribi traders (AER, 1993) treats a religious-ethnic coalition as a private contract-enforcement technology, with excommunication as the penalty, while Henrich (Science, 2010) found fairness toward strangers scaling with market integration and world-religion membership.

Is religion simply an insurance mechanism? A congregation providing in times of need? If so, how is it managed so all members contribute while free riders are eliminated or at least minimized?

The answer is: participation needs to hurt.

Iannaccone’s Sacrifice and Stigma (JPE, 1992) models congregations as club-goods producers whose output depends on average member commitment; dietary laws, dress codes, and Sabbath restrictions tax outside options, so only high-commitment types join, raising the club good for everyone. Berman (QJE, 2000) applied this to ultra-Orthodox Jewish communities, where extreme yeshiva study duration serves as the commitment signal underwriting a dense mutual-aid network. You want to benefit from a congregational insurance policy? Tax yourself.

But if religion “sells” insurance, welfare and enforcement, public provision should crowd it out. Gill and Lundsgaarde (Rationality and Society, 2004) make an interesting observation. Welfare-state spending is inversely related to religious participation cross-nationally. Seabright’s (The Divine Economy, 2024) extends the supply-side view to the present, treating religious organizations as platforms competing for attention and revenue.

This quid pro quo trick was not lost on nation states. Those with weak fiscal and legal capacity substituted religious authority for enforcement. Johnson and Koyama (Persecution and Toleration, 2019) argue that conditional toleration and religious law were low-cost governance tools that receded as state capacity grew, while Rubin (Rulers, Religion, and Riches, 2017) shows how rulers leaned on religious legitimation, rendering more veto power to clerics. Temples were themselves fiscal organs (much of the earliest writing from Uruk is temple accounting), and temples ran granaries, credit, and weights standards.

Can all be played down to economy? I think not. Let’s dig deeper.

Evolutionists

There’s broad agreement that religion has some evolutionary basis, but a genuine, unresolved debate over what kind. The debate thunders around the following: is religion an adaptation, a byproduct, or is it all about group natural selection?

The dominant view is that religion wasn’t selected for anything. It’s a byproduct, a spandrel of cognitive systems that evolved for other reasons and happen, as a side effect, to make supernatural belief easy to acquire and hard to shake. In some ways, it is the direct descendant of Hume’s projection theory in biological evolution terms.

The leading evolutionary theory is Hypersensitive Agency Detection (HADD). A system that evolved to detect predators with a heavy false-positive bias. Mistaking wind for a leopard costs you a jolt of adrenaline, while mistaking a leopard for wind costs you your life. As a result, we over-detect agency everywhere – and that same trigger-happy system, employed over weather, illness, and fortune, populates the world with invisible agents. Again, this is Hume’s “tendency to conceive all beings as like themselves,” with an evolutionary rationale.

Children are “intuitive theists” who see purpose and design everywhere by default, which is why they are so much more susceptible to religious faith, and once imprinted by it, find it tremendously difficult to shake. Boyer added the observation that minimally counterintuitive concepts, e.g., a being that’s like a person but invisible, or immortal, is maximally memorable and transmissible, which is why gods have the specific cognitive shape they do. In layman’s terms, make it supernatural so it is extremely hard to disprove.

The competing camp says the byproduct story can’t explain why religion is so costly. People fast, tithe, stay celibate, sacrifice animals, undergo painful initiations, sometimes die for their gods. Natural selection doesn’t usually tolerate that much waste unless it’s buying something. So maybe religion itself was naturally selected because it did some important adaptive work.

The leading candidate is cooperation. Ara Norenzayan (Big Gods, 2013) argues that “moralizing high gods”, deities who watch, judge, and punish, solved the problem of large-scale cooperation among strangers. A believed-in supernatural monitor makes people behave when no human is watching (the Critias/Sisyphus intuition from the philosophy answer, now as an evolutionary mechanism). This in itself lets societies scale beyond the face-to-face groups where order is rooted in reputation.

Then there’s costly signaling (Richard Sosis’s work), capitalizing on the very wastefulness that puzzles the byproduct theorists. Hard-to-fake, expensive displays reliably signal commitment to the group, screening out free-riders, viewed from the group’s side rather than the learner’s. Sosis found religious communes historically outlasted secular ones, and the more costly demands a commune made, the longer it survived. This is substantiated by the fact that, historically, groups with cohesive, cooperation-enforcing religions outcompeted groups without them, so religion spread by between-group selection. This is Durkheim’s “religion is society worshipping itself” recast as multilevel selection theory.

As appealing as this might sound, group selection in evolutionary biology is contentious at most, where many theorists think it reduces to gene-level or cultural selection.

The most defensible position is probably a hybrid: a cognitive byproduct gave rise to the capacity for supernatural belief, and then cultural (and possibly genetic) evolution co-opted it for group functions. So yes, there’s an evolutionary basis, in the sense that our capacity for religion clearly emerges from evolved cognition. What that basis is – side effect, adaptation, or both – remains an open question, to date.

Psychatrists

While philosophers, psychologists, and evolutionists bicker about the validity and underlying origin of religious belief, it would seem that psychiatry may have a better view into the religious mind. After all, it’s all in your head.

Is religion nothing but a (very common) mental disorder?

Psychiatrists say: no, religious belief is not a mental disorder. This is actually a settled question in psychiatry and psychology. The DSM-5 (the standard diagnostic manual) explicitly distinguishes between culturally or religiously sanctioned beliefs and pathological ones. A belief becomes clinically relevant only when it’s idiosyncratic (not shared by the person’s community), causes significant distress or impairment in functioning, or is accompanied by other symptoms such as hallucinations, disorganized thinking, or loss of contact with everyday reality (which you must admit is kind of funny, as religious belief claims to transcend reality).

The key distinctions clinicians use look something like this: billions of people share religious beliefs within organized traditions, they generally function well in work and relationships, and they can typically hold their beliefs alongside ordinary practical reasoning. Moreover, a delusion isn’t defined simply as a false or unfounded belief. The DSM-5 defines it as a fixed belief not amenable to change in light of conflicting evidence, and then adds the crucial qualifier: the belief is not one ordinarily accepted by other members of the person’s culture or subculture. That clause is doing enormous work. It means that “God exists,” “the Eucharist is the body of Christ,” or “the dead are reborn according to karma” cannot be delusions, no matter how unfalsifiable, because they’re consensually held within a community. The diagnostic criterion isn’t the content of the belief or its evidential support; it’s whether the person is an outlier relative to their own group.

Almost every DSM diagnosis requires that the symptoms cause clinically significant distress or impairment in social, occupational, or other important areas of functioning. This is a second gate that religion routinely doesn’t trip. Ordinary religious belief and practice generally improve functioning or leave it intact – people work, love, and parent normally, and often report the belief as a source of meaning rather than suffering. Devotion that structures someone’s life productively is the opposite of the impairment the criterion is looking for.

Even inside the psychotic disorders, the text keeps flagging culture. The hallucination criterion notes that hallucinations occurring while falling asleep or waking, and – critically – those that are a normal part of religious experience in certain cultural contexts, are not to be counted as symptoms. A visionary experience during a vigil, a voice heard in a culturally recognized religious frame, a trance state within a tradition that expects it – these are explicitly excluded from the symptom count. The manual is telling clinicians: a sanctioned religious experience is not a hallucination, even when phenomenologically it might resemble one.

There’s a non-disorder code specifically for religion.

So, following the DSM-5, it’s ok if it’s a common delusion (as Richard Dawkins framed it), held by many. Basically, you could be mad as a hatter as long as you go to work and pay your bills. Personally, I find that absurd; millions of people have accepted Nazi beliefs, and even more have taken on Communism as the theory to live by, none of which are considered to be acceptable solely on the premise of “that was a large group thing”. The fact that many believe something is a very weak argument, as many have strongly believed the Earth was flat and the center of the universe. I suppose the ancient psychiatrist could claim that believing in a flat Earth was not a mental disorder, as it was, well, common. Could it be that religious psychiatrists just gave themselves an escape clause?

And, maybe there’s another system at work.

A study (Zuckerman, Silberman & Hall, 2013), corroborated by many follow-ups, has found intelligence is an inverse predictor of religiosity. Simply put, the more intelligent you are, the less likely you are to possess religious belief. This is very much in line with the earlier debunking theories, as the less intelligent and informed you are, the more likely you are to attribute agency to phenomena beyond your mental capacity, which in turn makes you see the world as dangerous and inexplicable, leading to the need to cling to some all-knowing interventionist figure.

As robust as the correlations in the above research are, one has to admit that it is still lacking in the way it measured intelligence (i.e., via proxies, such as GPA, SAT/ACT scores, and not via direct IQ tests, which biases the sample towards education), and that it is heavily sampled from US protestant community, which in itself represent a single facet of religiosity.

A meta-analysis of 63 studies showed a significant negative association between intelligence and religiosity. The association was stronger for college students and the general population than for participants younger than college age; it was also stronger for religious beliefs than religious behavior.

The relation between intelligence and religiosity: a meta-analysis and some proposed explanations, Miron Zuckerman , Jordan Silberman, Judith A Hall, 2013

Neurologists

Neurotheology, a field that studies how the brain produces religious experience, introduced several interesting case studies.

A landmark study (Cognitive and neural foundations of religious belief, Kapogiannis, 2009) proposed that religious belief rests on several cognitive dimensions mapped to specific brain regions.

Harris (The Neural Correlates of Religious and Nonreligious Belief, 2009) scanned 15 Christians and 15 nonbelievers evaluating true/false statements. While somewhat questionable for various methodological reasons (which we will not cover for lack of space), this research provides a glance into the brain regions supposedly associated with religious belief.

The main finding was that while religious and nonreligious thinking engaged somewhat different regions, the core difference between belief and disbelief was content-independent – belief itself was associated with the ventromedial prefrontal cortex regardless of whether the content was religious. In other words, believing “God exists” and believing “2+2=4” used the same machinery. Despite believers and nonbelievers accepting and rejecting diametrically opposite statements in half the trials, the same neural systems were engaged in both groups throughout, which the researchers argue strongly supports the content-independence of belief and disbelief as cognitive processes.

Schjødt (Talking to God: Highly religious participants recruit areas of social cognition in personal prayer, 2009) found that personal prayer activates a network of brain regions involved in social cognition – the temporopolar region, temporoparietal junction, precuneus, and medial prefrontal cortex – the same regions used when communicating with another person, because the believer treats God as an intentional agent. A related study found that when believers thought a prayer came from a charismatic authority, they showed reduced activation of the frontal executive network, including the dorsolateral prefrontal cortex and anterior cingulate cortex.

This raises the question: even if the belief network in the brain is the same for all beliefs, religious or otherwise, why are some people more prone to believing in God, while others perceive it as a childish fantasy? This may suggest another network is in play, one that is yet to be uncovered. A set of networks either bolted on top, upstream, or somewhere else in the belief machinery that may explain this dichotomy between “believer brains” and those who are not.

The Kapogiannis follow-up (2014) to the 2009 paper we mentioned earlier used Granger causality to ask where religiosity differences live, and found it wasn’t in which regions activate but in the causal flow between networks, with theory-of-mind (ToM) regions driving downstream regions (those who are involved in the process of understanding someone else’s intentions), and this effect held true only for religious individuals. So the variance shows up in the routing, not the destination.

Or, is it simply the result of early-age indoctrination seeding the notion of a God (basically setting the neural network weights to lean towards such a concept) in the brains of those who would later become believers?

It does seem that a neurological foundation to common religious belief is at best shaky; however, neuroscience as well as psychiatric studies often benefit from researching pathological cases rather than the everyday ones, as such may reveal underlying mechanisms one cannot probe within normal human brains.

Fundamentalists and religious zealots take the notion of (their) God to extremes. They would be willing to kill, maim, and torture in his name without a second thought. Research published in the Proceedings of the National Academy of Sciences used lesion network mapping in patients with traumatic brain injury. Scientists identified a specific functional brain network, lateralized to the right hemisphere, which, when damaged or disrupted, is linked to higher levels of religious fundamentalism.

A functional MRI (fMRI) study by researchers at Harvard found that belief—whether religious or nonreligious—activates the ventromedial prefrontal cortex (vMPFC). However, religious thinking uniquely activated brain regions associated with the governance of emotion and cognitive conflict, whereas nonreligious belief relied more on memory retrieval networks (which is chiming on eariler research suggesting religious people are more inclined to rely on emotion rather than cognition). Another related study from the University of Utah published in Social Neuroscience used fMRI on subjects undergoing intense spiritual feelings. They found that religious experiences heavily activate the nucleus accumbens, a critical brain region for reward processing, as well as the frontal cortex (one must note, though, that other studies have shown that non-religious people find it rewarding to dismiss religious belief).

Hyperreligiosity: An overwhelming obsession with moral, theological, and philosophical ideas

It was also found that hyperreligiosity is deeply rooted in the temporal lobes and limbic system. This connection is most famously observed in patients with Temporal Lobe Epilepsy (TLE) and other neurological conditions that impact the frontal and temporal networks, including frontotemporal lobar degeneration, schizophrenia, and obsessive-compulsive disorder. In chronic cases of TLE, interictal hyperreligiosity often groups with other behavioral traits to form a distinct clinical profile known as Geschwind Syndrome (or Gastaut-Geschwind Syndrome).

This syndrome is defined by five primary characteristics, namely hyperreligiosity (described above); hypergraphia: A compulsive need to write excessively; hyposexuality: A markedly decreased interest in sexual activity, shifting physical drives toward cognitive and spiritual preoccupations; viscosity: Where conversational interactions become lengthy and rigid; and an intensified mental life: Deepened emotional reactions, a strong moral compass, and an heightened sense of personal destiny.

Sadly, as telling as these disorders are, they provide little concrete data pointing to the origin of common religious belief. Yes, some people are fanatics, some specific brain regions fire up when they manifest the zeal, but nothing robustly maps back to a specific brain sub-system or systems that underlie religion.

I beg to differ.

The Source

I’d like to suggest a bird’s-eye view theory encompassing all (or at least most) of the above, revolving around (the control of) anxiety.

Religion is not one adaptation but a culturally evolved stack of anxiety-management technologies, each layer solving the anxiety problem created by the layer of life beneath it. It persists because it lowered the chronic cost of uncertainty at every scale — body, mind, household, tribe, civilization — and it recedes wherever a cheaper anxiolytic substitute appears.

The predictive brain that made humans successful is a simulation engine, and simulation cuts both ways. The same machinery that plans a hunt can preview death, famine, betrayal, and every branch of the counterfactual tree. Foresight is the smart ape’s tax: humans became the only animal that is anxious about things that are not happening. Religion is the cultural answer to a biologically novel problem — unbounded, abstract, future-directed anxiety that no fight-or-flight response can discharge.

Hypersensitive agency detection (HAAD) converts random events into intentional acts. This looks like a bug; it’s a feature for anxiety regulation. You cannot negotiate with a drought, but you can negotiate with a rain god. Attributing misfortune to agents transforms unbounded uncertainty into a relationship — something with a protocol, a grievance procedure, terms. Anxiety over the unknown becomes anxiety with an interface.

Anxiety is physiologically an action-preparation state; it demands discharge through doing. Ritual is action with guaranteed executability: always available, always completable, immune to failure. Ritual converts free-floating dread into a finite procedural task with a completion state.

Daily life is a stream of micro-decisions. Religious law pre-computes them: what to eat, when to rest, whom to marry, how to bury. The decisions are made once, by an unimpeachable authority, forever. The believer trades autonomy for a massive reduction in per-decision cognitive and emotional load — and the trade is rational when computation is expensive, and stakes are opaque.

Scaling beyond kin groups created a new anxiety: the anonymous other. Two mechanisms solve it. Watchful, punishing gods make defection risky even when no human is looking, letting strangers cooperate. And costly signaling – visible, expensive, hard-to-fake piety – functions as a trust credential. Your observance is your deposit; the effort you sink (time, ritual, dietary restriction, tithe – your “tax”) is exactly what makes you legible and predictable to strangers. This layer doesn’t reduce the individual’s anxiety about the world; it reduces everyone’s anxiety about each other, which is what makes markets, cities, and armies possible.

At the top of the stack sits the one anxiety no action can discharge. Terror management: literal immortality (afterlife) and symbolic immortality (the eternal people, the tradition that outlives you) cap the stack. Empirically, mortality salience reliably increases religious commitment. Should you want to be routed through the gates of heaven rather than hell, pay your dues.

Religion doubles as mutual-aid infrastructure: the congregation is an insurance pool. Religiosity tracks existential insecurity. Where states provide welfare, healthcare, and rule of law, religion declines (and vice versa), not because people became smarter, but because the anxiolytic function was outsourced to institutions with better actuarial tables.

Every one of the above layers does the same thing: variance compression. Religion is a cultural prior that shrinks the world’s perceived entropy, over nature, over one’s own agency, over daily choice, over other minds, over death, over material fate. In predictive-processing terms, it’s a prediction-error minimization package: it doesn’t make the world more controllable, more modeled, and a modeled threat is a less intimidating one.

One question remains. As religion also manufactures anxiety sin, hell, impurity- how do those play in the above scheme? Have we stumbled on a counterexample? The answer is no.

Religion’s embedded anxiety pain points are its business model. The system induces anxiety in a legible, solvable form (you know exactly what you did wrong and exactly how to fix it), then sells the resolution. Converting diffuse existential dread into confessable, expiable, ritual-priced anxiety is itself the service. A protection racket, yes — but one that genuinely protects against the raw, unpriced dread it displaced.

Die Religion ist der Seufzer der bedrängten Kreatur, das Gemüth einer herzlosen Welt, wie sie der Geist geistloser Zustände ist. Sie ist das Opium des Volks

Karl Marx, Zur Kritik der Hegelschen Rechtsphilosophie, 1843

Religion, and religious belief. A genius brainchild of the human race, that as beautifully depicted in The Rocky Horror Picture Show ending scene, is “crawling on the planet’s face, some insects, called the human race… lost in time… and lost in space… and meaning … less”. But as we grow, as individuals, as a race, as we mature, it may be wise to accept the fact that we are alone, fatherless, and manage the rising anxiety, rather than resort to fantasy, to that opiatic experience that calms us down, but robs us of parts of our humanity.