What the science says

Consensus

Let-s begin from establishing what a "worldview" is, because the word does a lot of work here. The anthropologist Clifford Geertz gave the classic formulation: religion is a system of symbols that establishes powerful, pervasive moods and motivations by formulating a conception of the general order of existence and clothing it with such an aura of factuality that the moods and motivations seem uniquely realistic (Geertz, 1973). The key phrase is the aura of factuality: a worldview does not feel like a belief you hold, it feels like the way things simply are, which is what makes it a frame rather than an opinion. Religion, on this view, integrates several things at once, a picture of ultimate reality, a moral order, a set of rituals, an identity and a community, and it is that integration, not any single doctrine, that gives it its grip.

‍

The cognitive science of religion asks a narrower, strictly mechanistic question: given how human minds work, why does supernatural belief come so naturally? Its answer, and this is the important discipline, is about the machinery of belief and says nothing about whether any belief is true. Three tendencies of ordinary cognition do most of the work. We detect agents and intentions very readily, so readily that we routinely sense a "who" behind ambiguous events, which primes belief in unseen agents. We run the same theory-of-mind on those agents that we run on other people, giving gods beliefs, desires and watching eyes. And our memory favours "minimally counterintuitive" concepts, ideas that are mostly ordinary but violate one expectation, a person who is invisible, a statue that hears, which are both striking enough to remember and coherent enough to pass on (Barrett, 2000; Boyer, 2001). None of this tells you whether the agents are real; it tells you why the ideas spread easily through minds like ours.

‍

How does a worldview transmit across people and generations? Not chiefly through argument. Joseph Henrich's work identifies credibility-enhancing displays, or CREDs: people adopt a belief far more readily when they watch committed believers pay real, costly, hard-to-fake prices for it, the fasting, the tithing, the celibacy, the dangerous pilgrimage (Henrich, 2009). A cost you could not sensibly pay unless you really believed is evidence that you do, and that visible sincerity, not the doctrine stated aloud, is what the next person's mind takes as reason to adopt it. This is why religions that ask more of their members often transmit more strongly, and why worldviews of every kind, including secular ones, spread through visible commitment rather than through leaflets.

Controversies

The sharpest disputes are about the grand causal stories, the claims that religion built cooperation, or morality, or civilisation itself, and here the honest verdict is that the biggest of them are unresolved and the evidence is treacherous. The best-known is the "big gods" thesis, the argument that belief in powerful, moralising, punishing gods culturally co-evolved with, and helped enable, cooperation among strangers in large societies (Norenzayan, 2013). It is a genuinely interesting idea, and genuinely contested. Its empirical fate is a cautionary tale worth telling in full: in 2019 a high-profile Nature paper reported that moralising gods appear only after big jumps in social complexity, seeming to settle the direction of causation, but in 2021 that paper was retracted, after Beheim and colleagues showed that most of the outcome data, about sixty per cent, were actually missing and had been quietly recoded as "absent", and that handling the missing data properly removed and at points reversed the finding (Beheim et al., 2021). The lesson is not that either side won; it is that quantitative claims about religion across ten thousand years of history are extraordinarily sensitive to coding decisions most readers never see, so anyone who tells you the history of gods and cooperation is settled is overselling.

‍

A second, quieter controversy surrounds religion and wellbeing. Religious involvement does correlate with better mental health, more prosocial behaviour and even longevity, and syntheses sympathetic to the field read those associations fairly positively (Koenig et al., 2012). But the causal story is much weaker than "religion improves health" makes it sound. The evidence is mostly correlational and cross-sectional, it is wide open to self-selection, and it is heavily confounded with something mundane: religious people tend to be more socially connected, so much of the apparent benefit may be the community, the routine and the support rather than the belief itself. Untangling the doctrine from the congregation is genuinely hard, and the causal arrow is far from clean.

Limitations

Several limits bound the whole field. "Religion" is not one variable: belief, belonging, practice and personal experience come apart, and measuring "religiosity" by attendance, by stated belief or by self-rated importance produces different pictures that do not interchange. Almost all the psychological work is correlational, so claims that religion causes wellbeing, or prosociality, or anything else, run well ahead of the evidence. Much of the research is drawn from Western, and often Christian-inflected, samples, which is a serious problem for a global and enormously varied phenomenon, and a reason to distrust confident universal claims. And there is a category error that snares people on all sides: the science describes the mechanisms by which belief arises and spreads, and that is simply a different question from whether any belief is true. Explaining why an idea is easy for human minds to hold is not the same as showing it false, and treating a mechanistic account as a debunking, or as an endorsement, misreads what the science can do.

Open questions

The live questions are substantial. The timing and direction of the moralising-gods-and-cooperation relationship is now genuinely open again, awaiting better data and better handling of it. How much of religion's prosocial effect comes from the supernatural belief itself versus from the ritual, the community and the credibility-enhancing displays around it remains unclear. Whether the fast-growing "spiritual but not religious" category delivers the same functions, meaning, belonging, transcendence, without the institution is an open empirical question with real social stakes. And the global trajectory is genuinely mixed: secularisation has advanced in some rich, secure societies while religious populations grow demographically elsewhere, so the once-confident story of inevitable decline does not fit the world as it is.

So what

The single most useful idea for anyone who has to communicate, persuade or design across a diverse population is that a worldview is a frame, not a set of detachable opinions, and that facts which collide with the frame tend to bounce off rather than land. So the first move is always to understand the frame you are speaking into: what it holds sacred, what it reads as a threat, who it treats as credible, and what it will accept as evidence. This is not manipulation; it is the difference between being heard and being tuned out. Two further principles follow. Some values are held as sacred, and sacred values resist trade-offs: offering money or convenience in exchange for something a person treats as sacred can offend rather than persuade, because you have proposed a swap in a currency they refuse to trade in. And credibility runs through visible commitment, so a message lands better from a messenger who is embedded in the community and visibly pays the costs of belonging than from an outsider reciting facts.

‍

For companies and brands, the practical consequence is that whole categories, food, family, finance, sex, death, carry inherited moral and often religious framings, and those framings persist even among people who are not observant, so a brand is always operating inside a moral frame it did not set. Recognise it rather than blunder into it. Treat "spiritual but not religious" consumers as a real and growing segment with genuine needs for meaning, ritual and belonging, not as a marketing costume. And be very careful with the sacred: borrowing sacred imagery or language to sell something trivial reliably backfires, because it reads as a category violation, an attempt to trade in a currency that is not for sale.

‍

For political parties and campaigns, worldview substantially predicts which evidence people will accept, so arguments that attack the frame head-on tend to entrench it, while messages that connect a proposal to values the audience already holds sacred can actually move them. The honest caution here is heavy, because this is exactly the terrain that demagogues and manipulators work: manufacturing a sacred threat, staging loyalty tests, turning a worldview into an in-group weapon against an out-group. The line between speaking respectfully into a frame and weaponising it is the line between persuasion and demagoguery, and it runs through whether you are naming values people genuinely hold or manufacturing fears to exploit.

‍

For government and public health, messages that cut against a community's worldview, on vaccination, diet, end-of-life care, reproduction, fail not because people cannot understand the facts but because the facts arrive wrapped in a frame that feels alien or threatening. The reliable route is to work through trusted, embedded messengers who visibly share and pay the costs of the community's world, rather than to broadcast decontextualised facts from an official podium, and to frame the ask in terms the worldview already honours rather than in terms that read as an attack on it.

‍

There is a serious dual-use point here, heavier than for most topics, and it is worth stating plainly. The very mechanisms that make worldviews cohere, sacred values, in-group trust, credibility through costly commitment, are the mechanisms that cults, demagogues and manipulative marketers exploit to capture people. The responsible use of this knowledge is to respect sincerely-held worldviews as the deep and legitimate frames they are, to communicate honestly within them, and to refuse to manufacture sacred threats or to weaponise identity, even when doing so would work. And one honesty point runs underneath the whole article: to explain the psychological machinery of belief is not to debunk the belief, and treating a description of the mechanism as a verdict on the truth is a mistake this field asks you not to make.

‍

The summary is that religion and spirituality, and secular worldviews alongside them, are integrated frames that tell people what is real and what matters, built on ordinary cognition, transmitted through visible commitment, and doing real work in binding morality, identity and community. The science can describe how those frames form and function while staying silent on their truth, and it should stay humble about the grand historical claims, which have a habit of collapsing under scrutiny. For anyone communicating across worldviews, the enduring lesson is to understand and respect the frame rather than argue with the facts alone, and never to mistake an account of how a belief works for a judgement on whether it is right.

‍