The idea that kids accept what they are told uncritically, including religious beliefs, is intuitively satisfying. But then there's the research.

Reading Time: 13 minutes

The vast majority of people who ever lived have believed in the existence of one god or another. What are the evolutionary forces that made our species—and our species alone—a religious one?

Evolutionary thinkers have spilled much ink trying to answer questions like this. Some answers have proven to be much more successful than others. This article considers, and ultimately rejects, one account prominently given by the biologist Richard Dawkins for about 15 years now: that religion might ultimately be rooted in childhood gullibility. 

Millions of people have likely read Dawkins’s speculative story about religion (recounted below) and–given his stature as a public intellectual, science advocate, and outspoken skeptic–taken it at face value. It’s well worth taking Dawkins’s childhood gullibility account seriously and asking whether it fits well with our science. 

As we’ll see, the childhood gullibility model never really held together theoretically and is directly contradicted by our best current science on how children process information–including evidence that was readily available when Dawkins’s proposal first emerged.

Kids believe the darndest things?

In The God Delusion, Richard Dawkins offers a brief review of some then-recent research on what makes religion tick. He indicates a preference for theoretical models that posit that religious cognition is just an evolutionary byproduct of how our mental adaptations work together, rather than an adaptation in its own right. 

There’s much to like about this! After all, various flavors of byproduct thinking have met fairly widespread acceptance. Then Dawkins offers his own speculative evolutionary story of religion. He suggests that one key to religion might be a human psychological adaptation that makes us basically believe whatever our parents tell us:

More than any other species, we survive by the accumulated experience of previous generations, and that experience needs to be passed along to children….there will be a selective advantage to child brains that possess the rule of thumb: believe, without question, whatever your grown-ups tell you. [emphasis added]

To Dawkins, this leaves kids susceptible to religious indoctrination:

The flip side of trusting obedience is a slavish gullibility. The inevitable by-product is vulnerability to infection by mind-viruses.

Dawkins is appropriately cautious about the tentative nature of his conjectures in The God Delusion–he’s quite up front that he’s spitballing here. But some 15 years later in Outgrowing God, Dawkins again repeats his pet theory with perhaps more confidence: 

Our earliest ancestors lived in a dangerous place…There were pythons and leopards lurking in the trees, lions behind bushes, crocodiles in the river. Adults knew of these dangers, but children needed to be told….And natural selection would have favoured genes that built into child brains a tendency to believe their parents.

Gullibility is again key to Dawkins’s story of how religion works:

If adults ever gave bad advice alongside good advice, the child brain would have no way to distinguish bad advice from good….So if, for some reason, a parent were to give a child useless advice – like ‘you have to pray five times a day’ – the child would have no way of knowing that it was useless. Natural selection simply builds into the child brain the rule ‘Believe whatever your parents tell you.’ And the rule will come into force even when ‘what your parents tell you’ is actually silly or untrue. [emphasis added]

First, let’s acknowledge some strengths of the Dawkins childhood gullibility story.

The argument is straightforward and intuitively compelling. Kids really do need to learn how to get along in the world. Some threats really are too serious to rely on trial-and-error learning. Kids who were skeptical of their parents’ claims that there are jaguars in that bit of forest did not leave many descendants…they became lunch! So maybe it’s a good idea to just believe whatever your parents tell you. Maybe our species evolved in such a way that we are generally gullible as kids, at least when it comes to stuff our elders tell us.

Broad-spectrum childhood gullibility of the sort posited by Dawkins seemingly accords well with some other observations about religion. Kids usually grow up to have more-or-less the same religious beliefs as their parents. Baptist parents tend to have Baptist kids; Mormon parents, Mormon kids; Sunni parents, Sunni kids; so on.

On these two counts—intuitive plausibility and a general fit to observations about religion—the childhood gullibility model of religion appears okay. But let’s take these ideas seriously. Seriously enough to ask whether the childhood credulity account of religion is consistent with our best theories of how culture works, or with our best evidence from developmental psychology on how kids learn and think about religions.

Theoretical challenges

As intuitively satisfying as the childhood gullibility story seems, it doesn’t really hold together at a basic theoretical level. Over the past 40 years, the science of cultural evolution has rapidly matured. We’ve come to learn a great deal about how cultural information gets passed from person to person, and the specific learning strategies that we use to glean useful information from our fellow humans. This work—summarized in excellent books like Joe Henrich’s The Secret of Our Success and Pete Richerson and Rob Boyd’s Not By Genes Alone—makes a compelling case that our species has become so phenomenally successful largely because we are good at learning from each other. We pick up techniques and technologies from our elders, tweak them with our own innovations, and in turn pass them along to others. This capacity for cumulative culture is what let our species colonize the entire globe, from our humble origins as a peculiar African primate.

As intuitively satisfying as the childhood gullibility story seems, it doesn’t hold together at a basic theoretical level.

As Dawkins correctly notes, we do rely on each other for a substantial amount of our know-how. And we really do need to learn a lot from our caregivers—relative to the young of most species, kids are pretty useless and burdensome. But Dawkins is wrong to assert that at our cores, humans evolved to be gullible and credulous. We do not passively accept the information that others try to feed us. Quite the opposite: we need to be wary and epistemically vigilant against exploitation. Indeed, one of the more interesting puzzles in cultural evolution is figuring out what information learners need to believe—can we trust that our heroes and teachers truly believe what they’re telling us? In this regard, gullibility and credulity are self-defeating. Dawkins points out that there are lots of survival threats in our environment—crocodiles in the river, jaguars in the trees. But one of the biggest threats we face is each other. A gullible learner leaves themselves open to exploitation by unscrupulous teachers. Picture here a guru who tells their acolytes that the key to enlightenment and salvation is only available to those who renounce their worldly possession (conveniently, the guru offers to dispose of the possessions). 

Muslim man teaching his son to pray in sujud (Shutterstock)

Someone who believes whatever they’re told, even by a parent or elder, isn’t a survivor who avoids predators—they’re a sucker who can be manipulated and would quickly be outcompeted by more epistemically-vigilant peers. And like all interesting evolutionary puzzles, it comes down to tradeoffs: we need to learn, but we also need to avoid exploitation. What’s the right balance? This is such a thorny theoretical issue that Maciej Chudek, a graduate school contemporary of mine at the University of British Columbia, dubbed it the “evil teacher problem,” and it remains a central challenge in the cultural evolution literature.

Cultural evolutionary theory relies heavily on building mathematical models of selective learning —who do we learn from, under which conditions? In these models, gullible agents don’t usually prosper. An honest perusal of this body of theory gives no hints that we’d expect our species (or any other!) to use blind gullibility as its general learning strategy. Instead, we find ample theory and evidence suggesting that we are keenly strategic in our cultural learning, relying on a wide variety of specific learning strategies as situations demand. We pick and choose between different sources of information—we don’t just believe anything from anyone. We also stress-test what we’re told, and preferentially adopt beliefs that are backed by credible behaviors attesting to their value. Relying on these credibility-enhancing displays (in the terminology of Joe Henrich, the Harvard human evolutionary biologist who first developed models of this sort of learning) are a key part of our toolkit that helps cultural learners pick up on useful information while avoiding exploitation. No rigorous theoretical model that I’m aware of supports the plausibility of our species having evolved to just believe whatever our parents tell us. But a fairly vast body of theory tells us that broad-spectrum gullibility is an incredibly unlikely evolutionary outcome.

Someone who believes whatever they’re told, even by a parent or elder, isn’t a survivor who avoids predators—they’re a sucker who can be manipulated and would quickly be outcompeted by more epistemically-vigilant peers. 

The reasons that broad-spectrum gullibility fails as an evolutionary solution are somewhat analogous to another failed evolutionary idea that Dawkins ably skewered in The Selfish Gene. Naïve notions of group selection—the idea that some animals will make costly sacrifices “for the good of the species” was once popular in biology, but fell apart once biologists in the 60s and 70s started to really model the details. People rigorously worked out the implications of “for the good of the species” style group selection. And key to the dismantling of naïve group selection is the notion that uncritical cooperators will pretty much always be outcompeted by individuals who exploit them. Someone who always helps others is at a huge competitive disadvantage relative to the moochers and the cheats. 

Truly, Dawkins’s rhetorical takedown of shoddy group selectionist thinking in The Selfish Gene is masterful. This makes it all the more puzzling that his pet theory of religion relies on a cultural evolutionary analog of unconditional cooperation: unconditional childhood gullibility. Both cases (a group selectionist universal cooperator and a generally gullible cultural learner) posit an evolved psychology that is incredibly vulnerable to exploitation, and we shouldn’t expect either type to have ever really become an evolutionary success.

From the perspective of cultural evolution—as rigorously mathematically modeled over the years, and as observed in the world–broad-spectrum childhood gullibility is a bit of a theoretical non-starter.

Empirical challenges

Aside from being theoretically questionable, the childhood gullibility model also runs afoul of a wide range of evidence coming from developmental psychology—the subdiscipline of psychology that studies how kids learn about the world, represent ideas, and so forth. Just to highlight a few of these mismatches, let’s consider three sorts of questions:

1.     Do kids always prefer to learn from their parents?

2.     Do kids think about religion in the same way they think about more mundane facts?

3.     How do kids learn about things they can’t see?

1. Does mommy know best?

Parents (and other elders) hold a special position in the Dawkins childhood gullibility model. According to Dawkins, one’s parents have a vested interest in one’s own wellbeing, and it might behoove learners to just believe whatever their parents tell them (or so the story goes). And as a parent to two children, believe me: I truly wish it was the case that kids believe whatever they’re told. But kids don’t seem to be built that way.

Consider developmental psychological research on source credibility. The Dawkins model posits that kids just passively and uncritically accept information from certain people. But it turns out that kids are savvy and active in their information consumption. They recognize early on in life that some people are better to learn from than others, and they rely on a bunch of cues to assess source credibility, including cues like a source’s history of accuracy, or the degree to which they project confidence and certainty in their claims.

And they also seem to realize that different types of information require different cultural models. For a lot of information, kids seem to use age as a cue to credibility – they often prefer to listen to their elders, especially elders who seem reliable. But for other types of information, such as information about cool new toys, they recognize that similar-aged peers are actually more in the know.

Kids can treat different information sources as more credible than others, and know when to shift their attention to different sources depending on what type of information they’re after. This is quite at odds with a model positing that kids just blindly believe their elders.

The bottom line here is that kids aren’t blindly gullible when it comes to their parents or anyone else. They’re keenly attentive to who might know best, under which conditions. Kids love to see evidence. They love asking “why.” They are curious little critters, and active in their curation of knowledge. They are little scientists, rather than passive receptacles whose heads will fill up with whatever their parents pour in. 

Bottom line: if there’s a general psychological predisposition for kids to just believe their parents or elders, developmental psychologists have yet to find it.

2. Religion versus facts

Dawkins posits that kids can’t differentiate between the good factual advice their parents give them about predators and the iffy information their parents give them about gods: “If adults ever gave bad advice alongside good advice, the child brain would have no way to distinguish bad advice from good.” This implies a child psychology whereby kids are not very good at distinguishing factual claims from other sorts of claims. 

Once again, this isn’t how child psychology works.

First, it’s important to recognize that kids are quite adept at separating fact from fiction. There are substantial scientific literatures devoted to just figuring out the details of how kids think about things that actually exist versus things that only exist in pretense. They positively excel at entertaining fictional worlds, without any confusion about whether or not it is fact. They have imaginary friends, and even imaginary enemies, that they love to ‘play’ with but recognize are not real. One of my favorite papers in this area even looks at how children can mentally represent overlaps between different fictional worlds – they can think about what Batman might think about SpongeBob, without any confusion about the fact that neither character is real. 

The bottom line here is that kids are pretty good at sifting different sorts of claims into different mental bins: some bins hold facts, others fictions.

Beyond the fact/fiction divide, there’s some excellent research showing that children are pretty good at sorting factual claims from opinion claims. They realize that “the Colorado Avalanche won the Stanley Cup in 2022” is a verifiable factual claim about a hockey team, but that “the Colorado Avalanche are the best hockey team ever” is just someone’s opinion. Kids are quite sensitive to this fact/opinion distinction, and they readily apply it.

How then do kids represent religious claims: as fact, or as opinion? (Remember: the Dawkins gullibility model specifically says that kids will believe religious claims in exactly the same way as they believe factual claims about predators, because they can’t distinguish “good advice” from bad). Some excellent work by Larisa Heiphetz and her colleagues over the years has revealed that kids think about religious claims as sort of an intermediate category, hanging in the balance somewhere between fact and opinion. Some religious claims contain factual elements. But kids readily recognize that people can have different religious beliefs without just one person being absolutely correct with theirs. In this way, religions are viewed a bit more like opinions.

Kids seem adept at separating factual claims from claims of opinion, and they place religious claims somewhat in the middle. This rather strikingly contradicts Dawkins’s assertion that kids are susceptible to religion and superstition because they need to learn facts about dangers in the world. In fact, kids don’t even represent fact claims and religion claims as belonging to the the same epistemological categories.

3. Learning about the invisible

There are most definitely some things in the world that we can only learn about from others. Think of some of the myriad invisible threats we must beware of. Two-plus years into a global pandemic that has killed millions, this shouldn’t be a difficult task. There are innumerable tiny entities out there that can kill you…but have you ever directly seen these pathogens? No. Instead, we rely on the testimony of others to inform us about the unseen.

Testimony turns out to be an immensely important topic in the child development and learning literature. How do we decide whose testimony to trust, and for which types of information? As with above, it turns out that kids are quite savvy about this. They’re pretty good at figuring out that there are some things they’ll never be able to directly observe, but that are nonetheless important. Reliance on testimony turns out to be a great idea…but blind gullibility is not at all the expected outcome.

In what you may recognize as a recurring theme by now, it turns out that kids have a nuanced appreciation for the types of things that they can only learn about via testimony. For example, both germs and gods are things that we learn about from others telling us about them. But are we just as convinced by testimony about both germs and gods? Learners are actually more convinced by testimony about natural unseens (germs) than about supernatural unseens (gods, ghosts, and the like). They seem to have some intuitive skepticism of testimony about the supernatural—after all this testimony directly contradicts their core expectations about and experience with the world. As a result, kids seem to flag testimony about the supernatural as requiring special attention.

There’s a great deal of excellent developmental psychological work that specifically looks at how kids learn and think about supernatural agents, like gods. Interested readers should look up pioneering researchers in this area: folks like Drs. Kathleen Corriveau (Boston), Paul Harris (Harvard), Rebekah Richert (UC Riverside), Larisa Heiphetz (Columbia), Paul Bloom (Toronto), Telli Davoodi (Princeton), and others. These fine scientists, who directly study how kids learn about religion, present a psychological profile of children that is directly at odds with the childhood gullibility model Dawkins has presented over the past two decades.


In The God Delusion, and again 15 years later in Outgrowing God, Richard Dawkins presented a pet story about religion that depends on the notion that kids blindly believe whatever their parents tell them to do. Millions of readers may assume that this idea holds some scientific validity (or at least plausibility) – after all, it’s a claim made repeatedly over more than a decade by an incredibly prominent thinker who is a public advocate for science and skepticism.

Unfortunately, the idea doesn’t hold up. The childhood gullibility model of religion fails some basic tests of theoretical coherence–our most rigorous cultural evolutionary models suggest that a psychological predisposition for blind gullibility is unlikely to have ever evolved. Individuals predisposed to gullibility would be suckers, easily taken advantage of and outcompeted by more skeptical peers. The childhood gullibility model also conflicts with our best science on how kids actually learn and think about the world. This science portrays children not as passive receptacles of whatever hooey their parents dump into their heads, but rather as savvy and skeptical consumers who actively seek out and test cultural information. 

It’s time to listen to the scientists who actually study these topics. The psychological world they’ve uncovered is fascinating —a world where children aren’t passive receptacles but are rather little skeptics and proto-scientists.

A hallmark of science is its capacity for self-correction. We scientists try to stress-test our ideas, and are proud of our (occasional) ability to discard ideas that don’t work. In a previous story, I urged readers to discard the cherished notion that atheism ultimately boils down to rationality. Here, I’m asking readers to do the same with the notion that religion is rooted in childhood gullibility. It’s time to discard this one and instead listen to the scientists who actually study these topics. They’ve done some truly excellent research over the years, and it’s well worth engaging with their work on this topic. The psychological world they’ve uncovered is fascinating —a world where children aren’t passive receptacles but are rather little skeptics and proto-scientists. This psychological world is starkly at odds with any evolutionary story of religion that relies on the assumption that kids are blindly gullible to their elders’ teaching. 

Such is science—always there to poke holes in even intuitively plausible ideas, like the Dawkins gullibility model of religion.

Avatar photo

Dr. Will M. Gervais is a global leader in the scientific study of atheism and a Senior Lecturer of psychology at the Centre for Culture and Evolution at Brunel University London. Dr. Gervais’s research...