Field of Science

Mr Smart and Heroman

Let me introduce you to Mr Smart and Heroman. Mr Smart is really, really clever. So clever that he knows everything - like what's inside a closed box. Heroman is not so smart, but he does have a special power. Heroman has x-ray vision, so that he can see into the closed box.

Here's a picture of Mr Smart. He looks a bit like a lot of Professors I know.

Both Mr Smart and Heroman had a key role to play in a recent study by Jonathan Lane, of the University of Michigan, and colleagues, into how children come to understand magical beings.

There are basically two schools of thought on this. One is that they have to learn first about ordinary minds and then, building on that platform, they learn about extrardinary minds.

The other school holds that children are born with an inbuilt predisposition to think that all intelligent beings  have god-like omniscience. They then have to learn that, sadly, their parents and their friends are in fact limited in what they know.

A leading proponent of this idea is the cognitive psychologist Justin Barrett. His studies of the beliefs of young children have shown that the youngest (aged 3) seem to intuitively believe that all people (and God) are omniscient - they know everything that the child knows. Older children (aged 5) have learned that Mum has her limitations.

The distinction is important. At stake is the issue of whether we are born with an innate predisposition to believe in an Abrahamic god. As Barrett explains:
...early-developing conceptual structures in children used to reason about God are not specifically for representing humans, and, in fact, actually facilitate the acquisition and use of many features of God concepts of the Abrahamic monotheisms (Barrett & Richert, 2003)

But perhaps it's not that simple. Childhood development is a rapid, complex process. So Lane and colleagues set out to get a bit more granularity into the picture, by learning about the beliefs of children in the middle range, at around 4 years old.

The basic experiment is simple. The experimenter sits with the child and a box of crayons in a room. Except the box doesn't really contain crayons. it's got rocks in it instead.

The child knows that, because she's been shown them. But then the box is closed up again. The question for the child is this: who else will know what's inside the box, if they come into the room?

Would another girl her age know? Would her mum know? What about Heroman and Mr Smart? What about God?

Well, the results change with age, and in a fascinating way. On the graph, higher scores mean that the individual concerned would guess (wrongly) that there are crayons in the box.

The youngest age group, just under 4 years old. mostly think that everyone - Mum, Mr Smart, and God - would know that the crayon box actually has rocks in it.

The middle group, around 4 and a half years old, are more likely to think that they would be fooled, and think (wrongly) that the box holds crayons.

The exception is Heroman. The 4.5-year olds reckon that Heroman could see into the box, and so know that it contains rocks.

The oldest group, around 6 years old, have pretty much all figured out that Mum and the girl would be fooled, but that Mr Smart, Heroman and God would not be.

Here's what the researchers think is going on. The youngest children have what's known as 'reality bias':
When asked about what other people know or believe, very young children tend to answer by simply assessing reality and using that information to infer others’ knowledge and belief

The middle group, however, have developed enough to understand ignorance:
Soon after children develop an appreciation for the distinction between knowledge and ignorance, they begin to appreciate the distinction between reality and belief; they start to understand that others, misled by inaccurate perceptual cues or outdated information, can hold false beliefs
Only Heroman is not ignorant, because only he can see inside the box.

The oldest children have also learned that some agents - gods and the like - have (or are supposed to have) superhuman knowledge.

Lane concludes that childhood development proceeds in the exact opposite direction to what Barrett proposes. Rather than intuitively understanding the idea of omniscience, children naturally understand all agents - people and magical beings - to be limited in the same way as the people they know.

Realism, in short, is natural. The idea of the supernatural has to be learned.

ResearchBlogging.orgLane JD, Wellman HM, & Evans EM (2010). Children's understanding of ordinary and extraordinary minds. Child development, 81 (5), 1475-89 PMID: 20840235

Creative Commons License This article by Tom Rees was first published on Epiphenom. It is licensed under Creative Commons.


  1. That seems contra to what I have read that atheism is learned.

  2. (1) Sample since and significance levels would be nice. Especially since the article seems to confirms Tom's preferences. :-)

    (2) Tom says:
    "The idea of the supernatural has to be learned."
    I think the generalization from the study is way too broad. Instead, you could say:
    "The idea of the omniscience has to be learned."
    "A component of the supernatural has to be learned."

    Am I not right, for instance, that the cognitive illusion of dualism is not learned, and it is a huge player in experiencing the supernatural?

  3. [Barrett's] studies of the beliefs of young children have shown that the youngest (aged 3) seem to intuitively believe that all people (and God) are omniscient - they know everything that the child knows.

    I don't know if this is poor phrasing, if there is more to Barrett's argument that is being omitted, or what... but to me "know everything that the child knows" and "omniscient" are worlds apart!

    Even before reading this interesting study, it seemed to me that it's not so much that young children believe that people know everything about reality, but rather that they are unable to clearly distinguish between separate minds. They still understand that there are some things that are known to them and some that are unknown -- but they don't understand that each mind has its own set of knowns and unknowns.

    In any case, cool study -- though I agree with Sabio's comment under (2).

  4. Sabio: (1) Good question! There were 57 kids. Which sounds like a small number, but is fairly typical for this kind of study.

    Of more concern to me is how they split the children into the three groups. This seems to have been arbitrary, and might have been done after the study. They might have split them up so as to get the best results to fit their theory. Perhaps not, but they don't explain it in the paper.

    (2) yes, you're right, my concluding statement is too general. What specifically this study suggests is that children have to learn the concept that some minds have extraordinary levels of knowledge. But that's not quite so snappy :)

  5. James: It's poor phrasing on my part. What Barrett believes is that his research shows that children understand omniscience. In fact, at best it shows that children understand that some minds have extraordinary knowledge (per my comments to Sabio, above).

    But both of these are different to "knowing everything that the child knows". That's reality bias - what the youngest children have.

    I was thinking of writing about this for my next column in Free Inquiry, so I'm glad I piloted it here!

  6. Samuel: Good science gives unexpected results :) More seriously, this is just one aspect of the cognitive biases that contribute to different world views.

    The best way to look at it is that both atheism and religion are learnt, but that different cognitive biases contribute by buttressing environmental cues.

  7. I'm not sure I can agree that atheism is a "learnt" behavior. Atheism is simply the lack of belief in a deity. A child raised in the absence of belief in, say, superheroes hasn't learned asuperheroism. There are all kinds of things you don't teach kids, and their absence isn't a behavior they've learned.

  8. Tom:
    Since your are thinking about entering this in "Free Inquiry", let me pursue:

    (1) If the size of each group is too small for significance and the method of establishing the groups is non-random, then the study can not have meaningful results. If that is the case, then I would rephrase your sentence from:
    "...this study suggests ..."
    "...the authors of the study are unsuccessfully/deceptively trying to make it suggest ..."

    (2) "Snappy" is what commercial journalism is all about. Your blog is about knowledge, no. To sell your points with poor studies, making them say more than they can and not exposing the vacuous nature is reinforcing the type of biases you are trying to expose, no?

    Am I missing something here?

    (3) What I learned from this is that you really want "religion" to be learnt. Even after accepting my chastisement about generalizing, you respond to Samuel in generalized terms going not only from superstition to all of religion. Superstition is only part of religion. Superstition is also part of the non-religious Atheist's mind too. (as Bruce Hood has amply illustrated).

    [sorry, not good at cushioning my doubts -- please read generously]

  9. This merely seems to be a classic "theory of mind" experiment that has been done on children of this age group many times before. That is not to deny the application to the way in which some children are indoctrinated into a religious belief system. As Richard Dawkins has said, (I'm paraphrasing) "some of us just go one god (or tooth fairy, Santa Claus, elf, unicorn, or FSM) further." In other words, as a child grows and matures and learns to prune their list of possibly omniscient beings, perhaps some become convinced by the adults in their lives that even though they were right that all these other omniscient entities/beings do not exist, their intuition was right that there is one being that is omniscient, i.e. whatever "god" their parents and/or peers believe in.

    In the end, they get the satisfaction of saying to themselves and others that "I was too clever to fall for these silly, omniscient beings, but my intuition was right that there is one being that knows and sees all.

  10. @GoodDamon: It depends very much on how you define "atheism". It is well-known that children adopt teleological explanation by default (the skeptic community calls it "promiscuous teleology"; the child psychiatry community, as I recently learned, has a different phrase to describe it but I'll be damned if I can remember what it is...)
    While teleology is obviously not synonymous with theism, it does imply a "purposer", if you will.

    Whether a "teleological atheism" is possible I suppose depends on your definitions. By the definitions I feel most comfortable with, I think Tom's statement that "both atheism and religion are learnt" is more or less accurate. While most agnostics are essentially atheists (and vice-versa), I would argue that a child who has not been taught anything about theism would be a pure agnostic -- their teleological philosophy implies a "purposer", but they don't really have any thoughts about that implication.

  11. Sabio: all the results are statistically significant. The study is bigger than most in this area. In fact, they actually did the study twice - the second time with a somewhat different experimental set-up. Both experiments gave the same results.

    They didn't seem to have a statistical plan, but that's true of just about every piece of research in this area I've read. All of it should be considered 'exploratory'. All studies are flawed in some way - you can always dismiss the ones you don't like! However, this one is as sound as any that I cover, and better than most because it's prospective.

    Interesting point you make about me "really wanting religion to be learnt". It's perfectly possible - we all have inherent biases, and often they're unknown to ourselves. But actually I've written loads of posts making the opposite case - because that's what the evidence suggested.

    This is the first time I've written a post with this conclusion, because this is the first study to clearly show such an effect. Of course, as with any single study it can only examines one aspect the complicated mental process that goes into the complicated beast we call 'religion'. The same applies to all the other studies I cover, even thought I often try to draw some more general conclusions from them.

    What's more, no one study ever proves anything. This study builds on a number of previous studies and helps to explain why they seemingly conflict. In that regards, it's particularly interesting.

  12. Romantic Rationalist: but what this study suggests is that we start out with a theory of mind that only encompasses limited minds, and that we have to learn (or develop) more complex concepts.

    Interestingly, there's lots of evidence that even adults don't intuitively understand the concept of an omniscient god. For example, Barrett has shown that we seem to intuitively think of god as responding to prayers in sequence, as a human would, rather than simultaneously, as an omniscient, all-powerful god would. That's the case even for people who understand what omniscience means.


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS