Cristine Legare: The Development of Causal Explanation

Cristine Legare, who is visiting the Centre for Anthropology and Mind for the next six months, gave a talk last Monday on her research. She seems to have done an incredible amount of work already, encompassing experiments on children's causal reasoning and cross-cultural research, in Africa and the US, on explanations for AIDS. Her basic question of interest -  a very big question - is how we construct knowledge through explanation, exploration and hypothesis-testing. In particular, she is interested in how different explanatory frameworks - the scientific and the supernatural - may coexist. Her starting point in the talk was that questions are powerful problem-solving tools, and that the process of seeking, generating and evaluating explanations imposes constraints on learning that are often beneficial. Hence, students (in the broad sense; all children are students of the culture they grow up in) who explain things - even to themselves - learn material more effectively and generalise more readily to novel contexts. 

The first part of her talk dealt with her causal reasoning experiments, some of which were set out in a recent article:

Legare, C. H., Gelman, S. A., & Wellman, H. M. (2010). Inconsistency with prior knowledge triggers children's causal explanatory reasoning. Child Development, 81, 929-944.

I'm going to be lazy and post the abstract rather than summarize what she said - she puts it much better than I would have done anyway!:

"What events trigger causal explanatory reasoning in young children? Children's explanations could be triggered by either consistent events (suggesting that explanations serve a confirmatory function) or inconsistent events (suggesting that they promote discovery of new information). In 2 studies with preschool children (N = 80), events that were consistent with children's prior knowledge, and children were invited to explain either outcome (or both). Results demonstrate that inconsistent outcomes are an especially powerful trigger for children's explanations and that the explanations children provide for inconsistent outcomes refer to internal causal properties, overriding perceptual appearances. In sum, the data provide empirical evidence that inconsistent events motivate children to construct explanations, thereby suggesting that children's explanations function in the service of discovery."

One especially interesting point was that in about 50% of cases, children chose a causal-functional explanation for why one of the gadgets wasn't behaving as expected, while in 35% of cases they chose a "category switch" explanation, arguing that the gadget was "really" an example of a different category of things. Notably, these category switch explanations increased in frequency with age. Legare sees in them the possible origins of confirmation bias, whereby people discount evidence that something strange is going on and effectively see what they want to see.

Although this is important work, for me an even more interesting part of Legare's talk was the bit when she talked about explanations for AIDS in South Africa. Her main point was that science and religion need not be competitive ways of explaining the world: they may give rise to "synthetic" or blended models. Even five-year-olds in South Africa were more likely to give "bad blood" than contagion as an explanation for the spread of AIDS, so there is awareness of the science there. But in both SA and the US, there is a tendency also to look for supernatural explanations of AIDS - it's just that in America these explanations are couched in terms of God's punishment, whereas in Africa they are couched in terms of witchcraft. And whereas in America the punishment tended to be seen as a consequence of homosexuality or some other kind of "fornication", in South Africa witchcraft attacks were presumed to stem from envy, or from failing to show enough generosity to one's neighbours.

There were several interesting questions at the end of the talk. Legare had argued that the association of witchcraft with AIDS occurred because the latter had several features (notably, being associated with death) that made it especially salient to people. Harvey Whitehouse asked whether the unpredictability of AIDS might not also be a factor. Legare agreed and said that the prevalence of supernatural beliefs might ultimately stem from the fact that causal explanations are not much use in the social domain [because of second-order intentionality - or "free will", as some people might want to call it - though she didn't go into detail on this point]. People might be predisposed to see witchcraft because of a kind of "hyperactive hostility detection device" - if you think someone might want to do you harm, you can't be too preoccupied with the causal chain by which they might harm you. Interestingly though, in the causal reasoning experiments, social cues didn't make much difference to whether expectation violations were regarded as salient (but this was perhaps because the violations were so obvious in this case; she is planning new experiments in this area). Finally, Ryan McKay suggested that causal reasoning might be associated with the dorsolateral prefrontal cortex, as neuropsychiatrists, using similar experimental paradigms to Legare's, have found that people with delusions don't tend to distinguish very well between confirmatory and disconfirmatory trials.

My reactions:
I'm having lunch with Cristine tomorrow so I don't want to say too much about my initial reactions here: I'll maybe ask her some questions then and see how she answers them! I did ask her a question at the end of the talk about how initial curiosity becomes confirmation bias: when do we stop looking for causal answers and decide that our existing theory is good enough? She acknowledged that this is a major mystery of cognitive development, and suggested that there might be certain conditions under which further exploration is encouraged in a child (perhaps social conditions, I thought, though she was more equivocal about that).
Finally although she was careful to talk about the compatibility of "supernatural" explanations with scientific ones, I did still think she was a little guilty of making the easy division between the two that so many people do. The problem is that this is not necessarily a distinction that exists in the mind of the people that hold them: the explanations that we label "scientific" are simply the ones that happen to accord with our worldview; and whether there is actually any psychological difference between scientific and supernatural explanations is a moot point, it seems to me - both categories seem too big and diverse to compare in toto like that.