Being Nativist about Mindreading: more Demanding than you Might Think

Two distinct theses are sometimes argued for, or against, together in the debate on early mindreading: that infants mindread, and that this occurs thanks to innate mental notions. When this is the case, the underlying assumption is that mental notions cannot be learned early in infancy, and therefore at that stage they must be either innate or not present at all. I do not intend to directly argue in favour of the opposite claim that mental notions can be learned early in infancy. My more indirect route is to argue that a much stronger form of nativism than what is ordinarily thought is required if innate mindreading abilities are to be attributed to infants. This means that, if one defends early mindreading, the choice is between acknowledging that infants can learn to mindread, on the one hand, and being committed to a rather cumbersome version of nativism on the other. My argument is based on the claim that mental notions need to subsume – and therefore generalize over – a host of behavioural rules. This argument is assessed against the nativist position of Baillargeon and colleagues and the anti-nativist position of Ruffman. Finally, I consider a possible objection to the claim that mental notions essentially consist of statistical generalizations from experience.