When can children handle referential opacity? Evidence for systematic variation in 5- and 6-year-old children's reasoning about beliefs and belief reports.

Five- and 6-year-olds (N=51) heard stories in which a character sorted items into two locations. Either the character had a false belief about one of the items (e.g., thought a tin contained biscuits, not Lego), or was only partially informed of an item's dual identity (e.g., did not know that a tie was a present). Children found it easier to reject a report of the character's belief that described the true state of affairs when the character had a false belief (e.g., Is Fred's uncle thinking "where shall I put this Lego?"), than to reject one in which an object known to the character was described using a term of which she was ignorant (e.g., Is Mum thinking "where shall I put this present?"). Similarly, children found it easier to predict the character's incorrect sorting of the target items for false belief (with food not toys) than for dual identity (in the wardrobe not with things to take on a visit). Correct reasoning about beliefs and reports of beliefs that misrepresent an object does not imply mastery of the fact that beliefs represent an object in a particular way.