study has been criticised: Simon, Hespos & Rochat (1995) results explained by infants’ knowledge of physical objects in world, rather than numerical understanding. Added condition where identity of toy changed, but not number. - Replicated Wynn 1992. When screen comes down, can violate expectation in different ways: change identity or change in quantity. 4 conditions: possible, arithmetically impossible (Wynn), identity impossible, identity & arithmetically impossible. If infants can do simple additive reasoning, & Wynn was correct. Infants showed no increase in looking time in 2 add conds test identity – suggests only impossible arithmetic outcomes are surprising not identity changes. Xu & Spelke (2000): infants sensitive to ratios. Starkey, Spelke & Gelman (1983): cross-modal understanding of number.