How do we come to understand the world so that we can act in it, so that we talk about it? Where do our rich understandings – of causality and intention, of number and objects, of space and time, of language – come from? Many have looked at the diversity of human knowledge and at the certainty with which children acquire it all and concluded that the diversity is there to begin with. They have suggested that knowledge acquisition is driven by special mechanisms dedicated to specific domains (e.g. Fodor 1975, 1983; Chomsky 1980; Gleitman 1986; Keil 1989; Gelman 1990; Spelke, Breinlinger, Macomber, & Jacobson 1992; Leslie 1994).

The argument for the domain-specificity of cognitive development is easy to make along four lines.

1. What needs to be acquired is unique to each domain. Thus, nothing beyond the most superficial metaphors are likely to exist across domains.

2. One cannot get something from nothing. In cognitive development, this idea is often presented in terms of the problem of induction and the need for constraints. Briefly, there are an infinite number of objectively correct generalizations from any data set. Therefore, learning cannot happen without prior content-specific constraints on the generalizations that are formed.

3. The empirical evidence on children's learning strongly implicates domain-specific constraints on what is learned. Children exhibit learning biases specific to the content being learned.

4. […]