Commentary on Part II

Some of the most fundamental questions one can ask about cognitive development are related to the nativist-empiricist debate: How do we best characterize the initial state of the infant, and how do we explain the transition from the initial state to the mature cognition of adults? Carey and Markman (1999) argued that, at heart, this is a debate about whether there are domain-specific learning mechanisms. The empiricist position is that the initial state can be characterized in terms of perception and that general, all-purpose learning mechanisms, which can form associations and detect correlations, are sufficient to account for the acquisition of knowledge. Many of the outstanding issues in cognitive development center on questions of the domain specificity of learning and of the need to postulate core concepts or constraints on learning. Also central to the field are questions of whether cognition, especially the cognition of infants and young children, should be characterized as perceptual or conceptual, and whether attributing “theories” to infants and young children is necessary to account for their understanding of the world or whether it is an anthropomorphic overinterpretation of the evidence. These six chapters on the nature of children’s categories raise many of these most basic questions about human cognition and its development and provide thoughtful discussions, pertinent evidence, and elegant models for how best to characterize children’s classifications. We will emphasize two of these questions in our comments: (1) Are children’s classifications theory based or derived solely from computing correlations over perceptual features? (2) What role does language, in terms of naming categories, play?