The authors outline a cognitive and computational account of causal learning in children. They propose that children use specialized cognitive systems that allow them to recover an accurate “causal map ” of the world: an abstract, coherent, learned representation of the causal relations among events. This kind of knowledge can be perspicuously understood in terms of the formalism of directed graphical causal models, or Bayes nets. Children’s causal learning and inference may involve computations similar to those for learning causal Bayes nets and for predicting with them. Experimental results suggest that 2to 4-year-old children construct new causal maps and that their learning is consistent with the Bayes net formalism.

Words, kinds and causal powers: A theory theory perspective on early naming and categorization. For some twenty-five years, the prevailing theories of categorization in philosophy have invoked the idea of “kinds ” (Putnam, 1975; Kripke, 1972). When we look at how adults use words to refer to categories of things we find that they only rarely categorize objects on the basis of their common properties. Instead, adults seem to categorize objects together when they believe that they belong to the same “kind”; that is, that they share some common, abstract “essence.” Psychological investigations of adults have largely confirmed these philosophical intuitions, adults do seem to group objects together based on “kinds ” rather than properties (Murphy &

We outline a new computational account of learning in children using the causal Bayes net formalism. We also present evidence that children as young as two years old use something like causal Bayes net learning mechanisms to infer the causal structure of the world around them. This kind of learning may play an important role in the development of intuitive theories. Finally we contrast causal Bayes net and neural net learning mechanisms.