Answers are grounded in this paper and cite their sources.
When children learn categories, they do not learn isolated facts but rather systems of knowledge. These systems of knowledge are composed of property-property (e.g., things with wings tend to have feathers), property-role (e.g., things with eyes tend to eat), and role-role (e.g., things that eat tend to sleep) correlations. Research has shown that even young children have knowledge of property-property and property-role correlations, but there has been little direct evidence of their knowledge of role-role correlations or for how children might acquire this knowledge. We found that 5-year-olds have knowledge of role-role correlations and 2- and 3-year-olds have different degrees of knowledge of property-role correlations. Two-year-olds needed multiple properties to infer an appropriate role, whereas 3-year-olds needed only a single property (e.g., eyes or wheels). These results suggest that role knowledge may first be linked to object (or literal) similarity, then to clusters of properties, to a single property, and finally to other roles.
No takes yet. Share an insight, caveat, or question.
Sheya et al. (2006) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: