When do children begin to understand that things in the world are made of different kinds of "stuff," and that the "stuff" something is made of, has an effect of the behavior of the object? One theory suggested that young children categorize based on visual characteristics of objects, rather than material type. We hypothesized that as children begin to understand that objects made of different materials do different things (e.g. rubber stretches and bounces, cork floats), they will use that knowledge to categorize objects. The "thing" that a given material does is referred to as its causal property, what the material itself causes the object it comprises to do. In other words, the material causes the object to display a particular property (e.g. wood burns, plastic melts, metal conducts).

    We hoped to test whether there is a developmental change in whether children rely on objects' visual appearance or objects' causal property to categorize objects. To test our hypothesis, we devised a small machine (the material detector) which could be made to light-up from a remote switch, hidden from the subject. Children were shown the "material detector", but were not told what the material detector could do. After demonstrating that a target object placed on the machine would make it light-up, we asked children to make predictions about which of 24 novel objects (e.g. abstract wood, metal, and plastic pieces) would make the machine light-up. We hoped that the demonstration with the target object would allow the children to make predictions about the novel objects based on either their constituent material's causal property (making the material detector light-up),or on the visual appearance of the object, depending upon the child's natural preference for categorizing. Predictions based on the visual aspects of the objects would constitute categorization based on visual appearance (color, shape and/or size), while predictions based on the causal property of the objects (whether they turn-on the material detector) would constitute categorization based on material type.

    We found an increasing tendency to categorize (make predictions) based on the objects' constituent material/causal property as children get older: 4-year-olds made 36% predictions based on material, 5-year-olds made 50% predictions based on material, 6-year-olds made 67% predictions based on material, and 7-year-olds made 83% predictions based on material. This suggests that as children get older, they are more able to use their knowledge of material kind and causal property to evaluate and categorize objects in their world.