Sensitivity to the frequency of parts and kinds: two principles of organisation in semantic memory.

Are subjects sensitive to the frequency of parts of the same whole in the same way that they are sensitive to instances of the same category? Subjects read a randomised list of words and then estimated how many words on the list had belonged to particular groups, e.g. parts of a bird, or kinds of bird. Groups of words were either parts of the same whole, e.g. feathers, beak, wing, or members of the same category, e.g. robin, sparrow, eagle. Subjects were asked, either, "How many parts of a (bird) did you see?" or "How many kinds of (bird) did you see?". Half of the subjects expected the frequency estimation task, whereas for the other half, who performed a distractor task during encoding, the task was unexpected. In Experiment 1 subjects saw either parts of kinds. Sensitivity to frequency did not differ for parts and kinds nor as a function of whether the frequency estimation test was expected. In Experiment 2 subjects saw both parts and kinds belonging to the same superordinates, e.g. parts of birds and kinds of birds, and were asked to estimate the frequency of both types of groupings for each superordinate. The results were the same as in Experiment 1. In addition, subjects were able to separate the frequency of parts and kinds of the same superordinate. The results suggest that, in frequency estimation, part-of and kind-of associations can function similarly, and that the mechanism responsible for frequency estimation is sensitive to the type of association.