Natural concepts and generalization classes.

The results of an experiment performed by Herrnstein, Loveland, and Cable (1976) show that pigeons possess concepts that are not based on common elements. The conclusion they draw from these results is that these concepts are largely determined by genetic factors. In the present paper it is argued that, at least for some of the concepts, the conclusion is incorrect. These concepts-called here uncommon generalization classes-were acquired by the pigeons as a consequence of certain experiences through which they went during the experiment. The discussion will also suggest the fruitfulness of replacing the vague and misleading notion of concept by the more adequate notion of generalization class.