Conceptualization of Natural and Artificial Stimuli by Pigeons
Is conceptual behavior uniquely human? Or do other animals exhibit conceptual behavior? Since C. Lloyd Morgan first raised the possibility of conceptual behavior in animals, the matter has been hotly debated. Our own research on this issue has shown—with two quite different experimental methods—that pigeons readily learn and generalize four-category discriminations based upon full-color snapshots of natural (cats, flowers, humans) and artificial (cars, chairs) stimuli. Beyond these demonstrations, we have explored several factors that affect the speed of learning and the breadth of generalization. Learning is enhanced by arranging the collections of snapshots to coincide with human language categories, by repeating particular pictures, and by decreasing the number of pictures per category. Generalization is enhanced by increasing the number of pictures per category during original learning and by preserving the orientation of the snapshot during testing. Additional studies have shown that categorization is possible with nonrepeating stimuli, that test performance on previously repeated stimuli surpasses performance on novel test stimuli, and that confusions are more likely among stimuli from the same human language category than among stimuli from different language categories. These and other results not only persuade us that pigeons do learn concepts, but that the processes of conceptualization in animals resemble those of conceptualization in humans.