抄録
In Experiment 1, eight pigeons were trained to discriminate between two positive and negative human faces in a go/no-go discrimination procedure. Complete transfer was shown to the morphed images composed of the positive faces and of the negative faces, regardless of morphed proportions. Responding to the faces composed of one positive and one negative face increased as a function of proportion of the positive face. In Experiment 2, four pigeons were trained to discriminate two sets of morphed images created from pairs of four original faces designated as A, B, C, and D in each set. The morphed images composed of 50% of each of the paired original faces; AB, AC, AD, BC, BD, and CD in a set. The birds were then tested with the stimuli including the 50% morphed images used for training, the original faces, and the prototypes created by averaging all the four original faces in each of the two sets. The most pronounced discrimination, even slightly better than to the stimuli used for training, occurred to the prototypes: a finding supporting the abstraction of the central tendency as a prototype.
