Formation of equivalence sets in pigeons

Pigeons were reinforced for responding to a set of 20 out of 40 photographic slides, all 40 of which contained trees. The 40 slides were seen twice each day, in a different random order each time. After several sessions, the reinforcement contingencies were reversed: Previously positive slides were made negative, and previously negative slides, positive. After several more sessions, contingencies were again reversed, and so on throughout the experiment. Early in training, rho (the probability of ranking a positive over a negative) was less than .5 for the first 40 slides in the sessions in which the contingencies were reversed, indicating that the birds were continuing to respond approximately as they had in the previous session. Later in the experiment, rho for reversal sessions climbed above .5, indicating that with exposure to the reversed contingency for just the initial slides in the first session of reversal, the birds then responded correctly to most of the rest of the slides. The birds had learned arbitrary stimulus equivalences, so that the reversed reinforcement contingencies for slides at the beginning of a session predicted reversed contingencies for the other slides in the respective sets.