Pigeon's (Columba livia) paradoxical preference for the suboptimal alternative in a complex foraging task

Recent research has examined a task in which choice of 1 alternative A provides reinforcement and in addition, allows access to alternative B that also provides reinforcement. However, although initial choice of B also provides reinforcement, it does not also allow access to A. Thus, optimal performance would be to always choose A. Curiously, Salwiczek et al. (2012) reported that adult wrasse (cleaner) fish mastered this task within 50 trials, whereas monkeys and apes had great difficulty with it. The authors attributed the species differences to ecological differences in the species foraging experiences. However, Pepperberg and Hartsfield (2014) found that parrots too learned this task. In Experiment 1, using the manual presentation of stimuli, we found that pigeons actually showed a reliable preference for B, the suboptimal alternative. In Experiment 2, we replicated the suboptimal preference using an automated version of the task. We hypothesized that the pigeons may have been basing their preference on the frequency of reinforcement associated with each alternative (initially, all trials ended with choice of B, whereas only half of the trials involved choice of A). In Experiment 3, we tested the hypothesis that the pigeons' preference was influenced by the frequency of reinforcements associated with A and B. Thus, when the pigeon chose A, we replaced B with C, so reinforcement occurred to B only when they chose it first. With this procedure we found that B was no longer preferred over A. Thus, the data supported our hypothesis. (PsycINFO Database Record