The present study assessed the ability of a reinforcer to mediate an association between two stimuli that independently predict the occurrence of that reinforcer (acquired equivalence of cues). In Experiment 1, pigeons were trained on a shape (plus and circle) matching-to-sample task and a color (red and green) matchingto-sample task. Within each task, correct responses were systematically reinforced with corn on some trials and wheat on other trials so as to establish associations between one stimulus from each task and a "common" outcome (e.g., plus-corn; red-corn). Following training, birds were transferred to a symbolic matching-to-sample task wherein a stimulus from one training task was presented as the sample and the two stimuli from the other training task were presented as comparisons (e.g., plus sample; red and green choice stimuli). On the first transfer session, experimental birds made significantly more correct responses than controls, that is, birds "matched" stimuli previously associated with a common outcome. Experiment 2 replicated this acquired equivalence effect and controlled for food preference. Delayed matching-to-sample training in Experiment 2 also demonstrated enhanced memory performance for birds exposed to differential reinforcement contingencies (acquired distinctiveness cues), but this effect was confined to the shape task. Overall, these results indicate that (a) a reinforcer can serve as the basis for organizing otherwise unpaired predictive cues in memory, (b) animals will selectively use differential expectancies as cues for solving complex discrimination tasks, depending on the difficulty of the discrimination, and (c) overt behavior is not importantly involved in the learning or maintenance of reinforcer-specific expectancies.There is a substantial body of literature ulus is thought to be sufficient to evoke an to support the notion that animals can learn outcome-specific expectancy (cf. Bolles, to "expect" reinforcement. Apparently, this 1972). expectancy is based on the qualitative and Recent evidence indicates that under some quantitative properties of the reinforcing circumstances animals will use expectancies event, and presentation of a predictive stim-of different reinforcers as cues for solving complex discrimination tasks. In particular,

This research was supported by Grant MH24092 process; and the fact that in the case of from the National institute of Mental Health to the matching, the associated stimuli are of the first author. We thank Tammy Rayburn for help with same h ue whereas in the case of oddity they th RSrr SSTZSti sent to Thomas R. are <* different hues contributes neither Zentall, Department of Psychology, University of to the wa V ln whlch the tasks are learned Kentucky, Lexington, Kentucky 40506. nor to the relative difficulty of the two 70

No abstract

No abstract

No abstract