The data presented here indicate that stimulus-stimulus, stimulus-response, and stimulus-outcome relations can all affect conditional discrimination learning.
That the relation between sample and comparison plays a role in what is learned, suggests that identity is a meaningful concept for pigeons. Pigeons learn more than arbitrary S-R chains during acquisition of a matching-to-sample task. Similarly, that differential outcome can serve as a relevant cue for correct comparison choice suggests that expectancies can have stimulus properties not unlike those of external stimuli.
The relation between the sample stimulus and responses to the sample, also appears to be under cognitive control. Although the data can be described in terms of the simple additive effects of processing the visual stimulus and processing the feedback from sample-induced behavior, the fact that on each trial the visual stimulus must be processed sufficiently to identify the appropriate response suggests that birds are actively ignoring (and forgetting) the visual stimulus prior to the comparison response.
Finally, research we have described suggests that cognitive processing of stimuli during the acquisition of conditional discrimination learning tasks can also occur in other ways. The extent to which pigeons can use codes and can better define concepts when exposed to negative instance training further reflect their cognitive ability.
Research we have presented here supports the notion that cognitive processes are involved in pigeons' conditional discrimination learning. We have also identified a number of variables that might determine the extent to which pigeons will develop or demonstrate concept learning when training involves matching or oddity tasks. The goal of future research will be to identify the underlying cognitive processes involved, and to relate these findings to the development of conceptual behavior in other organisms, including humans.
We thank Joyce A. Jagielo for her contribution to the research presented. W. K. Honig for his helpful comments on an earlier draft, and Shirley Jacobs and Julie McKinzie for their help in preparing the manuscript. The research presented was supported by NIMH Grants MH 19757, MH 24092, and MH 35378.