A componential view of configural cues in generalization and discrimination in Pavlovian conditioning.

This paper describes three theoretical approaches to the representation of configural cues in generalization and discrimination in Pavlovian conditioning: that of the Rescorla-Wagner model, the Pearce model, and the authors' 'replaced elements' model. We summarize the results of a generalization experiment using the rabbit Pavlovian conditioned eyeblink response where animals were trained with cues A, AB, or ABC, and tested with A, AB, and ABC. The pattern of generalization decrement in testing supported the replaced elements model.