The effects of target salience in operant feature positive discriminations

The effects of the salience of the target stimulus on the nature of stimulus control acquired by the feature cue in operant feature positive discrimination learning were examined in two experiments. Each rat was trained with two discriminations, XA + /A − and YB + /B −, with a different response reinforced in each discrimination. When the feature (X) was trained with a target (A) of low salience, transfer test responding in the presence of an XB compound was of the type reinforced during XA in training, regardless of the salience of B. In addition. nonreinforced presentations of X alone after discrimination training eliminated responding in the preesence of XA and XB. In contrast, when the feature (X) was trained with a target (A) of relatively high salience, transfer to an XB compound depended on the salience of B. If B was of low salience, XB exerted little or no control over the rats′ responding, but if B were of high salience, XB controlled the response that had been reinforced during YB in training. Furthermore, nonreinforced presentations of X had no effect on responding during XA or XB. The results were discussed in terms of occasion setting and configural conditioning.