Abstract
The suppression of behavior that occurs in instrumental extinction is strikingly specific to the response. In contrast, Steinfeld and Bouton (2022) recently reported that inhibition developing in an operant feature-negative (FN) discrimination is not specific to the response. In two experiments, we tested two potential explanations of why inhibition in FN learning is relatively response-general. In each, we used Steinfeld and Bouton's method and concurrently trained two FN discriminations with different operant responses (AR1+/ABR1- and CR2+/CDR2-). We then assessed the extent to which the inhibitory cues (B and D) suppressed the response they were trained with (same-response inhibition) and the alternative response (cross-response inhibition). Experiment 1 tested the idea that FN inhibition might be response-general because it can create strong inhibition. Rats received either 3, 6, or 12 sessions of FN discrimination training (Steinfeld and Bouton's rats had received 12). Inhibition was response-general at every level of training. In Experiment 2, the inhibitors (B and D) were first trained as cues that set the occasion for R1 and R2 (respectively) before they were turned into inhibitors in the FN discriminations. In the end, there was less cross-response inhibition, and thus more response-specificity. We suggest that inhibition in FN learning may be response-general because the unambiguous inhibitory cue (B or D) can draw attention away from the response.