Operant or instrumental behaviors become inhibited in a number of experimental procedures. In extinction, a reinforced response is allowed to occur without reinforcement, and the response declines over time. The results of recent experiments from this laboratory and others have suggested that extinction of an operant response occurs because the animal learns to inhibit the specific response. Operant responses can also be inhibited in discrimination procedures. For example, in a feature-negative procedure, animals learn that responding is reinforced during cue A (AR+) but not when A is compounded with cue B (ABR-). Experiments using this sort of method have produced evidence of considerable cross-response inhibition. Given the contrast with extinction, six experiments in this dissertation attempted to understand the conditions that produce cross-response vs. response-specific inhibition. I developed a procedure in which rats learn two feature-negative discriminations with two different responses (AR1+, ABR1- and CR2+, CDR2-) and the effects of the inhibitory cues (B and D) are tested with both responses. The research found that the inhibitors suppress the response they were trained with (same-response inhibition) and the response they were not trained with (cross-response inhibition) about equally (Experiment 1). The results of Experiment 2 confirmed a role for response or prediction error in the development of inhibition, and ruled out a possible role for competing responses in inhibition. Experiment 3 ruled out the possibility the inhibitors produce cross-response inhibition by predicting the absence of the common reinforcing outcome. Experiments 4 and 5 found a role for previous inhibition learning with the response (inhibitability) to support cross-response inhibition. Experiment 6 tested the possibility that inhibitability could also produce cross-response inhibition in an extinction design, and failed to find such evidence. Overall, the results suggest substantial cross-response inhibition in feature-negative designs, much of which depends on the organism's experience with the inhibitability of the response. Furthermore, extinction and feature-negative inhibition may produce fundamentally different types of inhibition.