| Description |
The primary purpose of the present research was to investigate the aversive properties of s- following discrimination training with and without errors. A conditioned punishment procedure was used to assess the conditioned aversive properties of S- established through (a) an error nonreinforcement procedure, (b) an errorless nonreinforcement procedure, and (c) a response contingent electric shock procedure. Following discrimination training with these three procedures, responding to S+ produced briefly either S- or a neutral stimulus uncorrelated with nonreinforcement or shock. Compared to the effects of the neutral stimulus, S- produced (a) a decrease in responding for Ss trained to discriminate with errors, (b) no reliable alteration in responding for Ss trained without errors, and (c) an increase in responding for is trained with response dependent shock. In a subsequent phase of the experiment, generalization of conditioned punishment was investigated by punishing responses with a brief presentation of stimuli that varied along the S- continuum. A comparison of the nonreinforcement procedures showed both a steeper gradient and a greater decrease in responding to all values a l onq the S- continuum if the discrimination was trained with errors. In a final phase, the inhibitory control engendered by the three S- establishing operations was assessed by obtaining generalization of inhibition gradients. Relatively flater gradients were obtained for Ss trained to discriminate without errors as compared to is trained with errors. It was concluded that s-functions as an aversive stimulus following discrimination training with errors and as a neutral stimulus following errorless discrimination training. The failure of the shock procedure to produce a conditioned aversive effect was discussed. |
