ABSTRACT
In the science of learning there are two technical definitions of punishment: the response-contingent presentation of a negative reinforcer or withdrawal of a positive reinforcer; and the suppression of a response by a stimulus contingent on that response. The theoretical issue concerns whether punishment suppresses responses directly or through the reinforcement of responses that are incompatible with the punished response. Punishment suppresses a response more rapidly and completely if an alternative response that produces the same reinforcer at the same rate is available. The punishment of long IRTs on VI schedules of reinforcement provides a good illustration of the relationship between punishment and avoidance. Continuous punishment produces greater suppression than intermittent punishment does. When punishment is discontinued, rate of responding first exceeds the level it was at before the introduction of punishment and then returns to that level. Response-rate patterns produced by various schedules of punishment are typically opposite to those produced by the corresponding schedules of reinforcement.
