Some relations between positive reinforcement and punishment effectiveness were explored in three studies employing shock punishment of food-motivated leverpressing. Amount of prepunishment training on a variable-interval schedule of reinforcement was varied over 7, 21, 35, 49, and 63 h and was found to be unrelated to magnitude of suppression produced by shock punishment. A second study found no relationship between density of reinforcement delivered on a variable-interval schedule during prepunishment training and subsequent response suppression. In a third study, alternating cycles of punishment/no punishment (vacation from punishment) eventually resulted in a lessening of the initial disruptive effects of punishment, although contrary to previous reports, successive vacations did not enhance the suppressive effects of punishment.The bulk of research on punishment has concentrated on the study of behaviors established by positive reinforcement. Therefore, it is not surprising that many investigators have been concerned with the question of how punishment interacts with variables supporting the maintenance of the punished behavior. The purpose of the present paper is to report on the influence of three such variables: the amount of prepunishment training, the density of the reinforcement schedule supporting the punished response, and the alternation of periods of punishment and withdrawal of punishment (vacation from punishment). The reason these three particular variables were chosen for study is that they so often play a part in the apparently arbitrary selection of baseline parameters used in studies of the punishment of learned instrumental behaviors. Additionally, as indicated in the text of this paper, there is some evidence indicating that each of these variables may influence the outcome of any given punishment procedure.
EXPERIMENT 1The initial study investigated whether the strength of learning, as manipulated by the number of prepunishment positive reinforcement training sessions, influences the amount of response suppression produced by punishment. Both intUitively and in terms of a small amount of empirical evidence on overlearning, it may be expected that increases in strength of the prepunished behavior