Classical conditioning of attack behavior in squirrel monkeys.

Abstract

A review of the literature on classical conditioning of pain elicited attack suggested that the use of a bite-tube as a CS should produce strong conditioned attack in Squirrel monkeys. In Experiment I, however, it was found that backward-pairings of shock and the bite-tube produced conditioned attack more reliably than forward pairings. Experiment II demonstrated that forward-pairings could in fact produce strong conditioned attack but that such attack was independent of the duration of postshock attack. Experiment III showed that conditioning via the forward-pairing procedure did not depend upon the occurrence of shock-free periods. These data were compatible with the Pavlovian stimulus-substitution analysis of conditioning. Experiment IV found that conditioning via the backward-pairing procedure depended upon the presentation of trials on a fixed-time schedule. When trials were randomly distributed in time, very few attacks occurred. These data suggested that under the fixed-time schedule attack was elicited by the safety-signal properties of the bite-tube. A detailed examination of the data from all experiments suggested that neither the stimulus-substitution analysis nor the safety-signal analysis provided a consistent interpretation of the data. Finally, an opponent-process model of motivational phenomena recently proposed by Solomon and Corbit (1974) provided a consistent description of the present data and led to a number of testable predictions.