Pseudoavoidance responses in two-way avoidance learning

Abstract

In a two-way avoidance learning task, responses sometimes occur during the CS-UCS inter­ val prior to the initial receipt of shock. The number of such pseudoavoidance (P A) responses was found to be related positively to the number of avoidance responses made after the receipt of shock, implying that some of these avoidance responses were artifactual. The use of pro­ cedures that would decrease the occurrence of P A responses would also minimize this possible contamination of the avoidance data. Evidence was presented indicating that the frequency of P A responses was decreased when some amount of exploration of the apparatus, rather than none, was given prior to avoidance training, the shuttle compartments were separated by a small rather than a large guillotine door. a neutral rather than an intense auditory CS was used. and the CS-UCS interval was short (5 sec) rather than long (10 sec). In previous investigations of two-way avoidance learning in this laboratory, it was rare to observe a sub­ ject locomote from one compartment to the other during the CS-UCS interval before the ftrst shock was received. The compartments of the apparatus employed in these studies were separated by a small guillotine door resting on a hurdle, and the procedure involved some exploration of the apparatus prior to avoidance training. When the size of the door and the exploration treatment were altered in a recent study (McAllister, McAllister, Dieter, & James, in press, Experiment 2), a large number of such responses were observed. The purpose of this paper is to present the data from the above study that relate to such pseudoavoidance (P A) responses, as well as to identify in other studies the experimental conditions that have led to such responses. Knowledge of the variables leading to PA responses would enable investigators to choose conditions that minimize their occurrence. This is important because, although a PA response does lead to the nonoccurrence (avoid­ ance) of shock, it clearly is not made as a consequence of learning. Within two-factor theory, an avoidance response is assumed to be an escape-from-fear response. Obviously, a response made during the CS-UCS interval prior to the initial receipt of shock cannot be considered an escape from conditioned fear. Similarly, from a cognitive viewpoint (e.g., Seligman & Johnston, 1973), an avoidance response made prior to the receipt of the ftrst shock cannot be the result of the subject's expect­ ancy that shock will occur if the response is not made.