Establishment of an avoidance gradient under latent-learning conditions.

Abstract

The present expt. constitutes a test of the hypothesis that transfer of learned avoidance behavior may be mediated by indirect activation of memory traces under conditions which preclude explanation in terms of stimulus generalization, fractional anticipatory goal responses, or delay of reinforcement. The functional relationship between the magnitude of this transfer effect and directness of trace connection was also investigated. Four groups of 20 rats each were given 6 daily nonreinforced exploration trials in a runway divided by doors into 6 perceptually distinct compartments. After training, each rat was introduced directly into 1 end compartment, shocked, and removed; then after 15 min. was reintroduced into 1 of the 4 middle runway compartments. (N=20 in each compartment.) Initial direction of exit from these compartments and other behavior occurring during the 2.5-min. test period were recorded. Statistically significant evidence of transfer of avoidance behavior as indicated by initial direction of exit was obtained. Controls indicate the inapplicability of the alternative explanatory principles mentioned above. The hypothesis that the magnitude of the transfer effect is a positive function of the directness of connection between the compartment perceived on test and the end shock compartment (and between the corresponding memory traces) was confirmed by the differential avoidance behavior of the exptl. groups.