Goldfish, acquiring a passive avoidance response, showed substantially fewer responses in trial intervals than did their yoked controls. A passive procedure, where US reinforcement occurred immediately upon response, produced superior avoidance acquisition to a punishment procedure, where, if response occurred at any time during the trial interval, US reinforcement was administered at the end of the interval. This finding is consistent with the traditional delay of reinforcement gradient. Although goldfish acquired the passive avoidance response, it appeared to be situationally generalized and not stimulus specific, as indicated by a lack of differences between animals trained with a CS and those trained without a CS during trial periods. Also, response suppression between the trial periods was shown by the general ITI response rate suppression. These findings, as well as others showing marked differences between the behavior of goldfish and that of other standard laboratory animals, strongly recommend considerable hesitancy in conclusions of what and how this organism learns.
Bulletin of the Psychonomic Society
Zerbolio, Dominic and Wickstra, L, "Passive Avoidance in Goldfish: Lack of Evidence for Stimulus Specificity" (1978). Psychology Faculty Works. 43.