SOLUTION OF THE INTERMEDIATE SIZE PROBLEM BY PIGEONS¹

Abstract

Pigeons learned to respond to the middle-sized member (SD) of a set of three simultaneously presented stimuli with responses to the SD reinforced on a VI 1 schedule. They were then tested for several days with other sets of three stimuli. One procedure presented reinforcements on a VI 1 schedule during the test independent of the stimulus chosen when a reinforcement was programmed. The tests were also given under extinction conditions. With the testing carried out with extinction, preference consistently was for the test stimulus most similar in physical size to the SD. However, when the tests were with reinforcement, random responding resulted. Another effect of testing with reinforcement was an increase in incorrect responding with the training set. Such a test procedure was unsatisfactory for determining the effective aspect of the SD. The conclusion, based on the data of the extinction series, was that pigeons learned the intermediate size problem on the basis of the discrimination of absolute stimulus properties.In the intermediate size problem, an organism learns to respond to the middle-sized member of one set of stimuli and is then tested with a different set. The assumption is that stimulus preference in the test is controlled by the nature of the effective stimulus in the original discrimination. This may be the middle-size relationship, the absolute physical size of the SD, or some contextually determined property (Zeiler, 1963). Since the stimulus responded to in the test would differ depending on which of these properties controlled the discriminative behavior, the essential nature of the stimulus is identified by the test element chosen. The current investigation with pigeons represented the first in which sub-primates learned the intermediate size problem when the stimuli were presented simultaneously. As such, it provided an opportunity to compare the basis of solution of the problem by pigeons and by primates.The same subjects were tested repeatedly with a number of sets. Whenever there is more than a single test trial, a decision must be made about the reinforcement contingen-

SOURCES OF STIMULUS CONTROL DURING MULTIPLE DISCRIMINATIONS1

“…With the 1.4:1 probes, Birds 4 and 6 preferred different stimuli but both responded most to the same one of the two stimuli in each of the two sets. These findings completely replicated earlier results (Zeiler, 1965;Zeiler and Price, 1965) in showing that the similarity of stimuli to S+ controlled the responses.…”

“…Recent research also studied preferences with each test set but supplemented this information with quantitative data to indicate the degree of functional equivalence between the training and the test stimuli (Honig, 1962;Zeiler, 1965;Zeiler and Price, 1965). The experimental procedure enabling the acquisition of this additional information was a variation of the Guttman and Kalish (1956) technique for studying stimulus generalization.…”

“…These data indicated how equivalent the stimuli were to each other. 549 1968,11,[549][550][551][552][553][554][555][556][557][558] NUMBER 5 (SEPTEMBER) Zeiler (1965) and Zeiler and Price (1965) used the above procedure to study equivalence when pigeons learned to respond to one of three stimuli differing in size. Differential responding persisted in the tests and was always to the stimulus closest in size to S+.…”

CONTROL OF A CONTINUOUS RESPONSE DIMENSION BY A CONTINUOUS STIMULUS DIMENSION1

“…Any reinforcement during testing causes problems to some degree. A number of studies (e.g., Zeiler, 1965) have demonstrated that reinforcing responses to novel stimuli on tests with a primary reinforcer tends to bias -further testing. Extinguishing responses on tests also can create problems.…”