Stimulus Generalization as a Function of the Time Between Training and Testing Procedures

Abstract

The present experiment was designed, primarily, to study changes in the stimulusgeneralization gradient as a function of the interpolation of a time interval between training and testing procedures. A second purpose of the study was to explore the retention of a conditioned discrimination, as reflected by changes in the generalization gradient. Considerable evidence (e.g., Hanson, 1957;Honig, Thomas, & Guttman, 1959) has been obtained that discrimination training produces a steepening of the gradient and a shift in the peak of responding away from the SD value. It was anticipated that a decrease in the strength of the discrimination might be reflected by a reversal of these changes, i.e., a progressive flattening of the gradient and shift toward the SD value. METHOD SubjectsThe Ss were 15 White Carneau pigeons maintained by restricted feeding at 70-80% of their free-feeding weights.Apparatus Four Skinner-type automatic key-pecking apparatuses were used. One box was set aside for generalization testing. The light source for the illumination of the key in this box was a Bausch and Lomb monochromator, in which a diffraction grating is used to disperse white light into the spectral continuum. The other three boxes have Bausch and Lomb interference filters. The brightness level and transmission band width of the monochromator and filter colors were approximately equal. In addition, each S received part of its training in the test box so as to have experience with both kinds of stimuli. The apparatus was essentially the same as that described in other reports from the Duke University laboratory. Procedure After magazine training, Ss were trained to peck at the key, which was illuminated by a light of 550 millimicrons. When the response was well-established on a 100% schedule after 2 days of 50 reinforcements each, Ss were placed on a 60-second VI schedule for 2 days. Each training session consisted of 30 stimulus-on periods of 60 seconds, during which the data were recorded. Each of these periods alternated with 10-second blackout periods. In discrimination training the stimuli were changed.After the completion of VI training, discrimination training was begun. For all Ss a light of 550 millimicrons was the positive stimulus, and a light of 570 millimicrons was the negative. Reinforcement for responding to the positive stimulus followed the VI schedule previously used. The positive and negative stimuli were presented according to a prearranged random order. Fifteen 60-second intervals of SD and fifteen of SI were presented each day. These thirty presentations comprised three blocks of ten; and within each block,

STIMULUS GENERALIZATION AS A FUNCTION OF THE DELAY BETWEEN TRAINING AND TESTING PROCEDURES: A REEVALUATION1

“…It is possible that one effect of this training was to increase the "stability" of the generalization gradient over a time period. In addition, Thomas et al (1960) …”

“…Different species were used, a different type of task, a different stimulus dimension, etc. Unlike Perkins and Weyant (1958), Thomas et al (1960) tested for generalization after administering discrimination training. It is possible that one effect of this training was to increase the "stability" of the generalization gradient over a time period.…”

“…Subsequently Thomas, Ost, and Thomas (1960) weeks after reaching a criterion, subjects were tested for generalization carried out in extinction in a manner introduced by Guttman and Kalish (1956). A wide range of spectral stimuli was successively presented with subjects' rate of responding to each providing the measure of generalization.…”

“…It is possible that one effect of this training was to increase the "stability" of the generalization gradient over a time period. In addition, Thomas et al (1960) earliest test was given 24 hr after training was completed. Any changes in the gradient which take place in the first few hours after training would have been overlooked by such a procedure.…”

The recency effect in pigeons’ long-term memory

“…The interaction was not significant [F(2,18) =3.13]. The effect of the different delay intervals is particularly interesting in view of the fact that we have previously found no changes in wavelength gradient slope subsequent to 1 day's delay following single-stimulus training (Thomas & Lopez, 1962) or following experience with a single-discrimination problem (Thomas, Ost, & Thomas, 1960). However, u 10…”