Spectral generalization testing with pigeons using brief discrete trials

Abstract

The procedure of Guttman & Kalish (1956) for obtaining generalization gradients in extinction following variable interval (VI) training to a single stimulus value has proved a valuable tool for the analysis of behavior. The length of the VI trial, however, limits the usefulness of this procedure for certain types of questions. An alternative generalization testing procedure, explored in the present study, involves short, closely spaced discrete trials, each trial terminated by a single peck, and employs a probability measure of response. The relatively brief sampling time of the discrete trial procedure permits a more detailed tracking of the behavioral change during the generalization test. The present study was designed to examine the change in extent of generalization during extinction and recovery of responding. MethodTwelve White Carneaux pigeons, maintained at 80% of their ad lib weights, were trained to peck a Plexiglas key illuminated from behind by 550 mf.L from Bausch and . LOmb monochromators. All control, timing, schedule, and analysis functions were provided by a LINC computer operating on-line. Relays controlled by the LINC activated shutters, magazines, and magazine lights. The LINC in addition sensed switch openings resulting from pecks. During the first 14 days of training percent reinforcement was reduced from 100% to 10%, and trial time was reduced from 6 to 3 sec. For the last six days of training the animals were run on 3 sec trials and 10% random reinforcement. Throughout training and testing, trials were terminated either when the 3 sec had elapsed or when a peck occurred.The 11 stimuli used for the generalization test consisted of monochromatic values from 500 to 600 mf.L, at 10 mf.L intervals. Each sequential block of 11 3 sec trials was composed of the 11 test stimuli presented in random order separated by 2 sec blackouts. Four days of generalization testing occurred, conSisting of 4000 trials. Interpolated feedings ("free" feedings) occurred at regular intervals throughout

The Legacy of Guttman and Kalish (1956): 25 Years of Research on Stimulus Generalization

“…The 550-nm asymmetry had also been obtained by Hanson (1959Hanson ( , 1961, Honig, Thomas, and Guttman (1959), and Thomas and King (1959). It was later replicated by Butter (1963), Friedman (1963), Holland and Baker (1968), and Thomas and Switalski (1966). The 555-nm gradient reported by Tomie, Davitt, and Engberg (1976) is also similarly shaped.…”