Abstract
This study compared the efficiency of three procedures in forming stimulus classes. In a "linear" procedure, which had been used in a stimulus equivalence paradigm, stimuli were trained serially by using matching-to-sample (i.e., A [sampIe]→B [comparison], B→C, and C→D for A・B・C・D class). In a "circular" procedure, stimuli were trained circularly (i.e.,A→B, B→C, C→D, and D→A), while in a "symmetry-reinforced" procedure the roles of sample and comparison stimuli alternated with each other (i.e., A→B, B→C, C→D, B→A, C→B, and D→C). Experiment 1 compared the linear with the circular procedure. The proportion of subjects who showed the formation of two classes was larger in the linear procedure group (five of eight) than in the circular procedure group (three of eight). Mean reaction times of the former were significantly shorter than those of the latter. Using a within-subject design with eight subjects, Experiment 2 again compared the two procedures, but no significant differences were found. Experiment 3 compared the symmetry-reinforced procedure with the linear one, using a within-subject design with eight subjects. The number of classes formed was not much different between the two procedures. However, mean reaction times for the former were significantly shorter than those for the latter. The procedure used in stimulus equivalence (i.e., the linear procedure) may be efficient to form stimulus classes, although its efficiency may be limited.
