Effects of conditional discrimination training on the acquisition of symmetry and transitivity relations

Abstract

The present experiment analyzed the effects of conditional discrimination training (e.g., AB and BC) on the acquisition of symmetry (BA and CB) and transitivity relations (AC and CA). An arbitrary matching-to-sample task involving 9 four-stimulus sets was employed. Four normal adults were exposed to the training of 48 conditional discriminations distributed in three different groups. Group I included all possible 24 conditional relations established among the stimuli in A-B-C sets, that is, the AB and BC relations and their symmetric BA and BC and transitive AC and CA counterparts. Group II included only the relations A’B’ and B’C’. Group III included only B”A”, C”B”, A”C” and C”A” relations. Under continuous reinforcement conditions, the procedure compared the acquisition of conditional discriminations with (Group I) and without (Groups II and III) symmetric counterparts. Main results showed that the conditional relations were acquired earlier in training and accumulated fewer errors when accompanied by their symmetric counterparts (Group I versus Groups II and III). Similar effects were not observed on the training of transitivity relations, possibly indicating that different variables controlled symmetry and transitivity. These results suggest methodological alternatives to the procedures that typically evaluate conditional discriminations in the absence of reinforcement.