Development of Response Equivalence among Dissimilar Verbal Units

Abstract

Two experiments were conducted to test the hypothesis that response generalization could be induced among unrelated nonsense syllables if, prior to the test for generalization, these responses had typically produced similar outcomes (in terms of reinforcement vs no reinforcement) from one stimulus situation to the next. In Exp. I, Ss first learned to associate two nonsense-syllable responses with a common set of four one-digit stimuli; a second pair of syllables was associated with a different set of one-digit stimuli. When this preliminary task had been mastered, Ss were presented with a second list in which the original stimuli (digits) and responses (nonsense syllables) were re-paired. An analysis of errors on this second (test) list indicated that the preliminary task served to increase confusions between responses that had initally been associated with the same stimuli (digits). Exp. II demonstrated that such confusions could be temporarily enhanced by increasing the number of stimuli initially associated with the response elements that were to achieve equivalence.