摘要：Throughout the 25-year history of research on stimulusequivalence, one feature of the training procedure has remainedconstant, namely, the requirement of operant responding duringthe training procedures. The present investigation compared thetraditional match-to-sample (MTS) training with a more recentrespondent-type (ReT) procedure. Another consistent feature ofthe equivalence paradigm is the apparent stipulation that bothtraining and testing must occur before equivalence emerges. Inthis respect, a more idiosyncratic measure of class acquisitionwould be desirable. Multidimensional scaling, as a class ofexploratory techniques, is introduced as a possible addition to thestimulus equivalence paradigm.Results from 35 subjects in 3 experiments suggest that whilethe respondent-type training method can be an effectiveprocedure, the operant-based match-to-sample method wasclearly more effective in tests for symmetry, equivalence, andextended equivalence. The addition of a scaling procedure provedvaluable and showed that both training methods facilitated theemergence of derived relations to varying degrees. Results areevaluated in relation to the importance of broadening thenecessary and sufficient training conditions and responserequirements for the emergence of stimulus equivalence.
