Commentary: Analysing derived stimulus relations requires more than the concept of stimulus class

The study of derived stimulus relations has led to new and exciting perspectives on the nature of human verbal behavior (Barnes, 1994; Horne & Lowe, 1996; Sidman, 1994) that go beyond traditional perspectives on the topic within behavior analysis. There is a great deal of conceptual and empirical work to be done, however, to compare and contrast the different approaches being taken by behavior analysts to analyze derived stimulus relations. Our purpose in the present paper is to contrast the emphasis of relational frame theory on the concept of stimulus relation (Barnes & Holmes, 1991; Hayes, 1991, 1994; Hayes & Hayes, 1989, 1992; Hayes & Wilson, 1996) with what we see to be a basic commonality between Sidman and Horne and Lowe: the centrality of the concept of stimulus class. The concept of stimulus class has played a successful role in behavior-analytic interpretations of many phenomena, so it is not the utility of this concept in general that we wish to question. Rather, we wish to argue that (a) there are important differences between stimulus classes and stimulus relations, (b) popular research methods do not encourage clarity about this distinction, and (c) the attempt to interpret all derived stimulus relations in terms of the stimulus classes that may result is unnecessarily narrow and limits behavioral approaches to the analysis of language and thinking. Finally, we will suggest methodological alternatives that properly focus attention on the need for relational concepts in the analysis of derived stimulus relations.