Abstract: | Hayes and Barnes‐Holmes (2004) assert that the concept of a topographically unconstrained response class, the concept that carries the explanatory burden of relational frame theory, appeals to no new principles. Operants are properly defined functionally. I argue that they have stretched the concept of the generic response class beyond its appropriate limits. Skinner conceived of response classes as empirically defined units, mutually interchangeable in quantitative functions. The notion of overarching, generalized operants is an uncritical, analogical extension of this concept. I hold that the conceptual work of relational frame theory is incomplete, that a statement of principle is necessary, even if not new. Finally, I distinguish a supposed commitment to a philosophical “mediationism” from a valid inquiry about mediating behavior; that is, behavior with stimulus products that participate in the control of the behavior of primary interest. |