Meaning is more than associations: relational operants and the search for derived relations in nonhumans.

Zentall et al. distinguish between concept learning based on physical similarity (perceptual-based), relational-based in terms of relations between and among events, such as that seen on tests of same and difference responding, as well as associative concept learning in which the particular arbitrary stimuli are related along nonformal lines. At stake here is the necessity and usefulness of distinguishing between perceptualand relational-based concepts because both involve responding controlled by the formal properties of the relata. According to RFT, it may be appropriate to consider perceptualand relational-based concept learning as instances of nonarbitrarily relational responding, and associative-based concept learning as instances of arbitrarily applicable relational responding (Dymond & Roche, 2013; Hayes, Barnes-Holmes & Roche, 2001). Nonarbitrary relational responding involves responding based on the nonarbitrary or formal properties of the stimuli being related (e. g., such as the color or shape [perceptual similarity] of the stimuli; see Giurfa, Zhang, Jenett, Menzel, & Srinivasan, 2001; Stewart & McAlwee, 2009). On the other hand, arbitrarily applicable relational responding is “based not on any nonarbitrary or formal relations between the stimuli being related but on aspects of the context that specify the relation such that the relational response can be brought to bear on any relata regardless of their nonarbitrary properties.” (Stewart & McAlwee, p. 312). In this way, both nonarbitrary applicable relational responding and arbitrarily applicable relational responding involve responding to one event in terms of another and are generalizable to novel events. In Zentall et al.’s taxonomy, then, same and difference responding constitutes an instance of nonarbitrary relational responding, along with behavior classified as perceptual-based, while associative concept learning such as that seen on arbitrary match to sample (MTS) tasks may be considered examples of arbitrarily applicable relational responding. Approaching concept learning in this way may prove useful when investigating emergent relations in nonhumans, the interrelationship between nonarbitrary and arbitrary relational responding and the contextual factors responsible for facilitating the transition between types of relational responding.