Association for Behavior Analysis International

The current symposium will include studies that cover a variety of based and applied issues in the area of relational responding. In the first study, the experimenters evaluated whether the formation of equivalence classes of pictures of faces portraying happiness, neutrality, and anger in individuals with depression, and tested for transfer of function using a semantic differential. In the second study, the experimenters compared "direct" and "derived" extinction of avoidance responses in equivalence classes. In the third study, the authors assessed whether participants used verbal behavior to develop frames of comparison. Finally, the fourth study will review studies related to the application of relational training with individuals with autism.

Researchers and practitioners

A recent experiment with undergraduates verified that equivalence classes comprised of abstract stimuli and faces portraying happiness were more stable over time and exhibited stronger transfer of evaluative functions of faces to abstract stimuli, compared to classes with neutral or angry expressions. The present study is an ongoing replication with undergraduates with a high score in the Beck-depression scale. The Experimental Group (EG) learned matching-to-sample (MTS) relations to establish three four-member equivalence classes. Familiar stimuli from each class were pictures of faces portraying happiness (A1), neutrality (A2) and anger (A3). Remaining stimuli were abstract forms. After MTS tasks, participants evaluated the abstract D stimuli using a Semantic Differential (SD). A 30-day follow-up equivalence test and another evaluation of the D stimuli using the SD will be conducted for the EG. Preliminary results show that nine of the sixteen participants of the EG attained criteria for equivalence class formation. These participants rated D1 as more positive than the evaluation of A1 by the Control Group (participants with low score in the Beck-depression scale) and D3 as more negative than the evaluation of A3 by the Control Group. The maintenance tests are yet to be conducted.

Roche et al. (2008) compared "direct" and "derived" extinction of avoidance response in same/opposite derived relations. They found that "derived" extinction with C1 stimulus was more effective than "direct" extinction with B1 stimulus. However, C1 was presented in extinction more times than B1 and this could explain the observed effect. The present study compared "direct" and "derived" extinction of avoidance responses in equivalence classes, but with the same number of B1 and C1 presentations. In Phase 1, twelve undergraduate students established two four-member equivalence classes (A1, B1, C1, D1 and A2, B2, C2, D2). In Phase 2, B1 was paired with an annoying sound. In Phase 3, participants learned to emit an avoidance response to B1. In Phase 4, the other stimuli were presented to evaluate whether avoidance response would be transfer to stimuli of the same equivalence class. In Phase 5, participants were allocated in the "direct" or "derived" extinction group. Finally, the others stimuli were presented in the extinction transfer test. Only the "direct" extinction group showed transfer of extinction. These results differ from the Roche et al. (2008) findings. Future studies should control the number of stimuli presentation to compare "direct" and "derived" extinction.

The self-emergent properties of human language and cognition can be described in terms of a theory of relational density, with immediate implications for children with autism and related disabilities. The theory postulates that verbal networks demonstrate mass as a function of the molar properties of network volume (i.e., the number of relations) and density (i.e., the strength of relations). Consistent with Newtonian classical mechanics, mass exerts force that operates along with environmental events that can accelerate learning of new verbal relations. We will describe the utility of the theory with this population and present data from several studies that support its application. First, that the acquisition of new verbal skills can be modelled as a function of the volume and density of prior established verbal relations. Second, that by teaching new skills (i.e., increasing network volume and density), the acquisition of later skills is correspondingly increased. Third, that relating in general also exhibits greater acceleration through the direct reinforcement of multiple relational exemplars, or increases in overall relational network volume. We will conclude by describing immediate implications for how language development programming can be conducted considering new data, to accelerate learning through use of established behavior analytic training technologies.