Association for Behavior Analysis International

Recent empirical and conceptual developments (e.g., Urcuioli, 2008; Wasserman & Frank, 2005) have suggested that the failure to observe derived relational responding in nonhuman subjects in the result of the procedural factors such as the conditions of training and testing and not other factors that may limit learning to simple associative relations. This symposium proposes to bring together research presentations from three different laboratories investigating the particular role of procedural and other factors in precluding or facilitating the development of (derived conditional relations) in nonhuman subjects. Velasco and Tomanari ask if conditions of extinction arranged during typical tests are responsible for failures to find robust evidence of equivalence-consistent responding in pigeons. Swisher and Urcuioli ask if the fact that spatial location can become a part of the functional stimulus complex is necessarily a hindrance to the development of symmetrical relations in pigeons. Finally, Hinnenkamp & Vaidya examine the role that the accuracy of baseline relations plays in the development of symmetrical relations by manipulating a variety of factors designed to improve baseline accuracy prior to testing. Galizio will articulate the central theme of the papers and offer broader insights about the endeavors and the approaches.

Extinction often disrupts nonhuman's performance during equivalence tests. In our laboratory, we have worked on strategies to assess emergent relations under reinforcement conditions. In an ongoing study, after pigeons learn 2 arbitrary-matching tasks (AB | CD), they are given a reinforced symmetry test for half of the baseline relations (B1A1 | D1C1). To show that testing performances are not a result of rapid learning due to reinforcement, two novel relations are concurrently reinforced using the baseline stimuli (D2A2 | B2C2). Because novel relations are not inconsistent with the initial training, they can be incorporated on the baseline, thus enabling the assessment other emergent relations. For example, the baseline relation A2B2 along with the novel relation B2C2 allow for the assessment of A2C2 transitivity and C2A2 equivalence. To control for the reinforcement during testing, 2 novel relations are concurrently reinforced (A1C1 | C1A1). Results of a previous study in which 1 pigeon shows clear evidence of symmetry and 2 show questionable evidence of symmetry using a similar strategy encourage the prediction that equivalence can emerge in pigeons under the experimental conditions employed in our ongoing research.

Testing for evidence of symmetry with pigeons in 2 alternative matching-to-sample has met with limited success, but go/no-go procedures have yielded compelling evidence for this property of equivalence. In these procedures, stimuli are presented on the same key, avoiding any possible disruptive effects of spatial location. Could fixed spatial location be why results consistent with symmetry are more likely with go/no-go procedures? Six pigeons in a control group saw both color sample and form comparison stimuli presented on the left key in 3-key chambers. Six birds in an experimental group saw sample stimuli on the center key and comparison stimuli on the left key to assess the effect of varying location on responding to sample-comparison combinations on symmetry tests. Two birds in the control group and 2 in the experimental group responded significantly more to comparison stimuli on positive symmetry probe trials than on negative symmetry probe trials. These results confirm a seemingly counter-intuitive prediction from Urcuiolis (2008) theory that as long as the functional stimuli, which allow for class merger, are the same across the baseline relations, then the spatial locations of the sample and comparisons can differ.

The go/no-go procedure has recently attracted attention for its use in establishing symmetrical relations/responding in pigeons (Frank& Wasserman, 2005; Urcuiolli, 2008). In our attempts to utilize the go/no-go procedure to systematically replicate Urcuioli, we found that several of the pigeons used within our experiment failed to respond in accord with the experimenter defined baseline conditional relations, despite extensive training with visual stimuli. As a result, components of the go/no-go procedure were systematically manipulated to bring pigeons responding in line with the experimenter defined baseline conditional relations. Our results will illustrate the effects of inter-stimulus interval manipulations, DRO manipulations, and reinforcement density manipulations on visual sample, visual comparison go/no-go discrimination ratios. Furthermore, our results will also discuss the effects of using cross modal sample and comparison stimuli in the establishment of baseline conditional discriminations and derived symmetrical relations.