Difficulty in documenting class formation in non-human participants may be due to the use of standard training and testing procedures developed in the context of research with human participants. Differences among testing and training situation may produce a decrease in performance during test. Using capuchin monkeys (Cebus spp.) as subject, this study intended to compare acquisition curve of a symmetrical with non-symmetrical conditional relation tasks, with different baselines for each conditional relation trained. This comparative analysis may provide us with some evidence of class formation without the need of the stimulus equivalence standard test protocols. Two male capuchin monkeys (Cebus spp.) participated: a young-adult (M09) and an adult (M12), both with history of simple and conditional discrimination training. Six pairs of bi-dimensional stimuli were used (A1-B1, A2-B2, A3-B3, A4-B4, A5-B5, and A6-B6). The procedure comprised three phases. In Phase 1, a “preparatory” 0-delay arbitrary matching to sample training was carried out (A1-B1 and A2-B2). In Phase 2, a "consistent" or “symmetrical” arbitrary matching to sample training was carried out (A3-B3, A4-B4, B3-A3, and B4-A4). In Phase 3, a "inconsistent" or “non-symmetrical” arbitrary matching to sample training was carried out (A5-B5, A6-B6, B5-A6, and B6-A5). Subject M12 finished all phases of the experiment. The comparative analysis between symmetrical and non-symmetrical acquisition curves (subject M12) shows easier acquisition of symmetrical task than non-symmetrical relations, suggesting that the arbitrary related events are members of class. This data suggest that the comparison between acquisition curves, such as it is described here, is a promising way to evaluate class formation in non-human participants. Subject M09 was removed from experiment because his performance did not reach the criterion in phase 3.1. M09 results pointed the need of a refined control relations analysis during the MTS task, making possible to improve the training procedure.
|
