Ensino de discriminações condicionais por imitação: efeitos de dois arranjos de apresentação de tentativas.

Abstract

Imitation can be defined, in general terms, as the similarity of behavior between two organisms when the response topography of one, the imitator, is similar to nd controlled by the behavior of the other, the demonstrator. Along the years there has been a lot of controversies involving imitation`s conceptualizations and discussions on how it is produced and maintained. Some authors say that imitation, as any other behavior, needs to be directly reinforced in all its instancies to be established and maintained. Other authors emphasize the need of only some responses being reinforced, thus generalizing to others. Independently of the presence of direct reinforcement, imitation has been used as a teaching technique for many behaviors. Most behaviors that are taught through imitation are simple motor behaviors. Researches that use imitation as a teaching technique for complex behavior, such as conditional discrimination, are still new and raise lots of questions for investigation. One of these questions is the possible influence of trial arrangements on the participants` performance, when analyzing numbers of session to criterion. Research protocols on teaching conditional discriminations through imitation present, unsystematically, either one trial to the observer execute after each one presented by the demonstrator, or many trials to be executed by the demonstrator, for then, the observer do his. In this study, it was tried to compare these trials arrangements systematizing two protocols to be used with all subjects. Each set of pairs of conditional relations (BA or CA) was taught in either one of these ways: 1) for each trial demonstrated by the demonstrator, the observer executed one; 2) for each block of trials demonstrated, the participant executed another block of trials. Another difference from the protocols that have already been studied, was the presence o some identity trials with programmed consequences during the arbitrary matching training. As general results, all participants learned the conditional discriminations programmed by the study (BA and CA) and showed the emergence of equivalence classes. independently of trial arrangement and without planned consequences for the trials. This learning may be result of generalized imitative responses that could be maintained through other reinforced imitative responses that were not the arbitrary conditional discriminations (BA and CA). The large number of trials needed to criteria, may, however derive from this lack of reinforcement. In quantitative terms, trial arrangements did not result in different performance. However, data show that, when the trials were presented in an alternate way, the participant s behavior seemed to be only under the control of the comparison stimuli chosen by the demonstrator, without attending to the sample stimuli. Even though the results obtained by the two types of trial presentation were quantitatively similar, questions were raised on which trial arrangement would establish the best stimulus control when imitation is used as a teaching technique. This issue is specilly important because the better established the stimulus control is, the more effective is the conditional discrimination learning.