An analysis of the effects of experimental history on subsequent rule-following behavior

Abstract

Two experiments were conducted in order to identify variables that might interfere with rule-following behavior. Experiment I evaluated the effects of previous exposure to a correspondent rule on subsequent rule-following that was in discrepancy to reinforcement contingencies. Experiment II investigated the effects of previously reinforced behavior on subsequent rule-following. Sixteen university students (eight in each experiment) were exposed to a matching-to-sample procedure. On each trial, one sample and three comparison stimuli were presented to the participant who should select sequentially the three comparisons, according to its dimension in common to the sample: color (C), thickness (E), or form (F). In experiment I, condition I, participants were exposed to the correspondent rule (specifying that selecting the comparison stimuli in the CEF sequence, they would receive points exchangeable for money), and then to the rule discrepant to reinforcement contingency (sequence FCE). In condition 2, they were exposed to the following rule-following sequence : discrepant-correspondent-discrepant. On both conditions, the sequences CEF and EFC were reinforced in a FR 4 schedule. In experiment II, other participants were exposed to minimal instructions during Phase I, to discrepant rule (FCE) during Phases 2 and 4, and to correspondent rule (EFC sequence) in Phase 3. In Phase I, CEF sequence was initially reinforced on a CRF schedule, and gradually changed to FR 4. In Phases 2, 3 and 4, the CEF sequence continued being reinforced in FR 4. In Phase 3, EFC sequence was also reinforced in FR 4, concurrently with sequence CEF. The emission of any other sequence was not reinforced. In Phase I of experiment II, the CEF sequence was established in six participants. Of these, four showed and two did not showed rule-governed behavior in the subsequent phases. The last two participants, upon abandoning rule-following, proceeded to present the previously reinforced CEF sequence. These results suggest that following a rule discrepant to reinforcement contingencies may be maintained without a previous experimental history of correspondent rule-following, and even when control by reinforcement had already been established. However, the data also suggest that control exerted by a previous experimental history of behavior shaped by its consequences may substitute control subsequently shaped by rule- following. These findings have implications for drawing a distinction between behaviors controlled by rules and by contingences.