Blocking and reciprocal blocking in predictive and causal learning

Abstract

The blocking effect (Kamin, 1968) demonstrated that responding to a stimulus was attenuated (blocked) if it was presented in compound with a previously trained stimulus. Prior work with non-human animals, which examined the ability of the blocked stimulus to attenuate responding to the blocking stimulus (Arcediano, Escobar, & Miller, 2004), indicated that in some circumstances blocking is a reciprocal effect. That is, even though the blocking stimulus diminished responding to the blocked stimulus, the blocked stimulus in turn reduced responding to the blocking one. We sought to examine this reciprocal blocking effect in human learning. In a predictive learning task (Experiment 1) and in a causal learning task (Experiment 2) both the blocking and the blocked stimuli were found to mutually compete with each other. This observation seems to indicate that the blocking stimulus loses behavioral control because of its additional training with the blocked stimulus. This finding is not only at odds with the assumptions of traditional and modern associative learning models, but also with recent models of causal learning based on inferential reasoning. However, this finding seems to be consistent with the assumptions of the comparator hypothesis (Miller & Matzel, 1988).