Effect of training the correspondence to say-to do, to say-to describe, and to do-to describe on the acquisition, generality and maintenance of a conditional discrimination task in humans

Abstract

In this experiment, a second-order matching-to-sample procedure was modified to allow saying-doing, saying-describing, and doing-describing correspondence training. Trials included three types of screens: self-instructions, matching, and descrioptions. When subjects were trained on matching to sample before receiving correspondence training, the reinfrocement of one type of correspondence established all three types; subjects were able to select a self-instruction that corresponded to their matching performance and descriptions. However, starting the experiment with correspondence training seemed to interfere with task acquisition. Results suggest that the establishment of a correspondence between self-instruction and performance, performance and description, or self-instruction and description is an example of functional dominance, within the context of a task, of the reinforced elements over the others.