Common coding in symbolic matching tasks with humans: training with a common consequence or antecedent.

The present experiments explored the importance of training with a common antecedent or consequence for common coding in symbolic matching-to-sample tasks with human subjects using reversal and transfer tests. Experiment 1 assessed common coding following training in many-to-one (MTO) and one-to-many (OTM) symbolic matching-to-sample tasks by comparing learning of partial and total reversals of these contingencies. Experiment 2 assessed common coding in MTO and OTM tasks by comparing performances in transfer tests when the common coding relations in the transfer phase were either consistent or inconsistent with those present during initial training. Evidence for common coding in the MTO and OTM discriminations was obtained in both experiments, although the amount of common coding was greater in the MTO discrimination in Experiment 1. These data are discussed in terms of associative mediational approaches to common coding phenomena. It is suggested that a basic requirement of an associative theory is that common coding should be more likely to occur when the stimuli are trained with a common consequence than when they are trained with a common antecedent, and that this requirement has been at least partly satisfied in the present studies.