Retroactive inhibition without intentional new learning: Support for a stimulus-encoding theory

Abstract

The present studies provide additional results concerning an interesting effect discussed in a previous paper (Yang 1981) using a design in which retroactive inhibition (RI) was introduced without the interpolation of a new PA list. The recurrence of the A terms of previously learned PAs during a non-intentional learning interpolation task (NILIT) turned out to be a sufficient condition to give rise to RI. The present experiment 1 provides for a design in which the B terms of original learning (OL) are repeated during NILIT. In spite of the similarity relation between the material used during OL and NILIT, no RI is obtained. The previous and present results suggest that the detrimental effect of similarity on retention only becomes manifest when the common items function as retrieval cues. This suggestion is in line with a stimulus-encoding interpretation of the RI phenomenon. Experiment 2 was executed as a further test of the stimulus-encoding hypothesis. In this investigation the retention of OL is assessed by means of a backward recall procedure. Since under this condition the B terms function as retrieval cues, the stimulus-encoding theory predicts that RI will only be found in the group confronted with the B terms during NILIT, whereas the recall of a group working with the A terms of the PA list will not differ from that of a control group working with a set of new words during NILIT. These predictions are borne out by the results. A further analysis shows that the results discredit two other explanations of RI, viz., the Response-set Interference theory and the Unlearning hypothesis.