Abstract
To test the conceptual-peg hypothesis of Paivio, two experiments were performed by the use of stimulus-mediation-response (S-M-R) paradigm. In Experiment I subjects were instructed to learn pairs of six nonsense syllables using a picture or a word as the mediating item. The superiority of pictures over words was obtained for stimulus-relevant mediating items. This was consistent with the prediction of the hypothesis. Experiment II assessed the effects of pictures and words as the mediating items on the individual stages of S-M, M-R and S-R connections. The superiority of pictures over words as mediating items was obtained in S-M connection but not in the critical S-R connection nor in the M-R connection. These results were inconsistent with the hypothesis.