Conceptual Encoding by Sense Impression

Abstract

Previous research (e.g., Kroes & Libby, 1973) has shown that encoding by sense impression, that is, the classifying of words by the sensory qualities of the stimulus, is not as powerful as classifying by taxonomic or semantic differential characteristics. Because of the consistently poorer performance on sense impression, one may ask whether individuals actually encode by sense impression. Some support comes from Underwood and Richardson (1956) who found that presenting a sense impression category stimulus word such as “round” prompted responses such as barrel, doughnut and so on. From this finding, however, one cannot conclude that words are encoded and actually categorized by their sensory properties. For example, one might touch a satin gown and remark on its softness (an instantaneous sensory impression) but cognitively encode the word “gown” in a taxonomic category such as clothes. The present study examined this problem by use of the concept-shift method. In this method, short lists of symbolic information are presented for future recall. When information presented on succeeding trials is taken from the same category there is a reduction in recall. If information from a different category is presented on a later trial, recall improves to the level obtained in the first trial. This gain on the “shift” trial reflects the fact that the stimulus words are being encoded in a different conceptual category.