Semantic context and episodic memory

Harris, John Edward (1977) Semantic context and episodic memory. University of Southampton, Doctoral Thesis.

Record type: Thesis (Doctoral)

Abstract

The theoretical and experimental background to Tulving's encoding specificity principle on retrieval cue effectiveness is reviewed and possible theoretical and artifactual explanations are discussed and evaluated.The first study compiled association norms for the material used in Experiments 1-6 by a novel method in which subjects rated word pairs. The results are compared with those of conventional studies.In Experiment I the recall cues were strongly related to the ambiguous (homonym) targets. Hore homonyms were recalled with cues hich bad accompanied them at presentation than with cues equally strongly related to their alternative meetings. The use of act-establishing lists, as employed by Tulving and his colleagues, increased this context effect, but four such lists produced no greater effect than two.Experiment 2 showed the context effect to persist with forced responding, eliminating any explanation in terns of ignoring cues or criterion differences.In Experiment 3 none of the recall cues were presented with the study list. However, more target homonyms were recalled when the recall cues were related to the meanings primed by the presentation context than when they were equally strongly related to the alternative meanings. This suggests a semantic interpretation of encoding specificity.Experiment 4 was a cued recall experiment designed to test explanations of encoding specificity which claim that it results from multiple internal representations (nodes) of ambiguous words. The results are interpreted in terms of episodic theory and as incompatible with multi-node models. However, an alternative interpretation of Experiment 4 holds the results to be compatible with multi-node models. Experiments 6 and 6 were unsuccessful attempts to eliminate this interpretation by replicating the pattern of results of Experiment 4 using context recognition. Features of these recognition paradigms are identified which may have prevented the pattern from emerging. A paradigm which avoids these features is suggested.

This record has no associated files available for download.