Proceedings of the Annual Meeting of the Cognitive Science Society

Psychologists have used the semantic fluency task fordecades to gain insight into the processes and representationsunderlying memory retrieval. Recent work has suggested thata censored random walk on a semantic network resemblessemantic fluency data because it produces optimal foraging.However, fluency data have rich structure beyond beingconsistent with optimal foraging. Under the assumption thatmemory can be represented as a semantic network, we test avariety of memory search processes and examine how wellthese processes capture the richness of fluency data. Thesearch processes we explore vary in the extent they explorethe network globally or exploit local clusters, and whetherthey are strategic. We found that a censored random walkwith a priming component best captures the frequency andclustering effects seen in human fluency data.