Proceedings of the Annual Meeting of the Cognitive Science Society

The ability to recognize and make inductive inferences based onrelational similarity is fundamental to much of human highercognition. However, relational similarity is not easily defined ormeasured, which makes it difficult to determine whetherindividual differences in cognitive capacity or semanticknowledge impact relational processing. In two experiments, weused a multi-arrangement task (previously applied to individualwords or objects) to efficiently assess similarities between wordpairs instantiating various abstract relations. Experiment 1established that the method identifies word pairs expressing thesame relation as more similar to each other than to thoseexpressing different relations. Experiment 2 extended theseresults by showing that relational similarity measured by themulti-arrangement task is sensitive to more subtle distinctions.Word pairs instantiating the same specific subrelation werejudged as more similar to each other than to those instantiatingdifferent subrelations within the same general relation type. Inaddition, Experiment 2 found that individual differences in bothfluid intelligence and crystalized verbal intelligence correlatedwith differentiation of relation similarity judgments.