Linguistic Categorization and Hypothesis Testing in a Simple Concept-Identification Task

Abstract

An attempt was made to determine the degree of relationship between 48 Ss' verbal descriptions of three-dimensional, binary-valued stimuli, and the linguistic structure present in the hypothesis-testing sequences when these Ss were presented with the same stimuli in a simple concept-identification task. It was observed that Ss manifested a type of “gambler's fallacy” when testing hypotheses in such situations, and that this imposes a structure on Ss concept-identification performance which, to a large extent, obscures the linguistic patterns which were observed to occur.