This article presents two case studies in phonetic symbolism (or iconicity), opposing the statistical to the structural-cognitive approach. The former is frequently based on Stimulus–Response (S–R) Questionnaires, the latter attempts to reconstruct a plausible “mediating structure” between Stimulus and Response, based on hypotheses derived from empirical and theoretical studies in such disciplines as psychology, linguistics, psycholinguistics, aesthetics, and the philosophy of science. The thesis of this article is that rigorous statistical research that is not preceded by proper theoretical analysis is bound to yield inadequate results. It points out some of the fallacies of two statistical studies, and then it shows how a structural-cognitive approach would treat the same issues. I place my argument within a comprehensive theory of human information processing—phonetic, semantic, and visual.

You do not currently have access to this content.