Multidimensional Scaling and Experimental Aesthetics: Escher's Prints as a Case Study

Abstract

This article represents a case-study in the applicability of multidimensional scaling (MDS) to experimental aesthetics. MDS allows experimenters to ask observers to compare stimuli without specifying which criteria they should use. MDS plots data-the similarity judgements-in a "psychological space" whose number of dimensions reflects the number of ways the stimuli are perceived to differ. An experiment attempted to tap into the criteria used by non-specialists when viewing prints by the Dutch artist M. C. Escher, and to stUdywhether or not these criteria correspond to those Escher himself used to classify his work. Forty observers were shown every possible non-duplicate pair combination of twenty-six slides of Escher's prints, and were asked to rate the similarity of each pair, using any criteria they wished. Three dimensions were identified in this experiment ("dimensionality," "shape" and "degree of realism"). A neighborhood analysis (based on octants) suggests that observers' perceptions seem to correspond to Escher's groupings.