Shape constancy: a systematic approach.

Abstract

This work contains the analysis of one experiment and its results. It was carried out to demonstrate that: (i) Shape is a shape-constancy 'factor'. In other words, there must be a 'shape-constancy scale' where the various shpaes can be arranged; (ii) when the shape factor is isolated, the shape constancy of a visual pattern is directly proportional to the strength of the 'linear-perspective cue' and inversely proportional to its degree of 'simplicity'. A pattern such as the cross (which has many linear components and a high degree of complexity) should therefore exhibit a maximum of constancy. The results of the experiment have confirmed the above hypothesis except for an unexpected low degree of constancy of the cross. Another experiment is being performed to explain this last result.