Should Visual Images Be Defined as Depictive Neural Patterns, or as Conscious Sensations Constructed from “Imageless” Visual Thoughts?

Abstract

Greater vividness of visual imagery correlated significantly with improved performance on a task in which 112 subjects generated and tested hypotheses about the rules of three-point perspective (3PP), but not with improvement on a similar task in which another 116 subjects applied previously known rules of two-point perspective (2PP). Given the 100-year history of research finding zero correlation between image vividness and visuospatial reasoning, the 2PP results are interpreted as support for Külpe's thesis that the imageless thinker's existing knowledge of imageless rules allows for the solution of spatial problems, without any construction of visual images. The 3PP results are interpreted as support for Kunzendorf and Reynold's thesis that the vivid imager's construction of test images from newly hypothesized rules allows for the refinement of those imageless rules, with less reliance on overt mistakes for corrective feedback.