Abstract
It is demonstrated that when the two halves of the Müller-Lyer illusion are shown separately to the two eyes this can result in a perception of the binocularly viewed shaft as slanted in depth. For short fin lengths, the magnitude of this slant can be predicted by the extent of the ordinary two-dimensional illusion. This result is discussed in relation to Julesz's claim that stereoscopic matching of the images in the two eyes must precede whatever process is responsible for the Müller-Lyer illusion (whereas the reported illusion seems to imply the reverse). The cyclopean Müller-Lyer demonstration on which Julesz's argument is based is reexamined. It is suggested that the matching of coarse-scale features in the images in the two eyes may help to explain both the cyclopean and the reported three-dimensional Müller-Lyer illusion.
Get full access to this article
View all access options for this article.