A theory that can plausibly be 
applied to the Müller-Lyer 
figure concerns eye 
movements. The argument 
initially was that perceived 
extents are a function of the 
distance required to move the 
eyes from one end of the figure 
to the other. Because of the 
direction of the arrowheads, it 
was argued that eye movements 
would differ for the shaft 
lengths that were actually 
equal. No one believes this 
general principle anymore, and 
in fact the geometrical 
illusions occur in exposures 
much too brief to permit any 
eye movement at all.

    A more sophisticated theory 
proposed about eye movement is 
that perceived spatial 
properties, such as the 
separation (or extent) between 
points and the orientation or 
curvature of a contour, are a 
function of what "commands" a 
higher center in the brain 
issues to the eye muscles. 
(Information we obtain about 
what the eyes are doing derives 
not from proprioceptive 
feedback from the eye muscles 
but from a record of commands 
to the eye muscles.) Leon 
Festinger of the New School for 
Social Research once 
maintained that such eye 
muscle commands determine 
what we perceive.