Richard Gregory has 
presented striking illustrations 
(such as those shown in the 
photographs to the left) of how 
one might explain this illusion 
in terms of the depth-
processing theory. Consider the 
two photographs as 
representations of the junction 
of several surfaces. The line, or 
shaft, with the normal 
arrowheads seems to represent 
a convex corner with the shaft 
appearing closer than the 
shorter lines that form its 
heads. The line with the 
reversed arrowheads seems to 
represent a concave corner, 
with the shaft appearing 
farther away than the shorter 
lines. Gregory argues that the 
illusion is based on a simple 
application of Emmert’s law: 
The nearer-appearing shaft 
should look smaller than the 
farther-appearing shaft. But 
nearer and farther than what? 
The theory requires that the 
shafts appear nearer and 
farther than each other, but 
Gregory’s argument implies 
only that one shaft appears to 
be nearer than its inducing 
components and that the other 
appears to be farther than its 
inducing components. Some 
students of perception, 
including myself, have thus 
concluded that this theory is 
not applicable to the Müller-
Lyer illusion. An exception to 
this conclusion occurs when 
the two configurations are 
embedded in a scene in which 
each test line has a certain 
position in the third dimension 
relative to the other—–there 
they are nearer and farther 
than each other.