If constancy were a fact of 
knowledge rather than 
perception, the appearance of 
things ought to be governed by 
their retinal stimuli, and the 
achievement of constancy 
should be a corrective added by 
our reason. If appearance were 
governed by the retinal image, 
different objects of equal 
luminance should look as if 
they are the same shade.  But in 
fact, a black object in bright 
sunlight and a white object in 
shadow hardly look alike.  
Furthermore, in the earlier 
photograph of white stairs, the 
stairs in shadow do look like 
the stairs in sunlight, although 
their luminances are quite 
different. And, in the 
illustration to the left, the 
visual angle of the woman and 
man in the left foreground is 
the same as that of the same 
woman and man in the 
background, but they do not 
appear to be the same in size. In 
both these instances, if we saw 
the objects in life rather than 
in pictures, the differences 
would appear even greater.