Certain findings that at first 
seem to challenge the 
inference theory can in fact be 
reconciled with it. For 
example, Paul Kolers and James 
Pomerantz, then at the Bell 
Telephone Laboratories, 
demonstrated that apparent 
movement is easily seen when 
the two separate images are 
different shapes. If A is a circle 
and B a triangle, observers will 
perceive the circle changing 
shape as it moves, becoming a 
triangle by the time it reaches 
B. This effect suggests that the 
sudden disappearance of the 
object at A and the sudden 
appearance of an object at B 
creates such a strong 
presumption of movement from 
A to B that the movement is 
perceived despite the dissimilar 
shapes. The perceptual system 
thus seems to account for the 
dissimilarity in an ingenious 
way, by perceiving that the 
object deforms as it moves.

    One final experiment in 
support of the inference view is 
worth mentioning. Based upon 
an earlier finding by Arnold 
Stoper, then at Brandeis 
University, Eric Sigman and I 
performed the experiment 
illustrated at left. Two spots 
appeared and disappeared at 
spacings and intervals that 
normally would yield an 
impression of apparent 
movement. However, this 
appearance and disappearance 
was achieved by moving an 
opaque rectangle back and 
forth over the spots. When the 
rectangle itself was not visible, 
observers perceived the spots 
moving back and forth, as 
would be expected. When the 
rectangle was visible, however, 
they did not perceive the spots 
to be moving. Instead, they 
perceived them to be 
permanently present but 
alternately covered and 
uncovered by the rectangle. 
Although images of the spots 
stimulated the retina at the 
appropriate spacing and timing, 
the sequence did not yield the 
illusion of apparent movement. 
This finding suggests that we 
ordinarily perceive apparent 
movement not because it is an 
inevitable sensory outcome of 
stimulation but because it is 
the best explanation of the 
otherwise inexplicable sudden 
appearance and disappearance 
of objects. But in the present 
experiment another solution is 
available, namely permanently 
present spots undergoing 
covering and uncovering.