Often failure or partial 
failure of constancy is to be 
explained in just this way. For 
example, an airplane in the sky 
looks smaller than it would if 
constancy prevailed, and 
surely this is a consequence of 
poor cues to distance (see the 
earlier discussion of the 
elevated moon). The same 
explanation of 
underregistration of distance 
information has also been 
invoked for objects on the 
ground, where more distance 
cues are available. For example, 
when the airplane is seen on 
the ground from roughly the 
same distance as when it was 
airborne, it undoubtedly will 
look larger than it would in the 
sky, but it will also 
undoubtedly look much smaller 
than it would if seen close up. 
However, in experiments 
conducted in recent decades, 
researchers have found that 
observers report very distant 
objects to be even larger than 
they objectively are--an 
observation referred to as 
overconstancy--and that 
constancy often holds at very 
great distances, when distance 
cues are available. How can we 
make sense of these 
contradictory findings and 
explain why objects at great 
distances often look diminutive 
in everyday life?

    We typically do not take 
measurements of perception in 
everyday life, so investigators 
assume that their facts are to be 
trusted more than casual 
observations. However, my view 
is the opposite. We must start 
from observations in daily life 
and, if experiments yield 
different results, we should re-
examine our experiments. 
Overconstancy is a case in 
point.