In this discussion, I have 
been referring to knowledge in 
the sense of knowing 
something about the external 
state of affairs being viewed, 
and I have been arguing that 
such knowledge only affects 
perception under special 
circumstances. However, 
knowledge in the sense of 
stored memory of prior 
experience is another story. In 
previous chapters, we have 
seen many examples in which 
memory of prior experience 
affects perception. The 
successful organization and 
recognition of fragmented 
figures, the preference to 
perceive the white region as 
figure in the illustrations of 
the horse and profile, and, 
possibly, the tendency to 
perceive certain line drawings 
(such as those of cubes) as 
three-dimensional—–all are 
instances in which perception 
is enriched by past experience.

    In these cases, I would argue 
that the relevant past 
experience must be in the form 
of visual perceptions that have 
left behind visual memory 
traces. Prior perceptions via 
other sensory modalities will 
not be effective, nor will prior 
nonperceptual experiences. But 
a visual memory of a cubelike 
structure can be effective 
because of its partial similarity 
to the initial stage of 
perception of the line drawing 
seen later. Such similarity is 
crucial for the accessing of the 
appropriate memories and does 
not obtain between the initial 
visual percept and memories in 
other sense modalities or in the 
form of nonperceptual stored 
knowledge. Perceptual 
enrichment based on such 
visual memories is not 
incompatible with what I am 
saying here about the 
limitations of the effects of 
knowledge on perception. Past-
experience effects of prior 
visual memories can occur 
because they in no way do 
violence to the role of the 
stimulus or to other laws of 
perception. Rather, they are 
perfectly compatible alternate 
perceptions of a stimulus that, 
logically speaking, is 
ambiguous.