rse in the same manner: the qualities that interest one are unnoticed
by another.[3]

The image being a simplification of sensory data, and its nature
dependent on that of previous perceptions, it is inevitable that the
work of dissociation should go on in it. But this is far too mild a
statement. Observation and experiment show us that in the majority of
cases the process grows wonderfully. In order to follow the progressive
development of this dissolution, we may roughly differentiate images
into three categories--complete, incomplete, and schematic--and study
them in order.

The group of images here termed _complete_ comprises first, objects
repeatedly presented in daily experience--my wife's face, my inkstand,
the sound of a church bell or of a neighboring clock, etc. In this class
are also included the images of things that we have perceived but a few
times, but which, for additional reasons, have remained clean-cut in our
memory. Are these images complete, in the strict sense of the word? They
cannot be; and the contrary belief is a delusion of consciousness that,
however, disappears when one confronts it with the reality. The mental
image can contain all the qualities of an object in even less degree
than the perception; the image is the result of selection, varying with
every case. The painter Fromentin, who was proud that he found after two
or three years "an exact recollection" of things he had barely noticed
on a journey, makes elsewhere, however, the following confession: "My
memory of things, although very faithful, has never the certainty
admissible as documentary evidence. The weaker it grows, the more is it
changed in becoming the property of my memory and the more valuable is
it for the work that I intend for it. In proportion as the exact form
becomes altered, another form, partly real, partly imaginary, which I
believe preferable, takes its place." Note that the person speaking thus
is a painter endowed with an unusual visual memory; but recent
investigations have shown that among men generally the so-called
complete and exact images undergo change and warping. One sees the truth
of this statement when, after a lapse of some time, one is placed in the
presence of the original object, so that comparison between the real
object and its image becomes possible.[4] Let us note that in this group
_the image always corresponds to certain individual objects_; it is not
the same with the other two groups.