slightest opening for error. This part of the mechanism of perception
seems to work so regularly and so smoothly that one can hardly conceive
a fault in the process. Nevertheless, a little consideration will show
that even here all does not go on with unerring precision.

Let us suppose that the very first step is wanting--distinct attention
to an impression. It is easy to see that this will favour illusion by
leading to a confusion of the impression. Thus the timid man will more
readily fall into the illusion of ghost-seeing than a cool-headed
observant man, because he is less attentive to the actual impression of
the moment. This inattention to the sense-impression will be found to be
a great co-operating factor in the production of illusions.

But if the sensation is properly attended to, can there be error through
a misapprehension of what is actually in the mind at the moment? To say
that there can may sound paradoxical, and yet in a sense this is
demonstrable. I do not mean that there is an observant mind behind and
distinct from the sensation, and failing to observe it accurately
through a kind of mental short-sightedness. What I mean is that the
usual psychical effect of the incoming nervous process may to some
extent be counteracted by a powerful reaction of the centres. In the
course of our study of illusions, we shall learn that it is possible for
the quality of an impression, as, for example, of a sensation of colour,
to be appreciably modified when there is a strong tendency to regard it
in one particular way.

Postponing the consideration of these, we may say that certain illusions
appear clearly to take their start from an error in the process of
classifying or identifying a present impression. On the physical side,
we may say that the first stages of the nervous process, the due