manner he passes from the lower to the higher, without omitting the
intermediate stages. This appears to be the reason why he seeks for the
definition of knowledge first in the sphere of opinion. Hereafter we shall
find that something more than opinion is required.

False opinion is explained by Plato at first as a confusion of mind and
sense, which arises when the impression on the mind does not correspond to
the impression made on the senses. It is obvious that this explanation
(supposing the distinction between impressions on the mind and impressions
on the senses to be admitted) does not account for all forms of error; and
Plato has excluded himself from the consideration of the greater number, by
designedly omitting the intermediate processes of learning and forgetting;
nor does he include fallacies in the use of language or erroneous
inferences. But he is struck by one possibility of error, which is not
covered by his theory, viz. errors in arithmetic. For in numbers and
calculation there is no combination of thought and sense, and yet errors
may often happen. Hence he is led to discard the explanation which might
nevertheless have been supposed to hold good (for anything which he says to
the contrary) as a rationale of error, in the case of facts derived from
sense.

Another attempt is made to explain false opinion by assigning to error a
sort of positive existence. But error or ignorance is essentially
negative--a not-knowing; if we knew an error, we should be no longer in
error. We may veil our difficulty under figures of speech, but these,
although telling arguments with the multitude, can never be the real
foundation of a system of psychology. Only they lead us to dwell upon
mental phenomena which if expressed in an abstract form would not be
realized by us at all. The figure of the mind receiving impressions is one
of those images which have rooted themselves for ever in language. It may