t is _canalized_,
and the visual apparatus simply symbolizes the work of canalizing.
Therefore the creation of the visual apparatus is no more explained by
the assembling of its anatomic elements than the digging of a canal
could be explained by the heaping up of the earth which might have
formed its banks. A mechanistic theory would maintain that the earth
had been brought cart-load by cart-load; finalism would add that it had
not been dumped down at random, that the carters had followed a plan.
But both theories would be mistaken, for the canal has been made in
another way.

With greater precision, we may compare the process by which nature
constructs an eye to the simple act by which we raise the hand. But we
supposed at first that the hand met with no resistance. Let us now
imagine that, instead of moving in air, the hand has to pass through
iron filings which are compressed and offer resistance to it in
proportion as it goes forward. At a certain moment the hand will have
exhausted its effort, and, at this very moment, the filings will be
massed and coordinated in a certain definite form, to wit, that of the
hand that is stopped and of a part of the arm. Now, suppose that the
hand and arm are invisible. Lookers-on will seek the reason of the
arrangement in the filings themselves and in forces within the mass.
Some will account for the position of each filing by the action exerted
upon it by the neighboring filings: these are the mechanists. Others
will prefer to think that a plan of the whole has presided over the
detail of these elementary actions: they are the finalists. But the
truth is that there has been merely one indivisible act, that of the
hand passing through the filings: the inexhaustible detail of the
movement of the grains, as well as the order of their final arrangement,
expresses negatively, in a way, this undivided movement, being the
unitary form of a resistance, and not a synthesis of positive elementary
actions. For this reason, if the arrangement of the grains is termed an
"effect" and the movement of the hand a "cause," it may indeed be said
that the whole of the effect is explained by the whole of the cause, but
to parts of the cause parts of the effect will in no wise correspond.
In other words, neither mechanism nor finalism will here be in place,
and we must resort to an explanation of a different kind. Now, in the
hypothesis we propose, the relation of vision to the visual apparatus
would