Bergson begins just the other way up. He starts from the idea of a
whole field of direct knowledge vastly more extended than the actual
facts of which we are normally aware as making up our direct
experience. He calls this whole field of knowledge "virtual
knowledge." This field of virtual knowledge contains the whole of the
actions and reactions of matter in which our body has its part at any
moment, the multitude of stimulations which actually assail the senses
but which we normally disregard, together with all the responses by
which our bodies adjust themselves to these stimulations, and, in
addition, the whole of our past. For Bergson the problem is to
explain, not how we increase our direct knowledge, but how we limit
it: not how we remember, but how we forget. "Our knowledge," he says,
"far from being built up by a gradual combination of simple elements,
is the result of a sharp dissociation. From the infinitely vast field
of our virtual knowledge we have selected, to turn into actual
knowledge, whatever concerns our action upon things; the rest we have
neglected. The brain appears to have been constructed on purpose for
this work of selection. It is easy enough to show that this is so in
the case of memory. Our past, as we shall show in the next lecture, is
necessarily preserved, automatically. It survives in its entirety. But
it is to our practical interest to put it aside, or at any rate only
to accept just so much of it as can more or less usefully throw 'light
on the present situation and complete it. The brain enables us to make
this selection: it materialises the useful memories and keeps those
which would be of no use below the threshold of consciousness. The
same thing may be said of perception: perception is the servant of
action and out of the whole of reality it isolates only what interests
us; it shows us not so much the things themselves as what we can make
of them. In advance it classifies them, in advance it arranges them;
we barely look at the object, it is enough for us to know to what