invalidating knowledge can only serve to separate it from
incidental errors and to disclose the relative importance of truths.
Science would then be rehabilitated by criticism. The primary movement
of the intellect would not be condemned by that subsequent reflection
which it makes possible, and which collates its results. Science, purged
of all needless realism and seen in its relation to human life, would
continue to offer the only conception of reality which is pertinent or
possible to the practical mind.

We may now proceed to discuss these various attitudes in turn.

[Sidenote: Science its own best critic.]

A first and quite blameless way of criticising science is to point out
that science is incomplete. That it grows fast is indeed its commonest
boast; and no man of science is so pessimistic as to suppose that its
growth is over. To wish to supplement science and to regard its
conclusions as largely provisional is therefore more than legitimate. It
is actually to share the spirit of inquiry and to feel the impulse
toward investigation. When new truths come into view, old truths are
thereby reinterpreted and put in a new light; so that the acquisitions
of science not only admit of revision but loudly call for it, not
wishing for any other authority or vindication than that which they
might find in the context of universal truth.

To revise science in this spirit would be merely to extend it. No new
method, no transverse philosophy, would be requisite or fitted for the
task. Knowledge would be transformed by more similar knowledge, not by
some verbal manipulation. Yet while waiting for experience to grow and
accumulate its lessons, a man of genius, who had drunk deep of
experience himself, might imagine some ultimate synthesis. He might
venture to carry out the suggestions of science and anticipate the
conclusions it would reach when completed. The game is certainly
dangerous, especially if the prophecy is uttered with any air of
authority; yet with good luck and a fine instinct, such speculation may
actually open the way to discovery and may diffuse in advance that
virtual knowledge of physics which is enough for moral and poetic
purposes. Verification in detail is needed, not so much for its own sake
as to check speculative errors; but when speculation is by chance well
directed and hits upon the substantial truth, it does all that a
completed science would do for mankind; since science, if ever
completed, wo