seriously believe that there are atoms, but in their ignorance
of the precise composition of matter, they find it convenient to speak
of it as if it were composed of indestructible particles. But for
Democritus and for many modern men of science the atomic theory is not
a working hypothesis merely; they do not regard it as a provisional
makeshift; they regard it as a probable, if not a certain,
anticipation of what inspection would discover to be the fact, could
inspection be carried so far; in other words, they believe the atomic
theory is true. If they are right, the validity of this theory would
not be that of pragmatic "truth" but of pragmatic "fact"; for it would
be a view, such as memory or intuition or sensation might give us, of
experienced objects in their experienced relations; it would be the
communication to us, in a momentary dream, of what would be the
experience of a universal observer. It would be knowledge of reality
in M. Bergson's sense. Pragmatic "truth," on the contrary, is the
relative and provisional justification of fiction; and pragmatism is
not a theory of truth at all, but a theory of theory, when theory is
instrumental.

For theory too has more than one signification. It may mean such a
symbolic or foreshortened view, such a working hypothesis, as true and
full knowledge might supersede; or it may mean this true and full
knowledge itself, a synthetic survey of objects of experience in their
experimental character. Algebra and language are theoretical in the
first sense, as when a man believes in his luck; historical and
scientific imagination are theoretical in the second sense, when they
gather objects of experience together without distorting them. But it
is only to the first sort of theory that pragmatism can be reasonably
applied; to apply it also to the second would be to retire into that
extreme subjectivism which the leading pragmatists have so hotly
disclaimed. We find, accordingly, that it is only when a theory is
avowedly unreal, and does not ask to be believed, that the value of it
is pragmatic; since in that case belief passes consciously from the
symbols used to the eventual facts in which the symbolism terminates,
and for which it stands.

It may seem strange that a definition of truth should have been based
on the consideration of those ideas exclusively for which truth is not
claimed by any critical person, such ideas, namely, as religious myths
or the graphic and verbal mac