by better ones. All 'truths' must be _tested_ by something more
objective than their own self-assertiveness, and this testing by their
working and the consequences to which they lead may go on indefinitely.
In other words, however much a 'truth' has been validated, it is always
possible to test it further. _I.e.,_ it is never theoretically
'absolute,' however well it may practically be assured. For a
confirmation of this doctrine Pragmatism appeals to the history of
scientific truth, which has shown a continuous correction of 'truths,'
which were re-valued as 'errors,' as better statements for them became
available.

It may also be confirmed negatively by the breakdown of the current
definitions of truth, which all seem in the end to mean nothing.

The oldest and commonest definition of a 'truth' which is given is that
it is 'the correspondence of a thought to reality.' But Intellectualism
never perceived the difficulties lurking in it. At first sight this
seems a brave attempt to get outside the circle of thought in order to
test its value and to control its vagaries. Unluckily, this theory can
only assert, and neither explains nor proves, the connection between the
thought and the reality it desiderates. For, granting that it is the
intent of every thought to correspond with reality, we must yet inquire
how the alleged correspondence is to be made out. Made out it must be;
for as the criterion is quite formal and holds of all assertions, the
claim to 'correspond' may be false. To prove the correspondence, then,
the 'reality' would have somehow to be known apart from the truth-claim
of the thought, in order that the two might be compared and found to
agree. But if the reality were already known directly, what would be
the need of asserting an idea of it and claiming 'truth' for this? How,
moreover, could the claim be tested, if, as is admitted, the reality is