SEMANTICS: USES AND ABUSES
17
reducible are regarded as necessarily indefinable. We indicate what
they
mean by pointing. This interpretation of the results of semantic
inquiry has led to a number of extreme conclusions of the
u
x
is no–
thing but ..." kind. All abstractions have been considered as verbal
signs of concrete things. When we ask "what do they really mean"
they are to be ultimately discarded for the concrete things themselves.
And every concrete thing turns out to be either a "complex of sense–
data" or "a constellation of electrons" depending upon what the end–
terms are taken to be. On this view, the categories of psychology,
sociology, history, art criticism, etc., are regarded as having definite
meaning only if they can be
defined
without remainder in the terms
assumed to be ultimate.
If
these demands were insisted upon, very little could be mean–
ingfully said. The very demands themselves could not be com–
municated. Now the
sine qua non
of an empirical theory of meaning
is
that assertions must be such that they can be tested, directly or in–
directly, by some set of possible observations. What the nature of
these observations is and the character of the terms with which they
are to be described will be determined by the specific inquiry.
If
we
bear in mind the important point that we never begin an inquiry
from complete scratch, that no one can ask for information unless he
already has some information, that no problem arises save on the basis
of some knowledge accepted as valid, then we will readily see that
many propositions are meaningful which contain expressions not
definable in ultimate terms, whether these terms refer to sense-data
or to electrons. We know that such propositions are meaningful
because we know what to do in order to discover evidence for or
against them. We know that they are meaningful because their use
results in a congruence of behavior between us and others who hear
them or read them. So long as we know the conditions and context in
which an expression is being introduced, we do not need to eliminate
it by substituting a series of end-terms as its defined equivalent. There
is
no objection to saying "the personality (or even soul) of Mr. S.
is
diseased"
if
I can indicate the specific conditions in which his
behavior will be called the behavior of his "personality;" if I can
state the differentiating character between diseased and not-diseased,
and if I can describe what must be done and observed in order to
determine whether his personality is one or the other.
The allied notion that there must be some end-terms which are
intrinsically indefinable, and at whose objective referents we can but
dumbly point, is purely gratuitous. Certain terms need not be defined
in any specific inquiry because we understand them already; but
theoretically, all are definable if they enter into significant discourse.
