174 J.J. Haldane
archaeology aims to understand human history and not just to chart the causal
impact of those long dead, it requires the archaeologist imaginatively to
‘relive’ (nacherleben) the past by interpreting cultural and personal meanings.
There is an important counterpart to this requirement in semantic theory.
Contemporary philosophies of language divide into two broad camps. On
the one hand there are those which aim to give the meaning of speech acts
either by relating them to their behavioural causes and effects or by calibrat-
ing them against the states of the world with which they are correlated. Thus
one might hold that the meaning of the uttered sentence ‘snow is white’ is to
be given by the causes and effects of its utterance, or by specifying the con-
ditions under which it would be true (or by combining these two aspects).
Though theories of these two sorts differ from one another, they have in
common the assumption that it is possible to give the content of an utterance
by identifying something outside the sphere of meaning – behaviour or states
of the world.
Sometimes this attempt is two staged. For example, among those who
think that meaning is fixed by causes and effects, some attempt to specify the
content of a speech act by reference to the (typical) beliefs and communicat-
ive intentions of speakers who use it.^3 There are difficulties in identifying the
relevant psychological states (for speakers can be dishonest, distracted or
confused) but even if this can be done the account is regressive. If we say that
the utterance ‘snow is white’ means snow is white if and only if the speaker
believes that snow is white, intends to communicate this to a hearer and has
certain expectations about the hearer’s powers of understanding, then we still
have to say what it is to have beliefs, intentions and expectations with these
contents. The naturalist’s task is not complete until meaning has been explained
in terms of non-semantic(-cum-intentional) causes and effects.
Ingenious as they often are, theories of the sorts mentioned thus far invaria-
bly founder on the simple fact that nothing short of understanding its sense
can amount to grasping the meaning of an utterance. Non-intentional causes
and effects underdetermine meaning as do correlated states of affairs. This
gap can be demonstrated in various ways, but referring back to my own earlier
discussion (chapter 2, p. 107) the point can be made by observing that
descriptions outside the sphere of meaning are extensional while those inside
it are intensional. The diagram to which someone is causally related who uses
the sentence ‘that is a triangle’ to refer to a figure on the board, is also a
trilateral; but the meanings of the sentences ‘that is a triangle’ and ‘that is
a trilateral’ differ. This difference is one of sensenot causal influence, exten-
sion or truth conditions.
How then should we understand meaning? The question is ambiguous,
for it may be read metaphysicallyas asking how meaning is possible; or
epistemologicallyas asking how we can know what an utterance means. The