Page 116 - Encyclopedia of Philosophy of Language

Truth and Meaning
that a bachelor is an unmarried man then we lack a involved, would be committed to taking the statement grasp of the concept of a bachelor. There are other to express a truth.
analytic propositions which, though not self-evident,
weknowaprioribecausethereisproofofthemfrom 4. TheAttackonAnalyticity
self-evident premises which we know a priori. The
ability to appreciate the validity of such proofs is to
be explained in terms of abilities which constitute a
grasp of the relevant concepts. Having thesecon- cussion assumes a distinction between analytic ceptual abilities will not guarantee that we canpro-
duce proofs on demand. But the account has it that recognition that a proof is valid is explained by the interlocking exercise of conceptual abilities.
This application of the traditional view has the vir- tue of making a priori knowledge seem unmysterious. If it is along the right lines then it can be explained how we know some things to be true independently of experience, that is, in the absence of empirical groundsforbelievingthattheyaretrue.However,the account has to be supplemented if it is to show how a priori knowledge can be squared with classical empiricism. What is missing is an explanation of how it can be that we have a priori knowledge if all knowl- edge is in some sense derived from experience. In classical empiricism the required supplement is pro- vided by a theory of concepts. The theory has it that no concept is innate and that all concepts are acquired via the impact of experience. So, for example, we
would acquire the concept of redness by a process of abstraction from properties of the visual experiences we have as we look at red things. Thus even a priori knowledge would be derived from experience insofar as concepts are derived from experience.
3. AnalyticityandSemantics
No unmarried man is married
(1)
The concept of analyticity is of interest aside from its
links with the empiricist doctrines outlined above for
it has a place within a broad conception of what is
involved in understanding the meanings of
expressions in natural languages. There is some plausi-
bility in the idea that if one understands the English
term garage then one has a mastery of certain rules
or conventions governing the use of the term. For
example, there might be rules which require that if
garage applies to something then place for storing or
repairing motor vehicles also applies to it, and vice
versa. The totality of rules governing a given term
would determine its meaning and thus what concept
it expresses. This conception can be extended to cover
logical expressions like and,either... or... , and the
logical concepts linked with these expressions. In the
case of and there might be a rule which requires that
if you take a conjunction of the form P and Q to
express a truth then you must take P to express a truth
and Q to express a truth. An analytic statement would
be a statement the rules for whose constituent lexicalitem.
expressions are such that anyone who has a mastery In writings subsequent to 'Two dogmas' Quine of these rules, and thus understands the expressions attempts to make sense of analyticity in behavioral
94
In a classic article entitled 'Two dogmas of empiricism' (in Quine 1961) W. V. O. Quine argues that the con- cept of analyticity is irredeemably obscure. The dis-
statements which are logical truths like
and analytic statements which are not logical truths like
No bachelor is married (2)
Logical truths are statements which are true and remain true under all interpretations of their non- logical vocabulary. Nonlogical analytic statements can be turned into logical truths by substituting syn- onyms for synonyms. So (2) can be turned into (1) by substituting unmarried man for bachelor. Much of the discussion of Two dogmas' focuses on the idea of nonlogical analyticity. A major theme is that concepts commonly used to explain this notion are no clearer than what they are meant to explain, and further, that when one tries to elucidate them one finds oneself in turn falling back on the notion of analyticity. For example, the notion of synonymy just used to account for the analyticity of (2) is, according to Quine, obscure and its obscurity is not removed by account- ing for the synonymy of bachelor and unmarried man in terms of the analyticity of All and only bachelors are unmarried men. (This line of thought is submitted to close scrutiny in Grice and Strawson 1956.)
Quine is equally pessimistic about the idea of deploying the concept of a semantic rule to account for analyticity. Here the main target is work of Carnap on artificial languages (Carnap 1956). Carnap devised for artificial languages of a certain type a system of semantic rules which, roughly speaking, combine to fix the truth conditions for the sentences of these lan- guages. It turns out that there are certain sentences whose truth is guaranteed by the relevant semantic rules. These are the 'L-true sentences.' Quine argues that L-truth fails to provide the required elucidation for analyticity since, among other things, a satisfying theory of semantic rules is not available. J. J. Katz has developed an account of analyticity within the context of a general semantic theory which represents the meanings of lexical items in terms of semantic markers denoting conceptual constituents of mean- ings (see Katz 1972 for discussion and furtherref- erences). From Quine's standpoint, however, an account is still needed of what determines which con- ceptual constituents should be assigned to a given