In the positivist's project, the task of epistemology consists in logically reconstructing the sentences of theory into those that are grounded in the linguistic framework and those that are not so grounded. Now this procedure can work only when there exists already criteria for dividing a theory into sentences that are related to the framework and the sentences that are not. Without such a criteria, we cannot judge whether the reconstruction is correct or not. The construction will be correct if it clearly demarcates the analytical statements of the system to be true due to the nature of semantical rules and the contingent statements whose truths are not determined by any such rules. But the knowledge of this can be obtained only when we have a general notion of analyticity that defines 'S' as analytic in L for variable S and L. In the absence of such a general notion, a distinction between two types of statements cannot be made. Why is it that the availability of such a general notion of analyticity so important? In a particular reconstruction, semantical rules are constituted with a certain epistemological distinction and thus they cannot define such a distinction. For analyticity to be epistemologically important it must not be defined in terms of any given language. In fact, it should be prior to any such language. Without this, analyticity will turn into a declarative notion having no epistemological significance whatsoever.2 As a result, the whole problem of deciding whether semantical rules in a particular reconstruction correctly characterize all and only analytic statements remains unanswered and unexplained. This way the attempt to define analyticity in terms of semantical rules involves vicious circularity. Analyticity and semantical rules are so interdependent that one cannot be defined without the other.
The above discussion shows that language-epistemology nexus of positivism is inherently defective. The main source of its weakness lies in its conception of language. One of its major assumptions is that language is viewed independently of the experience of the world. As a result, a rigid distinction is created between language, on the one hand, and the world, on the other. This distinction is responsible for giving rise to the dichotomy between two notions of truth, i.e. truth by virtue of meaning and truth by virtue of experience. Quine's rejection of this distinction rests on his fundamental thesis that there is no separate place for linguistic component vis a vis our conceptual scheme due to which we can have truth by virtue of the linguistic component alone.
The next is the assumption central to language/epistemology nexus of positivism, i.e. the primacy of sentence over language. Sentences are taken to be independent units in a language which are true or false on their own. Sentences are thus not seen as forming a system expressing a totality. This gives rise to atomistic picture of language which is to a large extent responsible for the analytic/synthetic dichotomy. The atomistic picture of language is a well-established tradition in contemporary western philosophy defended by Russell, Early Wittgenstein, Carnap and so on. Quine's rejection of analytic/synthetic distinction thus presupposes the rejection of the atomistic view of language which is replaced by the view that language and conceptual scheme together form a holistic system.
| 2 | Carnap, R. (1952). Empiricism, Semantics and Ontology, In Semantics and the Philosophy of Language. Edited by L. Linsk, Urbana: University of Illinois Press. |