It is, of course, easy to forget that semantics has many precisely distinguished subdivisions—and, so forgetting, to submit irrelevant criticism.

It would be idle, for instance, to complain that in pure semantics the crucial terms receive no definition, being introduced into the discussion by means of rules or postulates which determine only how the words are to be used in the subsequent deductive elaboration of a system defined by those very same rules; or to lament the fact that in pure semantics the objects under discussion are whatsoever elements may happen to satisfy the axioms—“signs” only by proleptic courtesy. For such abstractness and indefiniteness of reference are necessary in any discipline which aspires to be “pure”; the hopes of positive achievement in semantics, qua formal algebra of meaning, rest upon just such self-imposed limitations; and to charge it, in the mathematical aspects of its work, with lack of empirical reference would be quite to misconstrue its purposes.

But semantics is hardly interesting enough as pure mathematics to be pursued for its own sake; and, but for the wider claims of philosophical relevance which have been made on its behalf, the subject would have aroused little interest outside the ranks of specialists in symbolic logic. When a well-wisher claims that “the recent discovery of semantics is of fundamental philosophical importance”¹ it must be presumed that the reference is to descriptive semantics rather than to a branch of pure mathematics. The department of semantics for which such claims of philosophical relevance are made must then be prepared, like any other empirical discipline, to demonstrate the adequacy of its basic concepts and the fruitfulness of its empirical and analytical procedures in the solution of the problems which define its purposes.

These quotations are typical of many statements in the Logical Syntax of Language in which logic is summarily identified with syntax. A related insistence, common to positivists from Hume onwards, is that logic is non-factual; or, as Carnap more explicitly says, logical and mathematical theorems

In Carnap’s more recent work, reference to semantics almost, but not quite, replaces the earlier reference to syntax:

But we are also told, in a similar context, that the equation of logic with semantics

As syntax is a special branch of semantics (distinguished by the absence of explicit reference to designata) there need be no inconsistency between the later and the earlier formulations. The earlier statements were made at a time when Carnap did not yet hold reference to extra-linguistic designata, which is characteristic of semantics, to be an acceptable part of the new disciplines which were to replace “the inextricable tangle of problems which is known as philosophy”¹ (L.S.L., 279). His gradual relaxation of the methodological restrictions necessitated by earlier positivist criteria of meaningfulness of utterance has clearly not shaken his acceptance of some form of what I am calling a “linguistic interpretation” of logic; the “theory of logical deduction” is still, as in the earlier days, regarded as “not dependent upon facts” because it deals, at least in part, with the consequences of linguistic rules. The intelligibility and plausibility of such a doctrine clearly depend upon the extent to which its advocates can make precise the senses of logic, language and dependence upon language to which they are committed; and for this purpose of mere clarification of the linguistic thesis, the shift of emphasis from semantics to syntax is no less significant than the constancy of the general mode of approach. It will be a major concern of this paper to discover what addition is made to the plausibility of the linguistic thesis by the admission of semantics into the company of reputable philosophical disciplines.