< previous page page_208 next page >

Page 208
(1) all bachelors are males
(2) all men are two-footed
are supposed to be true on the grounds of the meaning postulates assumed by a given system of signification (independently of the fact thataccording to a venerable tradition(1) is analytically true while
(2) is synthetically true). On the other side,
(3) this is a pencil
(4) this pencil is black
are true only if they are uttered in a given circumstance, where it is the case that the indicated object is a pencil and is black.
There are two domains of a truth-conditional semantics, one studying the requirements to be met by a proposition in order to be (logically or semantically) true or false on the grounds of a system of meaning postulates, the other studying the requirements to be met by a proposition in order to be (factually) true or false on the grounds of what is actually the case.
We see then that there are at least three sorts of theories that one can label as "semantics," namely,
(i) as theory of meaning, or a theory of semantic competence, or a cognitive semantics;
(ii) a theory of truth for nonindexical expressions, or for eternal propositions;
(iii) a theory of truth for indexical expressions such as the acts of mention (see for further distinctions Eco 1976:3.1 and 3.2).
None of these three semantics can avoid the pragmatic dimension.
Theory (iii). We must agree with Strawson (1950) when he says that "mentioning or referring is not something that an expression does; it is something that someone can use an expression to do." If it seems evident that "indexical expressions are standardly and naturally handled with truth conditional apparatus" (Gazdar 1979:2), it is equally evident that the truth of indexical expressions depends on the circumstances of the utterance, on the nature of both the utterer and the addressee (personal pronouns) as well as on the nature of the indicated object. Thus the pragmatic problem of deixis stands at the very core of the allegedly

 
< previous page page_208 next page >