Page 195 - Encyclopedia of Philosophy of Language

connection between a set of abstract syntactic and semantic rules, thought of as defining a particular natural language, itself regarded as an abstract entity, is forged by means of a highly general rule, or con- vention (in Lewis's sense), which the speakers actually follow.
2.3 Semantic Rules in Generative Grammar
In a grammar of modern linguistic theory the semantic component may look like this: on the one hand there is a so-called dictionary, the entries of which consist of lexical rules for primitive expressions, providing meaning (where possible by means of verbal meaning explanations) and grammatical category. On the other hand there are so-called projection rules for arriving at the meaning of complex expressions, and ultimately sentences, by way of selecting the readings of ambigu- ous simpler expressions that fit together within the more complex ones. In grammars of other kinds, how- ever, what are thought of as semantic structures (doubling as phrase structures) are generated directly, providing the basis for transformations and insertion of lexical items. Such generating rules are not semantic in the sense of (directly and overtly) providing interpretations of linguistic expressions. Within mod- ern linguistic theory the notion of a syntactic rule is clearer than that of a semantic rule, but for several reasons, in part connected with the existence of vari- ous constraints imposed on syntax by semantics or vice versa, the two notions are not sharply separated.
2.4 SemanticRulesandLanguage-games
In Hintikka's game-theoretic semantics (Hintikka and Kulas 1983;), developed for both formal and natural languages, a different conception of semantic rules concerning truth can be found. A sentence S is true in case there is a winning strategy, for the player Myself against the player Nature, in the semantical game associated with S. Such a game is defined by a number of rules, such as: the first move in a game associated with a conjunction A&B is Nature's choice of either A or B, whereupon the rest of the game is that associ- ated with Nature's choice (that is, A or B; since Nature makes the choice, Myself must have a winning strategy for A as well as for B). Hintikka connects this approach with Wittgenstein's notion of a language- game, claiming that a speaker's understanding of a sentence actually consists of his mastery of the rules associated with it and that semantic (word-world) relations are established in linguistic activity as governed by such rules.
In Sellars's writings (1963, 1974) a more abstract conception of the nature of semantic rules can be found. Sellars, too, employs the notion of a language- game, and makes the analogy with (some) ordinary games rather close. In using a linguistic expression one takes a position in the language-game. A move in the game is a transition from one position to another.
Rules of inference, material as well as formal (logical), govern such moves. One example (material) is the move from calling something red to calling it extended (the rule of which corresponds to a meaning postulate in Carnap's sense). Other rules, however, govern tran- sitions which are not moves proper but transitions into (language entry) and out of (language departure) the game. The transition from observing a red patch to calling it red is of the former kind, and the transition from uttering / am going out to going out is of the latter kind. Rules of these three kinds determine the meaning of expressions, but for epistemological reasons (compare Sect. 3) they are primarily to be thought of as ought-to-be rules, that is, as rules pro- viding standards for linguistic behavior, not as rules to be directly obeyed. Corresponding to these rules, however, there are ought-to-do rules for mature language users, requiring them to see to it that those standards are met, by training, teaching, and criticism (including self-criticism). Sellars's picture of linguistic practice as rule-governed is, of course, highly specu- lative, and it is doubtful that it can be borne out by more detailed considerations.
The notion of a language-game is originally Wittgenstein's (1958). The analogy with games stron- gly suggests a view of linguistic practice as rule- governed, and although it often arises in connection with other points, Wittgenstein repeatedly speaks of rules of language-games. It is open to debate, however, to what extent he acknowledges the existence of sem- antic rules as determining the meaning of linguistic expressions.
On the one hand there are in Wittgenstein references to grammatical rules, though 'grammar' is not here used in the ordinary sense but rather in the sense of a set of standards of description, which themselves give rise to rules of inference. Such standards can be ex- pressed in so-called grammatical statements, such as
White is lighter than black, or An order orders its own execution. These rules provide for inferences from A is white and B is black to A is lighter than B, and from A wasordered to Vand A didn't Vto A didn'texecute the order. They also exclude descriptions which are inconsistent with such inferences as nonsensical, coun- ter to grammar. These grammatical statements are normative; they do not flow from the meanings of the words involved but are part of the determination of meaning.
On the other hand it is not fully clear whether Wittgenstein's view was that there is, for example, a meaning-determining rule governing the use of red, or
just something closely analogous, namely (insti- tutional) standards of correctness of that use. The difference, if there is one, would be that if there were such a rule, then that rule would concern a certain transition, namely that from recognizing something as red to calling it red, and at least, so Wittgenstein argues, there is no rule of that kind. It may be,
Rules
173