As Lepore and Ludwig point at the beginning of the book, the most principal subject in Davidson’s early writings in philosophy of language in the 1960s is that we are limited beings whose sway of the indefinitely many expressions of our language must somehow come of our sway of limited resources. Otherwise, there would be an unlimited number of individual things to learn in learning a language, which would make language learning unrealizable for limited beings like us. The linguistic competence of a limited being of our kind must be consequence cooperation of a limited number of main competencies. This feature headed Davidson to point what I will name finite primitives constraint, videlicet, that cannot be an unlimited number of primitive significations of expressions in our language. Davidson reasoned that this constraint was disparate with certain proposals of the time. It rejected Israel Scheffler’s proposal that intentional verb phrases be interpreted as primitive unstructured units. In Scheffler’s opinion, the verb phrase, believes that Socrates was a philosopher, was to be interpreted as a primitive unstructured predicate-believes-that-Socrates-was-a-philosopher. Since this proposal would imply interminable many primitive expressions, it encroached the limited primitives constraint. A comparable objection can be made to Quine’s proposal to interpret quoted expressions as primitive wholes. Davidson reasoned thereto that the finite primitives constraint left out certain variants of Frege’s proposal that words utilize intentional contexts do not have their usual meanings but in place have a particular oblique meaning, and have various still doubly oblique meanings in doubly intentional contexts, and so on for deeper contexts. In this kind of opinion, philosopher has its usual meaning in Socrates was a philosopher, an oblique meaning in Mary thinks that Socrates was a philosopher, a doubly oblique meaning in Jack says that Mary thinks that Socrates was a philosopher, etc. If an oblique meaning of an expression is not defined by the normal meaning of the expression, and in kind for the doubly oblique and higher classification meanings, then this opinion appears to be committed to indefinitely many primitive meanings, which encroaches Davidson’s limited primitive constrain. Davidson also reasoned that the constraint left out a suggestion that might be made about certain types of prepositional phrases, pursuant to which verbs and prepositions unite to constitute semantically primitive relational predicates. In the rejected opinion John walked in the street would be interpreted as containing a semantically primitive relational predicate expression walked-in that did not include as a part the expression walked Davidson reasoned that there were indefinitely many types of adverbial modification by prepositional phrases, which would need indefinitely many primitive predicates, a consequence left out by the limited primitives constraint. These and other comparable early reasoning of Davidson had a main influence at the time and it is now widely consented that any acceptable analysis in these fields must concern some variant or other of the finite primitives constraint.
Semantic Structure and Theory of Truth
Specified that there is no bound to the number of expressions in a language, the limited primitives constraint suggests that most expressions in the language are not primitive. Non-primitive expressions are themselves constituted of primitive expressions and it seems that their significations must somehow be determined by the expressions, out of which they are formed, the significations of those expressions, and the methods they are put together. But how can we clarify the method that significations of compound expressions are formed from the expressions they include and the significations of those expressions? Frege had offered identifying significations with entities of a certain kind, Fregean senses. The entity appointed as a sense to a complex expression was considered to be a function of the entities appointed as senses to the expressions of which the complex expression was formed. On the other hand, Davidson reasoned that Fregean aspect had serious problems unless follow in a certain method. Davidson’s positive proposal was that an interpretation of how the significations of compound expressions in a language rely on the significations of their pieces could be reached though certain kind of “theory of truth”¯ for the whole language or at least for a fraction including the relevant expressions. The theory was to be patterned after Tarski’s theory of truth for a certain formal language. It was answered a version of Tarski’s Convention T, allowing arguments of relevant T sentences of the shape, x is true iff p, where x was to be displaced by something that directed to a sentence of the language and p was to be displaced by that sentence or a translation of that sentence. Not just any kind of theory of truth would execute, however. For instance, allow a fragment of a language with no indexical components and no doubtful expressions. Allow the theory of truth for that fraction with infinitely many axioms of the shape: “s”¯ is true iff s. Such a doctrine would not by itself shed light on how the significations of compound expressions rely on the expressions from they are formed and significations of those expressions. So there were two related scheme. One was to devise other prerequisites on a theory of truth, in addition to Convention T, to be seen if the theory was to serve as a key part of an interpretation of semantic competence. The other was to give theories of truth of the relevant kind for different fractions of normal language. With respect to the first scheme, that of discovering extra constraints on the relevant kind of theory of truth, one suggestion was that the theory of truth has only finitely many (nonlogical) axioms. Such a constraint must be supposed to be connected to the finite primitives constraint, on the ground that each principle of the doctrine should coincide with an individual aspect of a limited language user’s competence, an individual component that has to be studied in learning the language.
What to say about Davidson’s influence here? On the other hand, it is now widely accepted that proposals in semantics should include explanations of how those proposals allow relevant theories of truth for sentences of the kind being explored. Truly generally accepted textbooks in linguistic semantics order students to get important facility in comprehension how such doctrines are evolved. On the other hand, there appear to be differences referring admissible background logic, with some theorists following Montague in background logic allowing second or higher-order logic and not confining the logic to standard first order extensional logic. With respect to the second scheme, that of giving the important kind of semantics for different fractions of natural language, Davidson made essential proposals about adverbial modification and about indirect and direct quotation. Davidson’s incipient proposal about direct and indirect quotation was that the cited material to be interpreted as displayed material was not indeed piece of the containing sentence. So, Jack believes that snow is white was interpreted
as a sentence attended by a second sentence that was related to by the sentence: Jack believes that. Snow is white. Each of these sentences could then be presented a standard semantical treatment. Davidson’s more thought-out opinion was that cited material could be correctly interpreted both as something related to and also a piece of the containing sentence.