As a result of thinking (pace Tarski, wrongly) that it is propositions, not sentences, that are true or false, it has been supposed (also wrongly) that propositions such as that ‘Snow is white’ is true if and only if snow is white are necessarily true. But changing the rules for the use of the words in a sentence has no effect on the truth of the proposition, only on what proposition it formulates. Many similar statements, e.g., that ‘plus’ does not (...) mean plus, are only pragmatically contradictory: if this were true, it would be impossible to say so in these words. One should distinguish between sentences that express necessary truths, and sentences that necessarily express truths. It follows that many well known accounts of necessity are wrong, that the truth of an analytic proposition does not follow from the definitions of the words in the sentence that expresses it, that it is not helpful to define meaning in terms of truth, that truth is not relative to language, and that conventionalism is false. This paper is a move in the direction of establishing the eternity of truth. (shrink)

In the old days of descriptivism, fictional reference and non-fictional reference with proper names were treated on a par. Descriptivism was not an intuitive theory, but it meritoriously provided a unitary semantic account of names, whether referentially full or empty. Then the revolution of the new theory of reference occurred. This new theory is definitely more intuitive than descriptivism, yet it comes with a drawback: the referentially full use and the referentially empty use, notably the fictional use, of names are (...) semantically no longer on the same footing. How the use of names from fiction is referentially empty has even been seen as a symptom of its not being a use at all. In this paper, I ask the following question: can one keep the merits of the new theory of reference and still provide a unitary account of both fictional and non-fictional reference with names? My answer is positive, provided that one conceives proper names as a special kind of indexicals. (shrink)

The theme of this paper is a philosophical theory of communication; more specifically, a theory about the understanding of language. It is an old theory, whose classical exponent was John Locke, and in the form that Locke expounded it the theory is now generally rejected by philosophers. But it is far from being a mere museum piece. The view about language that Locke put forward was a plausible one, and it has continued to be put forward in various forms. My (...) main concern here is with a very recent version of the theory, contained in a book published early in 1975; however, it will be helpful to begin with a brief account of Locke's version. This is not only classical, but is also simple and clear, and as such it may serve as a useful stepping-stone towards an understanding of the more modern version. (shrink)

I am not sure how confident we can be about our knowledge of the meaning of a sentence, but I am sure that we should be much less confident than we often are about the meaning of a word or non-sentential expression. This paper is an attempt to whittle away our confidence about word-meaning, and it is to this end that I investigate the meaning of the word ‘know'. But the point of the investigation is to show that it is (...) not enough to investigate just the meaning of a word. We must investigate the syntax as well, something that philosophers have often ignored. A theory of meaning must be closely tied to a theory of syntax, and for one to be acceptable the other must be acceptable as well. A theory of language must be an integrated theory, and hypotheses about syntax and meaning stand or fall together. (shrink)