Abstract

According to the standard story W. V. Quine ’s criticisms of the idea that logic is true by convention are directed against, and completely undermine, Rudolf Carnap’s idea that the logical truths of a language L are the sentences of L that are true-in- L solely in virtue of the linguistic conventions for L, and Quine himself had no interest in or use for any notion of truth by convention. This paper argues that and are both false. Carnap did not endorse any truth -by- convention theses that are undermined by Quine ’s technical observations. Quine knew this. Quine ’s criticisms of the thesis that logic is true by convention are not directed against a truth -by- convention thesis that Carnap actually held, but are part of Quine ’s own project of articulating the consequences of his scientific naturalism. Quine found that logic is not true by convention in any naturalistically acceptable sense. But he also observed that in set theory and other highly abstract parts of science we sometimes deliberately adopt postulates with no justification other than that they are elegant and convenient. For Quine such postulations constitute a naturalistically acceptable and fallible sort of truth by convention. It is only when an act of adopting a postulate is not indispensible to natural science that Quine sees it as affording truth by convention ‘unalloyed’. A naturalist who accepts Quine ’s notion of truth by convention is therefore not limited to accepting only those postulates that she regards as indispensible to natural science