In the John Locke Lectures, included in Meaning and the Moral Sciences, Hilary Putnam argues that "the 'softness' of social facts may affect the 'hard' notions of truth and reference" ([19] *, p. 46). Without fully endorsing Putnam's argument, I hope to show that a similar argument could be constructed for a slightly different conclusion: that the 'softness' of ethics may affect the 'hard' notions of truth and reference.
Putnam's argument can be divided roughly into two movements. The first is an attempt to show that an understanding of notions of truth and reference requires an understanding of translation. Here Putnam's primary claims are as follow:
(1) Though Tarskian semantics adequately captures the formal logic of 'true' and 'refers', the concepts of truth and reference are underdetermined by their formal logic. Acceptance of a Tarski-style truth-definition for one's language does not itself dictate whether truth and logical connectives are to be understood realistically or idealistically, 'classically' (i.e., in terms of classical truth and falsity) or 'intuitionistically' (i.e., in terms of provability in a theory). (See esp. [19], (2) A second 'dimension' of underdeterrnination appears in contexts of translation. In applying 'true' and 'refers' with regard to a language other than our own,we require not only a concept of truth adequate to our own language but also a mapping of the other language on to our own. (See esp. [19], pp. 38--46.) The first movement of Putnam's argument is roughly from truth and reference to translation. The second movement is from translation to 'social facts'; wants, desires, and interests. Here a further set of claims is of importance:
(3) As Quine has shown, a principle of charity in translation-whereby we maximize the reliability of foreign utterances -is not enough; an infinite