One of the main logical functions of the truth predicate is to enable us to express so-called 'infinite conjunctions'. Several authors claim that the truth predicate can serve this function only if it is fully disquotational (transparent), which leads to triviality in classical logic. As a consequence, many have concluded that classical logic should be rejected. The purpose of this paper is threefold. First, we consider two accounts available in the literature of what it means to express infinite conjunctions with a truth predicate and argue that they fail to support the necessity of transparency for that purpose. Second, we show that, with the aid of some regimentation, many expressive functions of the truth predicate can actually be performed using truth principles that are consistent in classical logic. Finally, we suggest a reconceptualisation of deflationism, according to which the principles that govern the use of the truth predicate in natural language are largely irrelevant for the question of what formal theory of truth we should adopt. Many philosophers think that the paradoxes pose a special problem for deflationists; we will argue, on the contrary, that deflationists are in a much better position to deal with the paradoxes than their opponents.
The ProblemMany philosophers maintain that the truth predicate can serve certain expressive roles of a quasi-logical nature, the most salient of which is to enable us to express so-called 'infinite conjunctions'. These are sentences of the form All Ps are true.where P is a predicate of the language that applies to infinitely many sentences.
1They are considered to express the infinitely many Ps at once or, as is often said, their infinite conjunction, without turning to infinitary or higher-order resources (cf. Quine 1970;Leeds 1978;Putnam 1978; Horwich 1998;Field 2007).(The phrase 'expressing infinite conjunctions' is taken from the literature-e.g. Putnam 1978; Gupta 1993-and of course in need of clarification. For the moment, the reader should take it as a technical term.) We call this function of the truth predicate the 'infinite-conjunction' function.As Quine (1970, chap. 1) points out, the universal quantifier serves a similar purpose. If the infinitely many sentences we want to express differ in one or several individual terms-e.g. ''0 is divisible by 2'', ''2 is divisible by 2'', ''4 is divisible by 2'', etc.-and the class of objects these terms denote is definable in the language by a suitable predicate (e.g. ''is an even number''), we can express the infinitely many sentences at once just generalising over those terms using this predicate-e.g. uttering ''All even numbers are divisible by 2''. However, if the infinitely many sentences we want to express don't differ just in one or more individual terms, this strategy is no longer available. In that case, the truth predicate, interacting with the universal quantifier, might do the job, as long as the sentences at issue share a property definable in the language. For instance, we can assert al...