Truth, deflationary theories of

1. Problems with the Deflationary Thesis

What are we saying when we make utterances that appear to be ascribing truth to some truth bearer? According to the redundancy theory, whose invention is usually credited to F.P. Ramsey, we are saying nothing more or less than what is said in the statement to which we appear to be ascribing truth: ‘"Routledge editors are fine folks" is true’ is synonymous with ‘Routledge editors are fine folks’ (see Ramsey, F.P. §3). The redundancy theory, then, endorses the view that anything we can say or do with the predicate ‘is true’ can be said or done perfectly well without that predicate. The latter contention, which I shall call the Gratuity Thesis, is shared by the performative and prosentential theories of truth ascriptions discussed below, although these offer different accounts of what we are saying or doing when we ascribe truth. By itself, the Gratuity Thesis would be of importance only to linguists. Philosophical interest arises because redundancy theorists, and their performative and prosentential counterparts, make a breathtaking and rather cloudy inferential leap from the Gratuity Thesis to the Deflationary Thesis.

The description of the Deflationary Thesis given above seems to presuppose Platonism: that there are universal properties, and most predicates name properties. We can make it neutral with regard to Platonist–Nominalist debates by rewriting it thus: ‘"is true" is not a genuine predicate’.

It would be absurd to suppose that most predicates are not genuine predicates, so it would be equally absurd to take as a methodological principle that we should assume a predicate is not genuine until it is proven to be. Strangely, however, deflationists have adopted that principle, and anti-deflationists have let them get away with it. Indeed, it was not until the late 1970s and 1980s that the slow decline of ordinary language philosophy brought either contender to an awareness that linguistic analysis alone cannot settle the question (see Ordinary language philosophy §1). If indeed we do not need to postulate a property of truth (whether Platonically or Nominalistically construed) to explain what we are doing with our truth-ascribing utterances, then that is a point in favour of the Deflationary Thesis. But it is not decisive, for there may well be non-linguistic programmes which cannot be carried out without supposing that truth is some kind of property. Anti-deflationists have suggested, for example, that without such a postulation, we cannot explain why our scientific theories are so successful. To this, deflationists have responded that there is really nothing here that needs explanation: we have the theories we do just because they are successful. But this will not do. The fact that we would not have automobiles if they did not succeed in getting us from place to place does not preclude the need for an explanation of why they work, which would presumably be given in terms of the properties of automobiles and the forces at work in them.

Explaining the success of science is just a special case of the traditional epistemological programme of discovering the correct theory of justification. The task of the latter project is to refute scepticism by finding some property, called a ‘mark’ of truth, which correlates, perhaps imperfectly, with truth, and whose possession or non-possession by a given proposition is reasonably easy to detect. Coherence theories of justification (not truth) propose ‘coherence with other propositions’ as the mark. Foundationalists propose that it is ‘self-evident (perhaps defeasibly), or inferred from self-evident premises’. One cannot judge whether a proposed mark does indeed correlate with truth unless one has some notion of what truth is, so this programme must postulate that truth is some kind of property. Those few deflationists who have considered this proposal respond by insisting, in effect, that we are entitled just to beg the question against scepticism, and hence there is no need for traditional epistemology in the first place. Begging the question against scepticism has been popular in contemporary philosophy, but anyone tempted to think that this is philosophically respectable should consider the following analogy: to prove the soundness and completeness theorems for classical logic is, in effect, to prove that the results of the logic’s rules of inference correlate (perfectly) with truth (as defined in the logic’s model theory). How seriously would we take one who claimed that such proofs (and hence the model theory itself) are unneeded because we are entitled to assume that the inference rules of classical logic are correct? (And how would ‘correct’ as used here be defined without postulating a property of truth?) Even mathematical intuitionism does not want to say that there is no such thing as truth. It simply defines truth in terms of justification, that is, proof.