Reassessing Referential Indeterminacy

Quine and Davidson employ proxy functions to demonstrate that the use of language (behaviouristically conceived) is compatible with indefinitely many radically different reference relations. They also believe that the use of language (behaviouristically conceived) is all that determines reference. From this they infer that reference is indeterminate, i.e. that there are no facts of the matter as to what singular terms designate and what predicates apply to. Yet referential indeterminacy yields rather dire consequences. One thus does wonder whether one can hold on to a Quine-Davidson stance in semantics-cum-metaphysics and still avoid embracing referential indeterminacy. I argue that one can. Anyone adhering to the behaviouristic account pivotal to the Quine-Davidson stance is bound to acknowledge certain facts about verbal behaviour – that some utterances are tied to situations, that some utterances are tied to segments in situations, that some predicates have non-contextualised conditions of application, and that use involves causal dependencies. The restrictions from these facts ensure that only reference relations generated by means of rather exceptional proxy functions are compatible with verbal behaviour. I conclude that this allows one to rebuff the Quine-Davidson argument for the indeterminacy of reference, as it were, from within. I moreover tentatively conclude that the line of thought laid out provides good reason for just about anyone to hold that there are facts about reference after all. Reassessing Referential Indeterminacy 1 Reassessing Referential Indeterminacy 1. The Thesis Explored Celebrated ideas often share the fate of human celebrities: they become shrouded in a haze of publicity, adoration, and rumour. This might very well be true of the claim that reference is indeterminate. So let me begin with some clarifications. Firstly, the thesis concerns semantic reference, not speaker’s reference. It is a thesis about what singular terms as well as their utterances designate and what predicates as well as their utterances apply to. Secondly, the thesis is not an epistemological claim. What is claimed is rather that there are no facts of the matter as to what singular terms designate and what predicates apply to. Speaking of the ‘inscrutability of reference’ hence is misleading – as Quine acknowledges: “It is what I have called inscrutability of reference; ‘indeterminacy of reference’ would have been better”. Thirdly, the thesis is neither to be confused with, nor grounded in, the claim that many terms are vague and/or ambiguous. Neither phenomenon plays any role in the arguments Quine and Davidson put forth. Indeterminacy of reference is thought to arise even if we idealize away vagueness and ambiguity. Fourthly, the indeterminacy of reference is independent of any indeterminacy of logical form. Quine and Davidson agree that even if one assumes a determinate logical form, reference still comes out indeterminate. Finally, the thesis is meant to cover all languages. Indeterminacy is not confined to distant primitive tongues. As the saying goes, indeterminacy begins at home. Anyone who holds that reference is indeterminate is hence committed to the claim that there are no facts of the matter as to what singular terms designate and what predicates apply to. To render this idea more lucid, let me define an interpretation of a language L to be a complete assignment of extensions to the singular terms and predicates of L. Let us furthermore agree to call some interpretation correct for some language L if and only if it is compatible with all the facts that determine the semantic properties of L. Employing this terminology, the thesis that reference is indeterminate – that there are no facts of the matter pertaining to reference – unfolds thus: (IR) For every language L, there are indefinitely many fundamentally different yet equally correct interpretations. This is precisely what Quine and Davidson, the avowed champions of indeterminacy, seek to establish. Please note that the claim they aim to prove is neutral with respect to their general approach in semantics-cum-metaphysics. The indeterminacy thesis is neither wedded to any specific stance in semantics, nor is it intrinsically linked to the scenario of radical translation or interpretation, respectively. Since Quine and Davidson often run the thesis together with what they take to be the argument to establish its truth, this is easy to miss. Reassessing Referential Indeterminacy 2 The indeterminacy thesis is not perfectly neutral, though. It comes with two rather sensible background assumptions. The one assumption is that semantic facts are dependent facts. What sentences mean – what their truth conditions are – and what words refer to is in the end fixed by non-semantic facts. Assuming otherwise would run counter to what Quine and Davidson believe. It furthermore would turn at least some semantic facts into brute facts, and that is hardly appealing. The other assumption is that reference is a substantial relation between words and things rather than a purely disquotational intra-linguistic relation. That again makes sense. For it is very hard to see how reference could possibly be indeterminate if all facts about reference were exhausted by trivialities such as “‘rabbit’ denotes rabbits”. Disquotational theories of reference do not so much solve the indeterminacy problem as sidestep it. The purported indeterminacy poses a serious threat to anyone’s understanding of the interrelation of mind, language, and the world. If (IR) is true, neither any singular term nor any predicate stands in any determinate relation to things in the world. Imagine you assert “Gene Kelly was the greatest dancer ever”. Your assertion will be about very many different objects – Mount Everest and Khrushchev’s shoe, say – and it will put these objects in very many different classes. That will be true in every single case. We hence cannot use our singular terms to pick out certain things rather than others, and we cannot use predicates to classify certain objects rather than others. The indeterminacy of reference thus dissolves reference. As Leeds puts it: “Let us take inscrutability of reference as having shown that there is no such relation as reference”. Worse still, given that what you have said is just what you have thought, (IR) implies an indeterminacy of thought. The indeterminacy of reference thus dissolves intentionality, too. We consequently can rely neither on language nor on thought to relate to the going-ons in any specific region of the world. Yet it is very hard to envisage what else could allow us to do so. If reference is indeterminate, then, we are sort of losing the world. Not everyone will think that that is all too well. In fact, not even everyone who subscribes to a Quine-Davidson stance in semanticscum-metaphysics will think that this is all too well. The question thus is, can she do something about it? Imagine that you are a staunch adherent of the Quine-Davidson stance in semantics-cum-metaphysics. Can you hold on to these views and still escape (IR) and its dire consequences? Quine and Davidson think that you cannot. They argue that once you have endorsed their approach to semantics-cum-metaphysics, acknowledging that reference is indeterminate becomes inevitable. I beg to differ. I will argue that even a committed adherent of their general views can indeed avoid subscribing to (IR). More precisely, I will argue that a Quine-Davidson stance in semantics-cummetaphysics provides all the means needed to undercut the inference from the principles of that stance to the indeterminacy of reference. I thus maintain that one can rebuff the indeterminacy thesis, as it were, from within. I will moreover argue – though somewhat tentatively – that anyone adherent the Quine-Davidson approach should rather reject the indeterminacy thesis. For I think that the failure of the general argument Quine and Reassessing Referential Indeterminacy 3 Davidson rely on to infer (IR) provides good reasons to believe that reference isn’t really indeterminate at all. 2. The Argument for the Indeterminacy of Reference Quine and Davidson argue that the totality of all relevant facts radically underdetermines the interpretation of any given language. From this they conclude that reference is indeterminate. Suitably expanded, the Quine-Davidson argument for the indeterminacy of reference is best captured by the following scheme: (a) Semantic properties are exclusively determined by A facts. (b) For any language L, the totality of A facts is compatible with indefinitely many radically different interpretations.