An approach to occasion-sensitivity

Abstract

The aim of this dissertation is to explore the hypothesis that language is occasion- sensitive. To hold that language is occasion-sensitive is to hold that the truth-conditions of most of our utterances depend on the occasion of use in a way that is not determined by meaning. In chapter 1, I reconstruct what I take to be Travis’ main arguments. Travis casts doubt on the existence of representations that are free of occasion-sensitivity by presenting a number of examples involving shifts in truth-value. In chapter 2, I set out to defend the claim that language is occasion-sensitive against minimalist and indexicalists replies to Travis’ arguments. In order to do so, I distinguish the Principle of Compositionality from what I call Semantic Propositionalism. Advocates of occasion-sensitivity reject only the latter. I argue that neither minimalist nor indexicalist accounts succeed in their defence of Semantic Propositionalism vis-à-vis Travis cases. Minimalists find themselves in an unstable position. In order to secure minimal propositions, they need to dismiss common reactions to Travis cases. But this casts doubt on the possibility of finding out what the literal satisfaction conditions of the expressions used in Travis cases are. Indexicalism tries to secure Semantic Propositionalism by claiming that some predicates are context-dependent. If they are to be a defence of Semantic Propositionalism, indexicalist theories must fulfill two conditions. First, they must provide a set of necessary and sufficient variables. Second, they must not have recourse to pragmatic interpretation. I argue that current proposals do not succeed in fulfilling both conditions. Chapter 3 addresses the question whether mental representations are occasion-sensitive. I focus on Fodor’s arguments and on Carston’s theory of ad hoc concepts. The productivity argument has it that the best explanation to the productivity of thought is compositionality, and mental representations being compositional prevents them from being underdetermined. The argument from equivocation is based on the idea that only a non-equivocal mental representation can resolve a linguistic equivocation. I argue that neither argument work. The productivity argument only establishes Meaning Compositionality, something compatible with (truth-conditional) Underdeterminacy. As to equivocation, the context of use can solve the equivocation in absence of a non-equivocal mental representation. I also argue that if Carston’s ad hoc concepts are created on line, then she cannot avail herself of the productivity argument. An additional aim of chapter 3 is to distinguish Type-Underdeterminacy from Token-Underdeterminacy. I argue that there are reasons to think that even tokens suffer from some Underdeterminacy in the sense that they only determine a partial function from states of a fairs to truth-values. Instead of relying on occasion-insensitive mental representation or having recourse to occasion-insensitive structured propositions, occasion-sensitivity calls for a non-standard notion of utterance content. The aim of chapter 4 is to provide such a notion. I hold that Austinian propositions, conceived as including a lekton and an activity, can do the work. To different activities correspond different criteria of applicability for words. Thus, adopting a situationalist framework, we can think of the truth-conditional content of an utterance as including not only the conventional meaning of the sentence uttered but also the activity against which it is evaluated. This notion of content is compatible with Token- Underdeterminacy. After having put forward this notion of utterance content, I discuss a potential problem for the approach. If activities are very finely individuated, as the possibility of creating complex Travis cases recommends, then sharing content across contexts will be diffcult to achieve. I argue that this problem can be solved by having Austinian propositions with different granularities, thus adopting a form of multi- propositionalism. In chapter 5 I address the question whether phenomenon that Travis has detected is compatible with standard semantic theories (in the sense of theories of truth-conditions). Semantic theories have been seen as an explanation of our ability to interpret speech. Advocates of occasion-sensitivity and similar pragmatic views are under pressure to show that their rejection of certain theories is compatible with a plausible account of our ability to grasp truth-conditions. I argue that occasion-sensitivity, and in particular the notion of truth-conditional content introduced in chapter 4, is compatible with there being systematic connections between activities and truth-conditions, which can be used to account for our ability to interpret speech.