Eric Wallace
University of Vienna
Davidson’s method of truth, which is widely used in Natural Language Ontology (NLO), resolves metaphysical questions by appealing to the truth conditions of sentences. I raise a worry for the method of truth. Truth conditions are partially specified using sets. For example, the truth conditions of ‘Tibbles is a cat’ are given by assigning as semantic values [Tibbles] (i.e. the actual cat) to Tibbles’ and the function taking the cats to true is a cat’. Hence, as the truth conditions of Tibbles is a cat’ requires the existence of functions and `Tibbles is a cat’ is true, the method of truth tells us that there are functions. But claims about Tibbles do not mandate the existence of functions. The problem is exacerbated when we consider sentences that require more complex sets assigned as semantic values. Call this the problem of exotic commitments.
A natural solution arises from the role of semantic theorising. It is widely accepted that the role of semantics is to characterise the competence that enables us to use and understand language meaningfully. It consists in our ability to understand the meanings of concatenations of expressions in terms of the meanings of their constituents. I propose a compositional structure shared both by our semantic competence and the truth conditions of sentences. This structure accounts for semantic success. I claim that sets assigned as semantic values characterise this structure. Accordingly, semantic analyses do not require the existence of sets to make sentences true. These sets are not part of what combines to give truth conditions, but characterise how the non-sets of our analyses combine to give truth conditions. This analysis has three virtues: (i) It resolves the problem of exotic commitments, vindicating NLO; (ii) the resulting picture closely resembles a widespread view of syntax, unifying both disciplines; (iii) it vindicates semanticists’ liberality with sets in their analyses.
Chair: tba
Time:
Location:
SOPhiA 2026 