Abstract

Registration link https://forms.gle/jAiDPFv3tJ3XMiSb6

ABSTRACT

Ever since the Gricean turn scalar quantifiers (some, most), numerals, and connectives and and or have received under-informative core meanings, enriched by pragmatic inferences (more recently grammaticized according to some). For example, some and most are assumed to be semantically lower-bounded, namely also compatible with 100% values, while their upper bound (‘less than all’) is derived separately. Or is assumed to be semantically inclusive, hence also compatible with states of affairs in which both alternatives are true, and its supposedly common exclusive interpretation (only one option is true) is similarly separately derived.

I will argue that the quantifiers are semantically upper-bounded, although a truth-compatible inference may (but need not) bridge the gap between an upper-bounded ‘most’ and an ‘all’ state of affairs (Ariel, 2004, Ariel, 2015). Indeed, assertions on subsets may be judged true if these are seen as compatible with ‘all’ being the case. While my analysis of the quantifiers shows them to be more informative than the consensual assumption, my analysis of or (Ariel and Mauri, 2019) assumes that it has quite a different minimal meaning. Or by itself carries no truth-conditional content (see also Alonso-Ovalle, 2006) but rather, a procedural meaning, imposing an ‘alternativity’ relation between the disjuncts. The reading associated with the or proposition may be such that none, one or both disjuncts hold true. Once again, however, truth-compatible inferences may (but need not) bridge the gap between a speaker-intended ‘single option’ and a state of affairs in which ‘both’ is true (Ariel et al., in preparation).

The take-home message is that truth judgments are not solely accounted for by reference to ‘what is said’/explicatures. Truth-compatible inferences may (but need not) allow for certain discrepancies between these representations and the relevant states of affairs.