Covert operators are picked to minimize QuD-Ambiguity: The view from 'pex' and 'only'

Abstract

Silently derived scalar implicatures (SI) have a semantic effect similar to that of the overt operator only, differing primarily in how the relevant meaning components are distributed between presupposition and assertion (Fox 2007; Bassi et al. 2021 i.a.). Why is the division of labor between silent SIs and overt only so stable across languages? This paper argues that the answer lies in ambiguity minimization. Specifically, it argues that (i) SIs and only are used to address distinct Questions under Discussion (QuD; Roberts 2012 i.a.); (ii) because covert operations are optional, they create uncertainty about which QuD should be accommodated; (iii) when two operations are semantically similar enough, they compete for covertness, such that the covert operation (if any) should be the one that minimizes a QuD-related measure of ambiguity. This predicts SIs to be the preferred covert operation. The proposal provides a principled explanation for the semantic asymmetry between SIs and only and, in turn, yields an indirect argument for a grammatical, presuppositional view of SIs.