Some remarks on inference patterns involving epistemic modality
Brian Buccola
May 2012
 

This paper identifies and resolves a previously unnoticed puzzle emerging from recent literature on quantifiers and epistemic modality. Specifically, I look at inference patterns involving superlative quantifiers ("at least m", "at most m"), which give rise to epistemic implications, and comparative quantifiers ("more than n", "fewer than n"), which do not. I identify an asymmetry in the inference patterns involving superlative quantifiers that I argue is not readily explained by either of two competing theories of superlative quantifiers. The solution I propose is that subjects in inferential tasks make certain default (but defeasible) assumptions: anything that is not known to be false is assumed to possibly be true. This proposal makes a number of predictions regarding inference patterns with overt epistemic modals, which I show are broadly correct. The end result suggests a new point of divergence between the notion of semantic entailment and that of an intuitively valid inference.
Format: [ pdf ]
Reference: lingbuzz/003037
(please use that when you cite this article)
Published in: First evaluation paper, McGill University
keywords: modified numerals, ignorance inferences, scalar implicature, epistemic modality, nonmonotonic reasoning, default reasoning, semantics
Downloaded:108 times

 

[ edit this article | back to article list ]