Multiple pressures to explain the ‘not all’ gap
Jeremy Kuhn, Lena Pasalskaya
August 2022 

Languages frequently lexicalize the quantifiers 'all', 'some', and 'none', but rarely lexicalize 'not all'. Enguehard & Spector (2021) argue that this gap can be derived from probabilistic properties of the lexicon; on plausible assumptions, 'some' is more informative, and thus more likely to be lexicalized, than 'not all'. Here, we describe new predictions of this theory, which we test in an experimental setting. We present cross-linguistic data that suggests that the pressure to not lexicalize 'not all' is weaker for modal quantification than for individual quantification, and show that this can be explained by the informativity hypothesis. We then experimentally measure probabilistic properties of the lexicon by asking subjects to evaluate the surprisingness of quantificational statements. The results suggest that informativity alone cannot explain the 'not all' gap, but support the hypothesis that it plays a role in the differences between quantificational domains.

Format:	[ pdf ]
Reference:	lingbuzz/007194 (please use that when you cite this article)
Published in:	Manuscript, Institut Jean Nicod
keywords:	not all, quantification, modality, language universals, information theory, semantics
Downloaded:	380 times

 

[ edit this article | back to article list ]