Abstract

Every Quantifier Isn\'t the Same: Informativity Matters for Ambiguity Resolution in Quantifier-Negation Sentences

---

Abstract

Empirical work on quantifier-not sentences has focused primarily on universal quantifiers, exploring the ambiguity that arises when logical operators interact (e.g., Everyone didn’t go could mean No one went or Not all went). In their Rational Speech Act model of this ambiguity resolution, Savinelli et al.(2017) demonstrate that pragmatic factors (such as model priors over the likely world states) stand to explain divergent behavior between children and adults. We extend this work to a broader empirical base, exploring the model\'s predictions for sentences with a wider range of quantifiers (some, no); we then test those predictions against behavioral data collected in a series of experiments. We find that a straightforward extension of the Savinelli et al. model captures the range of quantifier-not behavior we gather, thereby providing strong support for this cognitive model of probabilistic ambiguity resolution. In particular, the model explains interpretation preferences on the basis of informativity and prior beliefs over world states, such that interpretations that are more informative are preferred.

Keywords: ambiguity resolution, informativity, scope, Rational Speech Act, quantifiers, negation

How to Cite:

Attali, N., Scontras, G. & Pearl, L. S., (2021) “Every Quantifier Isn\'t the Same: Informativity Matters for Ambiguity Resolution in Quantifier-Negation Sentences”, Society for Computation in Linguistics 4(1), 394-395. doi: https://doi.org/10.7275/3zax-6v78