When is Likely Unlikely: Investigating the Variability of Vagueness

Abstract

An important part of explaining how people communicate is to understand how people relate language to entities in the world. In describing measurements, people prefer to use qualitative words like `tall' without precise applicability conditions, also known as vague words. The use of vague language varies widely across contexts, individuals, and tasks (single reference vs. comparisons between targets), but despite this variability, is used quite successfully. A potential strategy for using vague language is to leverage the set of alternative descriptions to settle on the best option. To determine whether people use this strategy, we conducted an experiment where participants picked vague words from sets of alternatives to describe either probability or color values. We varied the set of alternatives from which participants could choose. Empirical evidence supports the hypothesis that people use the set of available options to pick vague descriptors. The theoretical implications of this work are discussed.


Back to Table of Contents