The utility of vagueness: does it lie elsewhere? Seminar at Institute for Language, Cognition and Computation, University of Edinburgh

Kees van Deemter, Matthew James Green

Research output: Other contribution


Much of everyday language is vague, yet standard game-theoretic models make it difficult to see how (borderline) vagueness can have benefits over crisp ones in co-operative situations. This talk will start by discussing the relevance of these issues for Natural Language Generation, and reporting on a range of attempts to pin down what the benefits of vagueness might be. Next, we report on a sequence of psycholinguistic experiments that were recently conducted in Aberdeen, which are starting to suggest a new, and less favourable view of vagueness: it appears that many of the benefits that vague terms can exert are caused by factors that tend to co-occur with vagueness, rather than by vagueness itself.
Original languageEnglish
TypeInvited seminar
Publication statusUnpublished - Mar 2014


Cite this