Toward a Rigourous Definition of Vagueness in Semantics

This paper deals with vague predicates. In particular, the goal is to arrive at a precise characterisation of a vague predicate without begging the question against a particular theoretical account on the matter. For instance, such a characterisation should not exclude by fiat a theory that exploits the notion of gappy or glutty truth values. Furthermore, it should be neutral with respect to locating vagueness in epistemology or language. Following Bueno and Colyvan (2012), I argue in favour of characterising vagueness purely in terms of sorites susceptibility. However, I caution that the sorites argument can also be used to diagnose epistemic uncertainty of other sorts (or epistemic vagueness, according to Sauerland and Stateva, 2007) if we allow for partially ordered sets. This seems like an undesirable result as it captures phenomena which are fundamentally different than pure (scalar) vagueness. I argue that so-called borderline cases resulting from general epistemic uncertainty differ from those resulting from scalar vagueness and the sorites can be employed as a metric to tease these apart.This paper deals with vague predicates. In particular, the goal is to arrive at a precise characterisation of a vague predicate without begging the question against a particular theoretical account on the matter. For instance, such a characterisation should not exclude by fiat a theory that exploits the notion of gappy or glutty truth values. Furthermore, it should be neutral with respect to locating vagueness in epistemology or language. Following Bueno and Colyvan (2012), I argue in favour of characterising vagueness purely in terms of sorites susceptibility. However, I caution that the sorites argument can also be used to diagnose epistemic uncertainty of other sorts (or epistemic vagueness, according to Sauerland and Stateva, 2007) if we allow for partially ordered sets. This seems like an undesirable result as it captures phenomena which are fundamentally different than pure (scalar) vagueness. I argue that so-called borderline cases resulting from general epistemic uncertainty differ from those resulting from scalar vagueness and the sorites can be employed as a metric to tease these apart