Abstract

I argue that our procedures for determining whether ascriptions of a predicate represent things as being a certain way are ultimately pragmatic. Pragmatic procedures are not subject to validation by the referential procedure – determining whether there is a property playing the role of its referent. Predicates can represent even if we can't provide an independent identification of its referent. For these predicates, the speakers’ knowledge of how they represent objects as being would have to be construed in terms of the ascription practices they associate with the predicates. The same approach can be applied to semantic predicates, such as ‘is representational’ or ‘refers to’. They can be treated as representational even if we can't provide an independent identification of their referents. The availability of this position undermines accounts of the referents of semantic predicates in naturalistic terms.