Iconic presuppositions

摘要

Why are some linguistic inferences treated as presuppositions? This is the 'Triggering Problem,' which we attack from a new angle: we investigate highly iconic constructions in gestures (speech-replacing gestures or 'pro-speech gestures') and in signs (classifier predicates in ASL) and show that some regularly trigger presuppositions. These iconic constructions can be created and understood 'on the fly,' with two advantages over lexical words: they suggest the existence of a productive 'triggering algorithm,' since presuppositions can arguably be generated with no prior exposure to the iconic construction; and they make it possible to minimally modify the target constructions to determine which do and which do not generate presuppositions. Our investigation does not just target standard presuppositions, but also 'cosuppositions,' initially defined as conditionalized presuppositions triggered by co-speech gestures. We show that pro-speech gestures and classifier predicates alike can trigger cosuppositions, which are thus an inferential class that goes beyond the confines of co-speech gestures (Aristodemo2017). Our data argue for a triggering algorithm that can generate presuppositions on iconic grounds, and we offer a generalization of cosupposition theory on which these can be triggered for reasons of manner (co-speech gestures), but also for conceptual reasons.