Abstract
Some sentences with presupposition triggers can be felicitously uttered when their presuppositions are not entailed by the context, whereas others are infelicitous in such environments, a phenomenon known as Missing Accommodation / Informative Presupposition or varying Contextual Felicity Constraints (CFCs). Despite an abundance of recent quantitative work on presuppositions, this aspect of their behavior has received less attention via experimentation. Here, we present the results from a semantic rating study testing the relative CFC strength of thirteen presupposition triggers, making this the largest cross-trigger comparison reported in the literature to date. The results support a three-way categorical analysis of presupposition triggers, based on imposing strong, weak, or no CFCs. We observe that strong CFC triggers are all focus-associating, suggesting that (at least some of the) variation in behavior arises due to naturally-occurring semantic classes. We compare our results to three previous proposals for CFC variation and argue that none yet account for the full empirical picture.
