Comparing theories of subcomparatives

The goal of my presentation shall be a reevaluation of the claim that the data on subcomparatives argues in favor of a degree approach to constructions of gradability, in contrast to a delineation approach. I show that this claim is build on an incorrect or incomplete set of empirical observations. Taking into account a bigger set of observations, the degree theoretical predictor of acceptability of subcomparatives - namely lexically encoded (in)commensurability - is shown to be insufficient. A more adequate predictor (based on a notion of informativity) is proposed and implemented within a delineation framework.