Measuring accuracy

For the most part, the book up until this point has focused on comparative constructions which, together, may be understood as the core, “regular” comparatives. Yet a variety of other types have been purported to exist. The goal of this chapter is to re-examine whether that purported variety should bear on the analysis of comparative morphology proposed in the book. Ultimately, the chapter concludes that there is at least a bifurcation in the data, but shows how the relevant distinctions can be accounted for based on an interaction between abstract syntax and “what is measured” in the two cases. In particular, the chapter suggests that while “regular” comparatives involve the measure and comparison of entities that are, for the most part, idiosyncratic to the lexical category targeted, “categorizing” comparatives involve comparing the extent to which a given predication is accurate.