The relative role of property type and scale structure in explaining the behavior of gradable adjectives

Abstract

Kennedy (2007) proposes a semantics for positive form adjectives on which the standard for ascribing an adjective A makes the individuals that are A stand out from those that are not. To account for the di erences between absolute and relative adjectives, Kennedy posits that the maximal and minimal degrees on closed scales naturally make individuals stand out in a way that degrees found away from the endpoints of a scale cannot. I argue that the ability of a degree to make individuals stand out is due less to scale structure than to the nature of the property the adjective describes. Thus, degrees that are not endpoints can behave like absolute standards as long as the application criteria for the property are clear. I relate the identi ability of such criteria to whether the property ascription can be modeled in terms of rule- vs. similarity-based classi cation (see e.g. Hahn and Chater, 1998).