Once again on norms and comparison classes

Abstract

A central assumption about relative adjectives (e.g. big, old) is that their positive form is interpreted vis-à-vis a class-specific reference point located in the mid-zone of a series (norm). More recently, functional-cognitive studies argued that other reference points (e.g. argumentative zero, endpoints) are more relevant in actual language use than a norm. This paper argues that the two positions are not necessarily irreconcilable and experimentally tests a hypothesis that a norm is a default reference point used for the interpretation of relative adjectives in zero-contexts. Experiment 1 addressed the location of a norm in the mid-zone of a series and its category-dependence. As predicted by the traditional semantic studies, the cut-off point between ‘big’ and ‘small’ is located around the midpoint of a scale. Furthermore, its location is category-dependent and sensitive to prototypicality effects. The results further indicate that adults easily compute a contextually relevant norm by integrating their world knowledge with the visually provided information. Experiment 2 investigated the relevance of a norm in the on-line processing of relative adjectives. The results suggest that language users indeed exploit norms for the interpretation of relative adjectives in real time.