An Attentional Constraint on Spatial Meaning

Abstract

Broad commonalities in human perception may constrain cross-linguistic variation in semantic systems (Berlin & Kay, 1969; Kay & Regier, in press).Regier's (1996) model of spatial term learning assumes such a constraint: preferential attention to the endpoints of spatial events, rather than their beginnings (see Arnold, 2001 for evidence of a similar constraint).Given this assumption, the model predicts that languages will make finer semantic distinctions at event endpoints than at their comparatively underattended beginnings.We tested both the perceptual assumption of preferential attention to endpoints, and the linguistic prediction of greater semantic specificity at endpoints.