Reference-point reasoning is a pervasive cognitive phenomenon intrinsic to many domains of human activity. However, very little is known about linguistic aspects of this phenomenon. This paper elaborates the reference-point model by applying it to lexical semantics and, more specifically, to the semantics of dimensional adjectives. It is argued that a panoply of reference points may be used to anchor conceptual specifications of adjectives, prototypes being only a special case of the reference-point mechanism. For example, dimensional adjectives may be interpreted vis-à-vis an average value of the property (norm), endpoints of the scale and dimensions of the human body (ego). Each of these reference points motivates crucial semantic and functional properties of dimensional adjectives.