Lexical opposition: “canonical ” and “peripheral ” antonymy

Whenever antonymy, or lexical opposition, is discussed, scholars refer to the phenomenon that some antonym pairs are better examples of antonymy than others and that these are reliably identified as such by native-speakers. The criteria upon which this distinction is based are only rarely discussed and while some isolated, and theoretically diverse, characteristics of “good ” antonyms are proposed in several studies of antonymy (Cruse 1986, Jones 2002, Murphy 2003) they have not been examined as an inter-related bundle of features of the category of antonymy. Starting out from these characteristics and the overwhelming evidence that the connection be-tween “good ” antonyms is particularly salient, the current project approaches the subject from a cog-nitive-pragmatic angle, incorporating the processing and structure of the opposition in the mental lexicon as well as a pragmatic, usage-based approach. In this paper I will take up this distinction between “canonical ” and “peripheral ” antonymy and evaluate the characteristics put forward in the literature for their applicability and usefulness in decid-ing whether an antonym will be judged “canonical ” or “peripheral”. The criteria considered are the three criteria (a-c) put forward by Cruse (1986) as well as two additional ones (d & e): a) purity of opposition → male:female vs. king:quee