On the distinction between metonymy and vertical polysemy in encyclopaedic semantics

In cognitive linguistics, metonymy is seen as a fundamental cognitive process where one conceptual entity affords access to another closely associated one. Cases of vertical polysemy have also often been treated as instances of metonymy (see e.g. Radden and Kövecses, 1999). In vertical polysemy a lexical form designates two or more senses that are in a relationship of categorial inclusion – e.g. dog ‘canine’, ‘male canine’. In this paper I present an account of cases of vertical polysemy from the point of view of domain-based encyclopaedic semantics as described in Langacker (1987). I claim that the domain configurations which underlie the broader and narrower meanings of vertical polysemes are very different from those involved in cases of metonymy. Croft (1993) argues that from a Langackerian viewpoint, metonymy involves a shift in the salience of two domains that form parts of a domain matrix against which a given concept is profiled. In cases of vertical polysemy, on the other hand, the relationship between the broader and narrower meanings may be effected in a number of different ways, none of which involve the kind of domain configurations found in metonymy. For example, the narrower ‘male canine’ sense of dog makes reference to an additional domain of SEX, a domain which is not an essential part of the domain structure of the broader ‘canine’ meaning.