Metonymy in Context and Communication

Imagine that you're working at a café. You're in the middle of the most hectic period of the day and several guests are waiting to be served their orders. While you're attending to a young couple, your colleague points out to you that 'The ham sandwich in the corner is getting restless'. Given that a ham sandwich is a possible order at your café, you'd be able to immediately identify the referent of the expression 'The ham sandwich' as the man in the corner who ordered a ham sandwich from you a while ago. You'd most likely also interpret your colleague's utterance as a prompt to serve this man his order as soon as possible. This ordinary example illustrates an intriguing phenomenon in human communication: Sometimes we use an expression to refer to something that falls outside its linguistically-specified denotation, something with a clear associative relation linking the linguistically-specified denotation and the contextually-determined denotation. Such use of referential terms is called metonymy. The research project 'Metonymy in Context and Communication' investigates how we understand such expressions in communication with each other. While metonymic uses by adults are well documented, little is known about how metonymy is understood or used by children. A central part of the current project has been to investigate whether young children are communicatively competent enough to cope with metonymy. In collaboration with colleagues at Stanford University, 47 English-speaking children (aged three to five years) and 27 English-speaking adults were tested on their comprehension and production of metonymy (Falkum, Recasens & Clark, revised and resubmitted). Results showed that although children were outperformed by adults, even three-year-olds could both understand and spontaneously produce metonyms in certain circumstances. Our results also suggest that young children may find it easier to produce a metonym than a full description as a referring description in many contexts, and that metonymy may serve as a useful strategy in referring to unfamiliar entities. Another important part of the project has been to develop a cognitively plausible account of how metonymy is processed in the language of adults. In collaboration with Deirdre Wilson I have worked on developing a novel, relevance-theoretic account of metonymy where it is treated as a form of motivated word coinage, based on existing words which provide clues to the speaker?s meaning. On this account metonymy has more in common with other word coinage processes such as noun-noun compounds, denominalisation and nicknames than it has with standard cases of 'figurative' language such as metaphor.

This project investigates the phenomenon of metonymy, i.e. the process whereby an expression is used to refer to (or 'stand for') something that falls outside of the conventional denotation of that expression, and where the conventional and the metonymic denotations stand in a certain relation to each other (e.g. 'The ham sandwich is getting impatient', where 'ham sandwich' refers to the person who ordered it, in the context of a restaurant). Such metonymies have been shown to raise deep issues for semant ic compositionality and cognitive theories of the process of utterance comprehension. This project seeks to investigate a new and largely unexplored area in research on metonymy, with the goal of extending our understanding of how metonymy works. The over all aim of this project is to develop a cognitively plausible account of the comprehension of metonymy. Towards this aim, I will (a) compare different conceptions of metonymy in the literature and consider whether we have to do with a single or several di fferent phenomena; (b) investigate, using the framework of relevance theory (Sperber and Wilson 1986/1995; Carston 2002), the hypothesis that referential metonymy involves a creative use of naming, similar to nicknaming; (c) develop a hypothesis about the role of mental processes of association and inference in the comprehension of metonymy; (d) propose a pragmatic criterion that is able to predict which metonymies are possible in a given context, and (e) investigate whether a relevance-theoretic approach to the comprehension of metonymy might be compatible with the cognitive linguistic hypothesis about the existence of so-called 'metonymic concepts' (e.g. Lakoff and Turner 1989; Radden and Kövecses 1999).