Keeping an eye on the data: Metonymies and their patterns

Abstract

This paper outlines a corpus-based method for the analysis of metonymic expressions based on a series of quantitative and qualitative analyses. While an intuitive approach to metonymy successfully identifies lexical items which have metonymic extensions, intuition alone cannot settle the question how these extensions map onto linguistic form. Consider the expression set all hearts on fire, which has been claimed to instantiate the conceptual metonymy the heart for the person. Intuitively, it is hard to tell whether the quantifier all has something to do with the figurative interpretation. In contrast, a corpus-linguistic analysis brings to light that quantified heart (some hearts, a few hearts, many hearts) is by default interpreted metonymically. This suggests that the figurative extensions of a given lexical item correlate with distinctive patterns. These patterns are solely determinable through analysis of authentic data. It turns out that in the investigated data, figurative usages made up more that 40% for all body lexemes; more that 65% of the data under investigation is organized in a limited array of patterns. These patterns expose a close correlation of form and meaning. Thus, collocation is a major clue to the interpretation of metonymic expressions. Furthermore, literal and non-literal examples contrast significantly with respect to neighboring word classes. I draw four conclusions from this pilot study: Metonymy can be analyzed through corpus analysis of source domain lexis. Metonymic expressions tend to be organized in patterns – these patterns trigger a specific metonymy. Metonymic expressions differ from literal expressions with respect to collocation and with respect to colligation.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 42 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.