Impact of Target Word and Context on End-to-End Metonymy Detection

TL;DR

This paper reformulates metonymy detection as a sequence labeling task and finds that context words are more influential than target words, especially for domain-specific entities.

Contribution

It introduces a new sequence labeling approach for metonymy detection and analyzes the relative importance of target words versus context.

Findings

Context words are more relevant than target words for detection.

Entity types linked to domain-specific context are easier to identify.

Target word usefulness is limited in the dataset.

Abstract

Metonymy is a figure of speech in which an entity is referred to by another related entity. The task of metonymy detection aims to distinguish metonymic tokens from literal ones. Until now, metonymy detection methods attempt to disambiguate only a single noun phrase in a sentence, typically location names or organization names. In this paper, we disambiguate every word in a sentence by reformulating metonymy detection as a sequence labeling task. We also investigate the impact of target word and context on metonymy detection. We show that the target word is less useful for detecting metonymy in our dataset. On the other hand, the entity types that are associated with domain-specific words in their context are easier to solve. This shows that the context words are much more relevant for detecting metonymy.