Metaphor Detection via Explicit Basic Meanings Modelling

TL;DR

This paper introduces a novel metaphor detection approach that models basic word meanings based on literal annotations, outperforming existing methods and reaching near the theoretical upper bound on benchmark datasets.

Contribution

It proposes explicitly modeling basic meanings from literal annotations, aligning with linguistic theories, and demonstrates significant performance improvements.

Findings

Outperforms state-of-the-art by 1.0% in F1 score

Reaches the theoretical upper bound on VUA18 benchmark

Highlights the importance of modeling basic meanings

Abstract

One noticeable trend in metaphor detection is the embrace of linguistic theories such as the metaphor identification procedure (MIP) for model architecture design. While MIP clearly defines that the metaphoricity of a lexical unit is determined based on the contrast between its \textit{contextual meaning} and its \textit{basic meaning}, existing work does not strictly follow this principle, typically using the \textit{aggregated meaning} to approximate the basic meaning of target words. In this paper, we propose a novel metaphor detection method, which models the basic meaning of the word based on literal annotation from the training set, and then compares this with the contextual meaning in a target sentence to identify metaphors. Empirical results show that our method outperforms the state-of-the-art method significantly by 1.0\% in F1 score. Moreover, our performance even reaches the theoretical upper bound on the VUA18 benchmark for targets with basic annotations, which demonstrates the importance of modelling basic meanings for metaphor detection.