Verifying Claims About Metaphors with Large-Scale Automatic Metaphor Identification

TL;DR

The paper addresses the problem of empirically verifying five cognitive-linguistic claims about metaphor usage in a large-scale corpus. It combines MisNet-based metaphor identification (utilizing MIP and SPV) with data from CC-100 and ancillary datasets to quantify concreteness, imageability, and familiarity of metaphorical object nouns, and to examine how emotion and subjectivity correlate with metaphor usage. Key findings show that direct objects of metaphorical verbs tend to be less concrete, imageable, and familiar, and that metaphor usage is elevated in emotionally charged and subjective sentences. The work demonstrates the feasibility and value of corpus-based verification for metaphor theories, while noting limitations related to language scope, automatic labeling accuracy, and proxy measures for familiarity that invite future refinement.

Abstract

There are several linguistic claims about situations where words are more likely to be used as metaphors. However, few studies have sought to verify such claims with large corpora. This study entails a large-scale, corpus-based analysis of certain existing claims about verb metaphors, by applying metaphor detection to sentences extracted from Common Crawl and using the statistics obtained from the results. The verification results indicate that the direct objects of verbs used as metaphors tend to have lower degrees of concreteness, imageability, and familiarity, and that metaphors are more likely to be used in emotional and subjective sentences.