How Can Metaphor not Handle Anomaly? Metaphor Detection with Anomalous Text

Anonymous

TL;DR: Dataset generation and classification detection for literal, metaphor and syntactic anomaly.

Abstract: Metaphor is essentially literal shifts in meaning, which is manifested as a mismatch between the literal meaning of the target word and its contextual context. In metaphor research, the theory of selection preference violation (SPV) is commonly used to identify metaphor in which the target word occurs less frequently in the surrounding words in its context, yielding a mismatch. Researchers are mainly concerned with considering such collocational mismatch as a metaphorical expression, yet they tend to overlook that collocational mismatch may also be a syntactic anomaly. Syntactic anomaly are mainly found in grammatical structures or grammatical rules, which are manifested as irregularities in sentence structure, non-compliance with grammatical rules, or deviations from usual linguistic expressions. In this paper, we integrate syntactic anomaly into the study of metaphor detection. Specifically, we craft a prompt. Based on this prompt, we use GPT-3 to generate a dataset containing literal, metaphor, and syntactic anomaly, called the LMA. We test our dataset in a series of related experiments. We explore the relationship between literal, metaphor and syntactic anomal, as well as the role of introducing SPV. We provide experimental analysis.

Paper Type: long

Research Area: Semantics: Sentence-level Semantics, Textual Inference and Other areas

Contribution Types: Data resources

Languages Studied: English