Identifying Speakers and Addressees of Quotations in Novels with Prompt Learning

TL;DR

This paper introduces prompt learning methods for identifying speakers and addressees in novel quotations, utilizing a newly annotated Chinese corpus and demonstrating superior performance over large language models.

Contribution

It presents the first Chinese quotation corpus with detailed annotations and proposes prompt learning techniques for speaker and addressee identification in literary texts.

Findings

Prompt learning methods outperform zero-shot large language models.

Annotated Chinese corpus enables better quotation element extraction.

Methods are effective on both Chinese and English datasets.

Abstract

Quotations in literary works, especially novels, are important to create characters, reflect character relationships, and drive plot development. Current research on quotation extraction in novels primarily focuses on quotation attribution, i.e., identifying the speaker of the quotation. However, the addressee of the quotation is also important to construct the relationship between the speaker and the addressee. To tackle the problem of dataset scarcity, we annotate the first Chinese quotation corpus with elements including speaker, addressee, speaking mode and linguistic cue. We propose prompt learning-based methods for speaker and addressee identification based on fine-tuned pre-trained models. Experiments on both Chinese and English datasets show the effectiveness of the proposed methods, which outperform methods based on zero-shot and few-shot large language models.