The Detection and Understanding of Fictional Discourse

2024/01/30 by Andrew Piper, Piper, Andrew, Haiqi Zhou +1
Arts and Humanities · Psychology · #Computation and Language (cs.CL) #Discourse Analysis in Language Studies #FOS: Computer and information sciences #Language, Metaphor, and Cognition #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2401.16678

openalex publication_date 2024/01/30 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

In this paper, we present a variety of classification experiments related to the task of fictional discourse detection. We utilize a diverse array of datasets, including contemporary professionally published fiction, historical fiction from the Hathi Trust, fanfiction, stories from Reddit, folk tales, GPT-generated stories, and anglophone world literature. Additionally, we introduce a new feature set of word "supersenses" that facilitate the goal of semantic generalization. The detection of fictional discourse can help enrich our knowledge of large cultural heritage archives and assist with the process of understanding the distinctive qualities of fictional storytelling more broadly.

Related

• Generative AI & Fictionality: How Novels Power Large Language Models
• Event Detection between Literary Studies and NLP. A Survey, a Narratological Reflection, and a Case Study
• Assessing Language Models' Worldview for Fiction Generation
• LFED: A Literary Fiction Evaluation Dataset for Large Language Models
• Information retrieval in folktales using natural language processing