A new research paper explores gendered language in science fiction and general fiction corpora, using a RoBERTa model to analyze how gendered terms are substituted. The study found that while both genres exhibit gender permeability, science fiction shows a greater tendency to cross the gender binary in both directions. The research also indicates that science fiction is more likely to degender terms, moving them toward neutral language, and that masked 'woman' prompts more diverse substitutions than masked 'man'. This work provides linguistic evidence for thematic claims in science fiction studies and offers a computational pipeline for analyzing other categorical vocabularies like race and class. AI
IMPACT Provides a computational pipeline for analyzing categorical vocabularies in literature, applicable to race and class.
RANK_REASON Academic paper published on arXiv detailing computational linguistic analysis. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →