Female Villain Language: Gendered Patterns Explored Through a Corpus-based Approach to Role Language Research

This study examines the linguistic characteristics of the speech of female villains in Japanese fictional dialogue, with the aim of determining whether female villains employ a distinct role language or yakuwarigo, that is, a distinct speech style that indexes characterization based on social stereotypes. An original, small-scale corpus was constructed consisting of dialogue from female villains, male villains, and female protagonists (heroines). Each group contains three characters drawn from films and television series with approximately 70-80 utterances per character. The corpus was manually annotated for linguistic features commonly associated with role language, including first- and second-person pronouns, verb conjugation, sentence-final elements, and gender-coded forms. Quantitative comparison of feature distributions across groups was complemented by qualitative analysis of marked forms. The results indicate that female villains do not exhibit a fixed speech style comparable to established role languages. Instead, female villain speech displays considerable intra-group variation and is characterized by a strategic combination of feminine-coded and masculine-coded forms. In particular, female villains avoid masculine forms in self-reference but adopt them frequently for second-person reference and occasionally for imperatives and negations. In contrast, male villains and heroines show more internally consistent patterns. These findings suggest that female villains do not constitute a distinct role language in the traditional sense but rather make flexible use of socially typified linguistic forms for characterization. On this basis, the study argues for a refinement of the conceptual framework of role language and proposes an approach that emphasizes dynamic indexicality and contextual interpretation of static form-category links, as inspired by gendered language research.