Unveiling stylometric signatures in translated children’s literature: Peter Pan in Chinese

View paper on Nature

This study employs a stylometric research design to examine linguistic discrepancies and explore the relationships among three Chinese translations of Peter Pan, a world-renowned work of children’s literature. Classification and clustering experiments were conducted to distinguish the three translations, and attribute selection algorithms were employed to identify the top 10% significant features (33 out of 335) spanning four dimensions: lexical, syntactic, readability, and N-gram. The results show that (1) classification and clustering experiments effectively differentiate the three translations, with Liang’s version consistently exhibiting a distinctive profile, while the Yang & Gu’s and Ren’s versions display a certain degree of similarity; and (2) the most discriminative features distinguishing the translations are primarily concentrated in lexical features and N-gram features.

Recommended citation: He, T., Kong, D., & Li, B. (2026). "Unveiling stylometric signatures in translated children's literature: Peter Pan in Chinese." Humanities and Social Sciences Communications. (Corresponding author: Delu Kong)
Download Paper