A Corpus-Based Study on the Translation Style of the Web Novel 'A Record of a Mortal's Journey to Immortality'

Against the backdrop of the large-scale international expansion of Chinese web fiction, the rapid adoption of AI translation alongside high-quality human translations has drawn attention to the linguistic features of translated works. Existing research predominantly focuses on macro-level industrial models or qualitative examinations of individual translation strategies, lacking systematic, linguistically grounded empirical analyses of classic translations that have undergone long-term market validation. This paper takes the English translation of 'A Record of a Mortal's Journey to Immortality' as its research subject and employs corpus linguistics methods to comparatively analyze the fan translation published on Wuxiaworld and the commercial translation published on Amazon, examining differences and similarities in lexical density, distribution of content words, handling of culture-loaded terms, and textual readability. The findings reveal that: (1) both translations exhibit higher lexical density than native English novels; (2) translation strategies for high-frequency key words reflect distinct orientation preferences; (3) regarding readability, both translations appropriately control average word length and sentence length, achieving standard reading proficiency, yet display stylistic differences; and (4) through word cloud and keyword analyses, the translations successfully construct a logically coherent Eastern xianxia (immortal cultivation) world. The study demonstrates that these linguistic features can serve as high-quality reference corpora for optimizing AI translation models and provide empirical support for future web fiction translation practice.