Ключевые слова
Аннотация
The growing role of artificial intelligence in literary translation has prompted renewed attention to questions of translator style and linguistic creativity.
While AI systems are now capable of producing grammatically sound and fluent translations, whether they can replicate the distinctive stylistic signatures of human translators remains an open question.
This study adopts a corpus-based approach to compare lexical choices in human and AI-generated English translations of literary texts.
A dataset of selected prose passages was compiled and analysed using both qualitative and quantitative methods, with a focus on lexical diversity, repetition patterns, synonym usage, collocational behaviour, and overall stylistic consistency.
The results show that human translators exhibit greater lexical variety and more contextually nuanced word selection, whereas AI-generated translations tend to rely on frequently occurring vocabulary and structurally conventional phrasing.
Although AI outputs maintain a high level of semantic coherence, their lexical preferences often diminish stylistic individuality and weaken the expressive quality associated with literary writing.
The study contends that translator style remains a meaningful measure of literary translation quality—one that cannot be fully captured by criteria such as accuracy or fluency alone.
By integrating corpus linguistic tools with translation theory, this research provides a workable framework for identifying stylistic differences between human and machine translation outputs and contributes to the broader scholarly conversation on AI's place in literary practice.