Ms. Faezeh Moteshaker, Dr. Hossein Bahri,
Volume 27, Issue 2 (9-2024)
Abstract
This study investigates syntactic and thematic strategies by a translation model based on artificial intelligence, ChatGPT, in English to Persian literary translation. Mona Baker's (2011) theory of linear dislocation aims to assess whether and to what extent four major strategies (voice change, verb change, nominalization, and extraposition) occur in AI-generated text. The data set includes Thomas Hardy's short story Absent Mindedness in a Parish Choir and its full Persian translation produced by ChatGPT. A qualitative comparative method was adopted, in which Baker's scheme was taken as the starting point for text analysis. Forty four segments were identified and examined, each demonstrating the use of at least one of the mentioned strategies. The findings show that ChatGPT makes considerable use of all four strategies implicitly: verb change (31.82%), voice change (27.27%), nominalization (22.73%), and extraposition (18.18%). These results preserved thematic structure and communicative coherence in most cases. The study identifies ChatGPT's capacity to respond to functional translation in literary translation with idiomatic, rhythmic, and rhetorical fidelity. It suggests Baker's strategies are still relevant not only for human translators but also as a valuable instrument for evaluating and post-editing AI translations, especially literature. This study investigates syntactic and thematic strategies by a translation model based on artificial intelligence, ChatGPT, in English to Persian literary translation. Mona Baker's (2011) theory of linear dislocation aims to assess whether and to what extent four major strategies (voice change, verb change, nominalization, and extraposition) occur in AI-generated text. The data set includes Thomas Hardy's short story Absent Mindedness in a Parish Choir and its full Persian translation produced by ChatGPT. A qualitative comparative method was adopted, in which Baker's scheme was taken as the starting point for text analysis. Forty four segments were identified and examined, each demonstrating the use of at least one of the mentioned strategies. The findings show that ChatGPT makes considerable use of all four strategies implicitly: verb change (31.82%), voice change (27.27%), nominalization (22.73%), and extraposition (18.18%). These results preserved thematic structure and communicative coherence in most cases. The study identifies ChatGPT's capacity to respond to functional translation in literary translation with idiomatic, rhythmic, and rhetorical fidelity. It suggests Baker's strategies are still relevant not only for human translators but also as a valuable instrument for evaluating and post-editing AI translations, especially literature.
Publisher
