Intertextuality in Human and AI Literary Translation: A Comparative Study Through Kristeva’s Theory and Hatim & Mason’s Framework

Abstract

Grounded in Kristeva’s (1966) theory of intertextuality and Hatim and Mason’s (1990) intertextual framework, the present research addresses the representation of intertextual references in Persian translations of six English novels. The corpus includes translations by three human translators—Hosseini, Badiei, and Karimi Hakkak—and an AI language model, GPT-4. To this end, using literary criticism, the most frequent intertextual elements were identified and compared with the translations. The findings reveal that human translators often employ strategies such as accommodation and explanation, drawing on classical Persian literature to recreate the intertextual layers of the source text, mirroring the original authors’ reliance on earlier literary works. The results further indicate that GPT-4, when effectively prompted, utilizes strategies like adaptation and explanation, showing similar performance to human translators in many cases, though its use of classical Persian literature remains limited. The outcomes highlight that both translation and writing are intertextual processes. While AI language models like GPT-4 have made significant progress, further development is needed to more effectively leverage the potential of classical Persian literature in these models. Utilizing classical literature in translation not only enriches translated works but also plays a key role in preserving target language and culture in the face of global linguistic shifts.