Research Article

ChatGPT and Human Translation: Evaluating Linguistic Accuracy, Stylistic Fidelity, and Cultural Resonance in Translating John Steinbeck’s The Pearl

Authors

  • Ameena Zeyad Bani Mustafa MA Student, Department of English Language and Literature, Faculty of Arts, Jerash University, Jordan
  • Haitham Mohd Khair ALYousef Assistant Professor, Department of English Language and Translation, Faculty of Arts, Jerash University, Jordan

Abstract

Artificial intelligence has expanded the use of machine translation in multilingual communication. Although ChatGPT produces fluent translations rapidly, its suitability for literary translation remains contested because literature requires more than semantic transfer. It depends on rhythm, imagery, symbolism, narrative voice, and culturally embedded meaning. This study compares ChatGPT’s Arabic translation of John Steinbeck’s The Pearl (1947) with Yousef Ibrahim Al-Jahmani’s professional human translation. It evaluates the AI output across linguistic accuracy, stylistic fidelity, and cultural integrity. The study uses a mixed qualitative and quantitative design. Qualitative analysis applies Halliday and Hasan’s cohesion theory to examine reference, conjunction, and lexical cohesion, while Nida’s dynamic equivalence informs the assessment of emotional and cultural effect. Representative extracts were assessed for semantic fidelity and stylistic and cultural difficulties. ChatGPT achieved semantic accuracy in 84% of the examples. Cohesion-related problems occurred in 28% of translated sentences, while 16% showed loss of imagery or figurative language and 12% contained cultural misunderstandings or inaccuracies. The findings indicate that ChatGPT preserves much of the novella’s propositional content, especially in direct passages. Its limitations emerge when meaning relies on literary form rather than lexical equivalence. Repetitive conjunctions, literal lexical choices, inconsistent treatment of names, and weakened figurative language can disrupt narrative rhythm and diminish the source text’s emotional and cultural force. Comparison with the human translation suggests that literary translation requires interpretive decisions not captured by semantic accuracy alone. The study therefore supports a hybrid model in which ChatGPT generates preliminary drafts and human translators provide the stylistic, cultural, and interpretive revision necessary for literary translation.

Article information

Journal

International Journal of Linguistics, Literature and Translation

Volume (Issue)

9 (9)

Pages

42-58

Published

2026-09-01

How to Cite

Ameena Zeyad Bani Mustafa, & Haitham Mohd Khair ALYousef. (2026). ChatGPT and Human Translation: Evaluating Linguistic Accuracy, Stylistic Fidelity, and Cultural Resonance in Translating John Steinbeck’s The Pearl. International Journal of Linguistics, Literature and Translation, 9(9), 42-58. https://doi.org/10.32996/ijllt.2026.9.9.5

Publication history

  1. Submitted
  2. Published

Author information

Ameena Zeyad Bani Mustafa
MA Student, Department of English Language and Literature, Faculty of Arts, Jerash University, Jordan
Jordan
Haitham Mohd Khair ALYousef
Assistant Professor, Department of English Language and Translation, Faculty of Arts, Jerash University, Jordan
Jordan

Downloads

Views

47

Downloads

5

Keywords:

Artificial intelligence, ChatGPT, literary translation, Arabic translation, linguistic accuracy, stylistic fidelity, cultural integrity, human translation