Article In: FORUM: Online-First Articles
The role of prompt engineering in achieving functional equivalence in Arabic-English institutional translation
This content is being prepared for publication; it may be subject to changes.
Abstract
The expansion of Generative AI has introduced diverse architectures for
interlingual tasks. This study evaluates DeepSeek and ChatGPT-5 in translating
governmental correspondence from Arabic into English using a Palestinian Ministry of
Culture corpus. Employing House, J. 2015. Translation Quality Assessment: Past and Present. London: Routledge.
Translation Quality Assessment model, it examines functional equivalence through Field,
Tenor, and Cultural Filter. Findings indicate that architectural differences are less
influential than prompting strategies; both models required context-aware instructions to
avoid literalism and adapt Arabic honorifics to English administrative conventions.
However, because these systems are probabilistic, prompting shapes rather than guarantees
consistent performance. Results show that prompt engineering enhances functional
translation quality, yet its effectiveness depends on users’ expertise in translation and
bureaucratic discourse. The study argues that successful institutional translation depends
less on model selection than on strategic mediation through prompt engineering, which
encodes register and institutional norms while recognizing that prompting remains an
interpretive act capable of privileging particular cultural readings.
Résumé
L’essor de l’IA générative a donné lieu à diverses architectures
pour les tâches de traduction interlinguistique. Cette étude évalue DeepSeek et ChatGPT-5
dans la traduction de correspondances administratives gouvernementales de l’arabe vers
l’anglais, à partir d’un corpus du ministère palestinien de la Culture. S’appuyant sur le
modèle d’évaluation de la qualité de la traduction de House, J. 2015. Translation Quality Assessment: Past and Present. London: Routledge., elle examine l’équivalence fonctionnelle à travers les dimensions
du champ, de la relation et du filtre culturel. Les résultats indiquent que les
différences architecturales sont moins déterminantes que les stratégies de formulation des
invites (prompting) ; les deux modèles nécessitent des instructions tenant compte du
contexte afin d’éviter le littéralisme et d’adapter les formules honorifiques arabes aux
conventions administratives de l’anglais. Toutefois, étant donné le caractère probabiliste
de ces systèmes, le prompting façonne les performances sans pour autant en garantir la
constance. Les résultats montrent que l’ingénierie des invites (prompt engineering)
améliore la qualité fonctionnelle de la traduction, mais que son efficacité dépend de
l’expertise des utilisateurs en matière de traduction et de discours bureaucratique.
L’étude soutient que la réussite de la traduction institutionnelle dépend moins du choix
du modèle que d’une médiation stratégique par l’ingénierie des invites, laquelle encode le
registre et les normes institutionnelles, tout en reconnaissant que le prompting demeure
un acte interprétatif susceptible de privilégier certaines lectures culturelles.
Article outline
- 1.Introduction
- 2.Literature review
- 2.1The evolution of machine translation in institutional contexts
- 2.2Translating institutional discourse: Genre and register constraints
- 2.3Methodological approaches to evaluating translation quality
- 2.4Prior evaluations of AI in Arabic-English translation
- 3.Theoretical framework
- 3.1House’s translation quality assessment model
- 3.2Register analysis in institutional contexts
- 3.3Functional equivalence and the comparative process
- 3.4The cultural filter and AI adaptation
- 4.Methodology
- 4.1Research design and corpus selection
- 4.2Experimental procedure and prompting strategies
- 4.3Data analysis and reliability measures
- 5.Data analysis
- 5.1Lexical choices in relation to field
- 5.1.1Analysis of lexical choices
- 5.2Tenor/level of formality
- 5.2.1Analysis of formality and tenor shifts
- 5.3Cultural filter and functional equivalence
- 5.3.1Analysis of cultural adaptation
- 5.1Lexical choices in relation to field
- 6.Discussion
- 7.Limitations of the study
- 8.Conclusion
- 9.Recommendations
- 9.1For the academic field
- 9.2For the industry and AI developers
- Disclosure statement
- Author queries
References
References (37)
Abbaszade, M., V. Salari, S. S. Mousavi, M. Zomorodi, and X. Zhou. 2021. “Application of Quantum Natural Language Processing for Language Translation.” IEEE Access 9: 130434–130448.
Abudalfa, S. I., F. J. Abdu, and M. M. Alowaifeer. 2024. “Arabic Text Formality Modification: A Review and Future Research Directions.” IEEE Access 12: 185117–185148.
Ajax, R., A. Luz, and B. Bamigboye. 2025. “Natural Language Processing (NLP) in AI-Driven Memory Recall Systems.” ResearchGate. [URL]
Aldawsari, A. H. 2023. “Comparing the Performance of Google Translate and SYSTRAN on Arabic Lexical Ambiguity.” AWEJ for Translation & Literary Studies 7(3): 19–34.
Alenezi, A. M. 2024. “Error Analysis of Neural Machine Translation in Technical Texts: Google Translate as a Case Study.” Journal of the North for Humanities 9(2): 167–181.
Al-Khresheh, M. 2025. “A Back-Translation Analysis of AI-Generated Arabic–English Texts Using ChatGPT: Exploring Accuracy and Meaning Retention.” Dragoman Journal of Translation Studies 17: 97–117.
Almahasees, Z., S. Meqdadi, and Y. Albudairi. 2021. “Evaluation of Google Translate in Rendering English COVID-19 Texts into Arabic.” Journal of Language and Linguistic Studies 17(4): 2065–2080.
Al-Sahlany, Q. A., and H. A. Al-Husseini. 2010. “Kinship Terms in English and Arabic: A Contrastive Study.” Majallah Universitas Babilonia 18(3): 709–726.
Alzain, E., K. A. Nagi, and F. Algobaei. 2024. “The Quality of Google Translate and ChatGPT English–Arabic Translation: The Case of Scientific Text Translation.” Forum for Linguistic Studies 6(3): 837–849.
Chatzikoumi, E. 2020. “How to Evaluate Machine Translation: A Review of Automated and Human Metrics.” Natural Language Engineering 26(2): 137–161.
Freitag, M., et al. 2021. “Experts, Errors, and Context: A Large-Scale Study of Human Evaluation for Machine Translation.” Transactions of the Association for Computational Linguistics 9: 1460–1474.
Gao, R., Y. Lin, N. Zhao, and Z. G. Cai. 2024. “Machine Translation of Chinese Classical Poetry: A Comparison among ChatGPT, Google Translate, and DeepL Translator.” Humanities and Social Sciences Communications 11(1): Article 835.
Ghafouripour, S., and R. Eslamieh. 2018. “A Translation Quality Assessment of Two English Translations of Rubaiyat of Omar Khayyam Based on Julian House’s Model.” International Journal of English Language & Translation Studies 6(2): 217–226.
Hutchins, W. J. 2003. “The Development and Use of Machine Translation Systems.” International Journal of Translation 15(1): 5–26.
Jiang, X., F. Li, H. Zhao, J. Qiu, J. Wang, J. Shao, and T. Chen. 2024. “Long-Term Memory: The Foundation of AI Self-Evolution.” arXiv.
Jiao, W., W. Wang, J. T. Huang, X. Wang, S. Shi, and Z. Tu. 2023. “Is ChatGPT a Good Translator? Yes, with GPT-4 as the Engine.” arXiv.
Kang, J. 2019. “Institutional Translation.” In Routledge Encyclopedia of Translation Studies, 3rd edition, ed. by M. Baker and G. Saldanha, 256–261. London: Routledge.
Khoshafah, F. 2023. “ChatGPT for Arabic–English Translation: Evaluating the Accuracy.” Research Square.
Kimera, R., Y. S. Kim, and H. Choi. 2024. “Advancing AI with Integrity: Ethical Challenges in Neural Machine Translation.” arXiv.
Kocmi, T., and C. Federmann. 2023. “Large Language Models Are State-of-the-Art Evaluators of Translation Quality.” In Proceedings of the 24th Annual Conference of the European Association for Machine Translation, ed. by M. Nurminen et al., 193–203. Tampere, Finland: European Association for Machine Translation. [URL]
Kojima, T., S. S. Gu, M. Reid, Y. Matsuo, and Y. Iwasawa. 2022. “Large Language Models Are Zero-Shot Reasoners.” Advances in Neural Information Processing Systems 35: 22199–22213.
Liu, C., C. Xie, H. He, Y. Zhu, J. Li, and C. Zhang. 2022. “Application of J. House Translation Quality Assessment Model in Corporation’s Web Page Translation.” Journal of Language Testing & Assessment 5: 60–64.
Lockhart, E. N. S. 2025. “Creativity in the Age of AI: The Human Condition and the Limits of Machine Generation.” Journal of Cultural Cognitive Science 9(1): 83–88.
Metwally, A. A., W. R. A. Bin-Hady, and E. Asiri. 2024. “Assessing AI-Powered Translation Quality: Insights from the Translation of the Farewell Sermon.” International Journal of Language Education 8(4): 836–853.
Mohammed, T. A. S. 2025. “Evaluating Translation Quality: A Qualitative and Quantitative Assessment of Machine and LLM-Driven Arabic–English Translations.” Information 16(6): Article 440.
Shehab, E. 2004. “The Translatability of Terms of Address in Najib Mahfouz’s Ziqaq Al-Midaq into English.” An-Najah University Journal for Research — B (Humanities) 19(1): 315–327.
Siregar, R., H. Subagiharti, D. Syafitri Handayani, D. Hutagaol, A. L. Hasibuan, N. Nuraida, and E. Sipahutar. 2025. “A Comparative Analysis of Google Translate and ChatGPT Results Based on Translation Quality Assessment (TQA).” International Journal of Educational Research Excellence 4(1): 328–337.
Song, C., and Q. Kao. 2025. “Comparing Generative AI and Advanced Learners in Translation Tasks Using the House Model: Similarities and Pedagogical Implications.” International Journal of Chinese Language Teaching 6(3): 25–52.
Tahseen, W. M., & S. H. Hussein. 2024. “Investigating Machine Translation Errors in Rendering English Literary Texts into Arabic.” Integrated Journal for Research in Arts and Humanities 4(1): 68–81.