The Limits of the Machine and the Aesthetics of Languages: A Study on the Inability of Artificial Intelligence to Translate Arabic Rhetoric

Musyirah Saleh; Andi Abdullah Hamzah; Kamaluddin Abu Nawas

doi:10.36668/7aet2m35

Authors

Musyirah Saleh Universitas Muhammadiyah Makassar, Indonesia
Andi Abdullah Hamzah Universitas Islam Negeri Alauddin Makassar, Indonesia
Kamaluddin Abu Nawas Universitas Islam Negeri Alauddin Makassar, Indonesia

DOI:

https://doi.org/10.36668/7aet2m35

Keywords:

Artificial Intelligence, Machine Translation, Arabic Rhetoric, Arabic Language Pedagogy, Dynamic Equivalence.

Abstract

The rapid development of Neural Machine Translation (NMT) and Artificial Intelligence (AI) poses epistemic challenges in translating culture-bound text, particularly Arabic rhetoric (Balaghah). This study investigates the cognitive and structural limitations of AI engines in translating Arabic rhetorical structures into Indonesian, and explores its implications for Arabic language pedagogy. Employing a descriptive-comparative method, a purposive sample of thirty rhetorical texts consisting of ten Quranic verses (isti'arah), ten classical Arabic poems, and ten popular proverbs was evaluated. These texts were processed through three prominent AI engines: Google Translate, DeepL Translate, and ChatGPT (GPT-4), and subsequently compared against authoritative human-vetted translations. The findings reveal a significant inverse correlation between the rhetorical density of the text and the semantic accuracy of the machine translation output. While Google Translate and DeepL suffered from rigid literalism by processing tokens as isolated lexical data, ChatGPT displayed relatively better pragmatic adaptability but fundamentally failed to grasp dynamic equivalence and contextual appropriateness (muqtadha al-hal). This failure stems from the machines’ lack of socio-cultural awareness and embodied experience. From a pedagogical perspective, this study moves the scholarship forward by demonstrating that linguistic intuition remains an irreplaceable human bastion against total digital automation. Consequently, this study realistically recommends that Arabic language educators and curriculum developers integrate machine translation criticism into Balaghah instruction to foster students' critical thinking and rhetorical appreciation in the digital era

Downloads

Download data is not yet available.

References

Al-Amri, M. S. (2025). Semiotics of religious texts in the digital age: A critical review of neural machine translation on Quranic rhetoric. Journal of Arabic Theological and Linguistic Studies, 18(1), 45–67. doi.org

Al-Sallab, A., Baly, R., Hajj, H., Shaban, K., El-Hajj, W., & Badaro, G. (2015). Deep learning models for sentiment analysis in Arabic. Proceedings of the Second Workshop on Arabic Natural Language Processing, 9–17. doi.org

Ali, A., & Al-Onaizan, Y. (2020). Challenges and advancements in Arabic machine translation: A comprehensive review. Journal of Artificial Intelligence Research, 68, 432–456. doi.org

Creswell, J. W., & Creswell, J. D. (2018). Research design: qualitative, quantitative, and mixed methods approaches (5th ed.). SAGE Publications. [1]

Emran, M. A., Al-Kabi, M. N., & Al-Ayyoub, M. (2021). Arabic machine translation: A survey with challenges and future directions. IEEE Access, 9, 142310–142328. doi.org

Guellil, I., Adeel, A., Azouaou, F., Benali, R., & Hachani, A. (2019). Arabic natural language processing: An overview of modern tools, techniques, and challenges. Cognitive Computation, 11(5), 611–632. doi.org

Hadi, M. S., Abrori, M. S., & Zahra, D. N. (2020). Pengembangan Multimedia Interaktif Macromedia Flash Profesional 8 Pada Pembelajaran Al-Qur’an Hadis Kelas X Semester Genap Di Man 1 Yogyakarta. At-Tajdid : Jurnal Pendidikan Dan Pemikiran Islam, 3(2), 148. https://doi.org/10.24127/att.v3i2.1123

Hadi, M. S., Amalia, M., Rose, S., Tinggi, S., Islam, A., & Syuhada, M. (2025). Heutagogical Self-Directed Learning in an Indonesian Islamic Senior High School : Teacher Scaffolding , Digital Practices , and Policy Implications. 5(2), 239–252. https://doi.org/10.14421/hjie.52-08

Hadi, M. S., Amrullah, A. K., & Anwar, K. (2024). Journal of Islamic Scriptures in Non-Arabic Societies Membentuk Peserta Didik Menjadi Pribadi Ulul al-bab Prespektif. 1(1), 61–62.

Hadi, M. S., Setiawan, V., & Hidayah, P. M. (2025). The Effect of Implementing Deep Learning on Improving Students ’ Cognitive Abilities in the Independent Curriculum at Yogyakarta High Schools. 7(1). https://doi.org/10.25217/jrie.v7i1.5637

Hassan, H., Ansari, A., & Chang, M. (2018). Achieving human parity on automatic Arabic to English news translation. Microsoft Research Technical Report, MSR-TR-2018-22, 1–18.

Hussein, A. Y. (2024). Beyond large language models: The necessity of embodied experience in cultural and idiomatic translation. International Journal of Translation and Computational Linguistics, 32(3), 211–229. doi.org

Mansour, R., & Roberts, P. (2024). Deciphering sarcasm and irony: Cognitive limitations of artificial intelligence in Middle Eastern rhetorical translation. Lingua Digitalis, 29(2), 154–176. doi.org

Zaidan, O. F. (2023). Evaluating neural machine translation for Arabic cultural and rhetorical texts. Computational Linguistics and Arabic Language Journal, 15(2), 89–104. doi.org