An AI Enabled Comparison CDA Using ChatGPT and DeepSeek to Identify Discourse Strategies in Imam Ali’s Al-Shaqshaqiyya Sermon in Nahj al-Balagh
DOI:
https://doi.org/10.31185/wjfh.Vol22.Iss3/Pt2.1754Keywords:
Artificial Intelligence - CDA - Al-shaqshaqiyyahAbstract
This study aims to analyze an old religious text, Imam Ali sermon Al-Shaqshaqiyya by using three critical discourse analysis models. The three models VanDijk’s, Fairclough’s, and Wodak’s have been used to analyze both the original Arabic sermon with its English translation. The study examines AI-generated outputs across three AI systems. The results show that ChatGPT-4 is the most accurate in terms of accuracy, coherence and human-guided CDA in English text. DeepSeek demonstrates significant expertise translating culturally and historically embedded elements of the Arabic text while less skillful in doing so for the English. ChatGPT-3. 5, on the other hand, yields only surface-level interpretations and there is no analytic depth. Cross-linguistic findings suggest that English translations are more accessible, but involve sacrifices of rhetorical force and cultural subtlety found in the Arabic text.
Downloads
References
References
OpenAI. (2023). ChatGPT: Optimizing language models for dialogue. Retrieved from https://openai.com/research/chatgpt
Van Dijk, T. A. (2008). Discourse and context: A sociocognitive approach. Sage Publications.
Fairclough, N. (2015). Language and power (3rd ed.). Pearson Education
Wodak, R. (2001). The discourse of politics in action: Politics as usual. Sage Publications.
Pires, T., Schlinger, E., & Tetreault, J. (2020). How multilingual are multilingual transformers? Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), 3786–3797. https://doi.org/10.18653/v1/2020.acl-main.338
Lippi, M., & Torroni, P. (2016). Evaluation of AI-based systems in natural language processing tasks. Springer.
Bender, E. M., Gebru, T., McMillan-Major, A., & Mitchell, M. (2021). On the dangers of stochastic parrots: Can language models be too big? Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (FAccT). https://doi.org/10.1145/3442188.3445922
Al-Shawi, A., & Mahadi, T. (2012). Difficulties of translation and evaluative idioms in English and Arabic. International Journal of Humanities and Social Science, 2(9), 143–149.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Prof. Basim Muftin Badar , Asst. Prof. Muna AbdualHussein, Prof. Muayyed Juma
This work is licensed under a Creative Commons Attribution 4.0 International License.
