A Study of the English Translation of Arabic Sharia Terms in Legislative Documents: Human versus Artificial

Legislative texts are drafted in a special-purpose style that is indivisible from the countries’ cultures and legal systems. Translation of authoritative text-related rules and regulations necessitates precision of mediating the normative/informative binding content (Šarčević 1997). A Translator of legislative documents such as religious texts, laws, constitutions, conventions or treaties is required to have enough knowledge of the legal system used in the country or the culture he is translating from. Meanwhile, a close examination of cultural, linguistic and textual aspects of the source text (ST) will help translators decode ST complexities. The aim of this research paper is to compare the Arabic legal Sharia terms with their English equivalents, with special focus on cultural and linguistic accuracy. Then, the English human translation will be compared with two AI translations generated by Deepl and Gemini. The corpus of this study is extracted from Arabic Family Laws promulgated in the United Arab Emirates (UAE) and Qatar, and Law of Personal Status issued in Morocco. The paper will discuss and analyze the Artificial intelligence (AI) efficiency and its drawbacks in rendering Sharia terms into English. It will also describe the strategies used in human translation of texts (TT) by applying theoretical approaches to legal translations as special-purpose texts. In dealing with the topic, I will start with a theoretical part that explains the proposition of the paper. After this theoretical part, I will start the application of my proposition to actual legal codes dealing with the four selected SL terms. A comparison of the AI translations with human translations will be attempted. This application will be preceded by an explanation of the methodology used in the analysis and the comparison.