Analyzing Machine Translation Based on Arabic Translation Theory: A Systematic Literature Review

Abstract

This study aims to systematically review various studies related to machine translation (MT) in the context of Arabic, with an emphasis on the translation theory approach. The method used is Systematic Literature Review (SLR) with reference to the PRISMA guidelines. The review was conducted on ten scientific articles published between 2020 and 2025 and relevant to the topic of Arabic MT, both in terms of systems, text types, and theoretical approaches. The results show that Google Translate is still the most reviewed translation tool, although other systems such as DeepL and ChatGPT are starting to show superior performance especially in terms of idioms, grammatical structures, and pragmatic contexts. The evaluation was conducted using various approaches, including Newmark's theory, Nababan, as well as readability and user perception metrics. The study also identified some key challenges, such as mismatches in nahwu-sharf structures, idiomatic errors, and limitations in understanding cultural contexts. The study concludes that MT systems cannot yet replace the role of human translators completely, especially in complex Arabic texts. The main contribution of this study is to provide a mapping of the current literature as well as recommendations for future research directions in the development of theory-based MT systems and Arabic contexts