نوع مقاله : مقاله علمی پژوهشی
نویسندگان
1 دانشیار، گروه زبان و ادبیات عربی، دانشگاه خوارزمی، تهران، ایران
2 دانشجوی کارشناسی ارشد مترجمی زبان عربی، دانشگاه خوارزمی، تهران، ایران
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
This research undertakes a comprehensive comparative evaluation of the performance exhibited by three prominent Machine Translation (MT) systems—Google Translate, Microsoft Translator, and DeepL—in the specialized domain of translating complex legal texts from Arabic to Persian, utilizing Human Translation (HT) as the primary benchmark for quality assessment. To ensure both representativeness and the necessary terminological precision required in legal discourse, the study systematically utilizes a curated selection of diverse legal passages derived from Shaker Kasrai’s Translation of Documents, Certificates, and Legal Texts; this specific corpus was selected due to its recognized structural rigor and the author’s established judicial expertise. The evaluation process rigorously focuses on three pivotal criteria: the accuracy of specialized legal terminology, the maintenance of semantic coherence across sentence boundaries, and strict adherence to the formal, prescriptive tone inherent in legal documentation. The findings indicate that, while the examined MT systems generally succeed in preserving the overarching structural integrity of the source text, they consistently demonstrate a qualitative inferiority when compared to human translation, particularly in their capacity to reproduce precise legal nomenclature and convey the subtle semantic nuances embedded within the source material. A critical deficiency identified throughout this analysis is the persistent inability of these automated systems to effectively comprehend and translate the complex etymological and contextual foundations that underpin specialized legal vocabulary.
کلیدواژهها [English]
)