Challenges of Post-Editing in English to Arabic Machine Translation of Technical Texts: A Study of Technological and Linguistic Barriers

Abstract

The increasing reliance on machine translation (MT) for English-to-Arabic technical texts presents significant linguistic and technological challenges, necessitating extensive human post-editing. This study examines these challenges by analyzing machine-translated technical texts and assessing the post-editing process undertaken by professional translators. Despite advancements in neural machine translation, English-Arabic translation remains problematic due to syntactic, morphological, and terminological discrepancies between the two languages. The study employs House’s (1997) Translation Quality Assessment (TQA) Model to evaluate machine translation quality and the impact of post-editing interventions. Methodologically, ten technical texts were selected from car and hair dryer manuals and translated using Google Translate. Two professional translators, each holding a PhD in translation, post-edited these texts in a two-stage process, producing a single collaboratively refined version. Semi-structured interviews were then conducted to explore the translators' experiences, the challenges they faced, and their perspectives on the effectiveness of MT tools. The analysis of the interviews revealed key technological and linguistic barriers, including inconsistent terminology, unnatural sentence structures, and difficulties in maintaining semantic and pragmatic accuracy. The findings highlight that MT tools struggle with context-sensitive technical terms, resulting in inaccuracies that demand significant human intervention. Additionally, issues such as word order mismatches, poor handling of Arabic morphology, and ineffective recognition of formal registers contribute to the post-editing workload. The study recommends improvements in MT systems, including enhanced AI-driven context recognition, customizable glossaries, and adaptive learning mechanisms to refine MT accuracy over time. By addressing these gaps, MT tools can better integrate into professional translation workflows, reducing post-editing efforts while improving the quality of English-to-Arabic technical translations.