MACHINE TRANSLATION POST-EDITING THROUGH A HERMENEUTIC LENS

This article explores how hermeneutic principles – contextual reading, the hermeneutic circle, and interpreter subjectivitycan enrich the post-editing of statistical and neural machine translation (SMT/NMT). A mixed-method study combined hermeneutic textual analysis, think-aloud protocols, and error annotation on parallel SMT/NMT outputs. Results show that while NMT reduces surface errors, both paradigms still require human intervention to restore omitted meanings, pragmatic tone, and cultural references. Experienced post-editors drew consciously on hermeneutic reasoning, supporting the claim that post-editing is an interpretive – not merely corrective – task. Recommendations include integrating translational hermeneutics into post-editor training and refining quality guidelines to foreground meaning fidelity.