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 posteditors 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.