Machine Translation Limitations And Post-editing Solutions

With the development of computer science,machine translation has been widely applied in translation activities.Despite its high efficiency and cost-effectiveness,machine translation produces target texts with a quality far below that of professional translators and the standard required by clients.In this regard,post-editing is an indispensable part of machine-aided translation projects,which ensures a balance between translation efficiency and quality.The thesis delves into machine translation limitations and post-editing solutions based on the translation of UN Chronicle.The analysis is conducted from three aspects,namely,lexicon,syntax and text.It is found out that typical errors made by machine translators include: on the lexical level,inaccurate terminology,wrong semantic meanings and improper collocations;on the syntax level,unadjusted passive voice,word-for-word translation of long sentences and inflexible word order as well as a lack of coherence throughout the text.The thesis,therefore,suggests corresponding post-editing solutions: on the lexical level,ensuring correct terms according to termbases and style guides,choosing context-specific meanings and referring to trustworthy resources like UNTERM and Linguee;on the syntax level,converting passive to active voice,dividing long sentences into shorter ones and adjusting word order;in terms of coherence,emphasizing primary information and using cohesive devices.