"Pre-editing+MT+Post-editing" As Applicable Translation Model

Machine translation(MT),first put forward in the early 1930 s,has been maintaining an upward momentum despite twists and turns,especially since the advent of computers.In recent years,the neural machine translation(NMT)approach has emerged to improve the quality of MT,but still the quality is non-comparable to that of human translation(HT).To combine the advantages of both MT and HT,the "MT + post-editing(PE)" model has been introduced to translation practices and plenty of studies have been conducted on PE.By contrast,pre-editing has not got enough academic attention.By taking the translation of Korean Perspectives on International Law as an example and adopting comparative analysis,this thesis explores the application of the "pre-editing + MT + PE" model in E-C translation and mainly focuses on specific strategies on pre-editing and PE.As for pre-editing,it is suggested to ensure that the source text uses proper spelling,punctuation,grammar,and sentence structure,and make a glossary if necessary.As for PE,it is advisable to revise machine translation results from three levels: 1)at the word level,the commonest errors are pertinent to mistranslated words,unintended omissions and overtranslation;2)at the sentence level,structure division,tense and voice;3)at the passage level,coherence and cohesiveness.By adding pre-editing to the “MT + PE” model,the thesis hopes to provide some useful insights from a new perspective on how to improve machine translation results with human intervention.