Machine Translation And Post-editing Of The Military Sci-tech Literature From The Perspective Of Relevance Theory

With the prevailing of computer and Internet technology,the combination of efficient machine translation(MT)and post-editing(PE)mode is gaining ground in the translation services.However,in practice,due to the lack of relevant termbase(TB)and translation memory(TM),there are multiple challenges ahead for the translation of professional scientific literature.Just because of this,there still remains the difficulty of translating such texts and providing guidance for translators.From the perspective of the Relevance Theory,this thesis is an exploration of the translating process in MT+PE mode in the military sci-tech English-Chinese translation.It consists of 6 chapters.Apart from the introduction and the conclusion,Chapter Two reviews the relevant studies at home and abroad.On the basis of the Relevance Theory proposed by Sperber and Wilson and Gutt’s relevance-theoretic approach to translation,Chapter Three tentatively constructs “the author-MT+PE-reader” information-processing model.With the guidance of this model,Chapter Four and Chapter Five probe into the original author’s ostension in MT and the translator’s inference in PE respectively.In addition,given that the military sci-tech literature is full of terms,abbreviations and long and complex sentences without parallel texts for reference,this thesis intends to clarify the MT’s identification process to the author’s informative intention and guide the post-editor to make inference about the author’s communicative intention,which finally achieves the optimal relevance among the original author,MT+PE and the target reader.According to the new information-processing model,MT system makes lexical choice and identifies syntactic structure by identifying the original author’s ostension in the part-of-speech,meaning of words,context and preposition in the source text,then the translator makes different inferences at lexical level and syntactic level based on the MT output including logical inference,knowledge inference and grammatical inference,thereby the reader is able to spend the minimal cognitive efforts to obtain the maximal relevance in translation.Based on the theoretical deduction and example analysis,this thesis aims to provide a feasible theoretical framework for the MT+PE mode and make reference for the study of translating military sci-tech texts in the future.