A Corpus-based Study On The Translation Of Chinese Political Documents At The Sentence And Word Levels

Political documents are of their own characteristics.Words and expressions in political documents are precise and formal.Political document translation should be completely faithful to the original expressions.The translation of political documents is important for China in international communication and external propaganda.Political document translation has been studied in vast ways.But most of them are done in a qualitative way,focusing on the translation of terminology and analyzing the source and target texts by studying extracted words or sentences.This method is prescriptive and evaluative and may not provide the complete features of the translated works.The present study approaches political document translation by employing corpus-based methodology.This thesis studies the translation patterns of political documents at the sentence and word levels,using Selected works of Deng Xiaoping as the language material.With the specific aim of investigating the translation patterns,a parallel corpus DXPARA,which contains Selected works of Deng Xiaoping and its translation,was complied and used.DXPARA is aligned at sentence level with Trados, and aligned at word level with the bilingual dictionary and character string intersection method.The following conclusions are reached from the study:(1) At the sentence level,81.87%of the sentences have the 1:1 alignment.As a political document translation,Selected works of Deng Xiaoping adhere to the faithful principle.The 1:N sentence alignments is 15.56%.Chinese sentences of this type are long and translated by means of division.The N:1 sentence alignments is 2.57%.These Chinese sentences are parataxis or logically connected sentences.(2) At the word level,the 1:1 word alignment accounts for 66.7%.These Chinese words have word-for-word correspondence.The 1:N word alignments are 33.1%. These Chinese words tend to be polysemy or multi-category,and they have two or more translation equivalents.The rest 0.2%is 1:Φword alignment.They have no translation equivalents.