| The translator’s style has always been a hot topic in corpus translation studies.With the rapid development of technology in recent years,the study of translator’s style has been expanded at the lexical,syntactic,and semantic or discourse levels.Sentiment analysis techniques are introduced to the discourse level.The sentiment is the soul of literary works,so when translating literary works,the translator should also consider how to reasonably convey the original text’s sentimental message.Researches,by far,seldom choose sentiments as the research object,and most adopt qualitative methods.Quantitative methods are scarcely used in this domain.Based on this,this article adopts Appraisal Theory as the theoretical framework and combines quantitative and qualitative methods to compare the sentimental resources in the original English text of Lolita and two Chinese translations and observe whether or not the translation can accurately convey the sentimental message of the source text.The self-built parallel corpus of this research consists of Lolita,written by the Russian-American writer Vladimir Nabokov and the translations of Zhu Wan and Yu Xiaodan.Sentiment words in the source text and target texts are identified according to the Attitude System of Appraisal Theory.This study compares the original text and the two translations in terms of linguistic features of sentiment words(such as density,high-frequency words,etc.),the sentimental tendency of texts,and sentimental flow patterns to reveal the differences in sentimental resources between the original text and the translated texts,as well as between different translations.The qualitative analysis summarizes situations where inaccuracies of translation emerge.The research results show that: first,there are significant differences between the original text and the translated text in terms of part-of-speech proportion and highfrequency words of sentiment words.The original text is higher than the two translations in the number and density of sentimental words,and Yu’s is higher than Zhu’s.Adjectives account for almost half of all sentimental words,and content words account for 90%.What’s more,the number of idioms in the original text is significantly less than in the two translations.The high-frequency words in the original text are almost all adjectives used to describe the female protagonist,and most of their corresponding Chinese translations also appear in the list of highfrequency words.Second,the overall sentimental tendency of the two translations is consistent with the source text.However,the negative tendencies of the translated versions are stronger than those of the original text,and Zhu’s version is stronger than Yu’s version.Third,the target texts are also consistent with the source text in sentimental flow but differ in intensity.The above three points show that the two translators,Zhu and Yu,can accurately grasp the sentimental tendencies,but the accuracy of the intensity could be better.The situations that lead to the inadequacy of sentiment conveying are summarized into three categories: adding sentiment that is not present in the original text,missing sentiment that is present in the original text,and misunderstanding the sentiment.In addition,the factors that cause the differences in translating sentiments are mainly subjective factors,such as translators’ personal language usage habits,translation purposes,and translation principles,and objective factors,such as the intervention of editors of publishing houses,and different perspectives to understand the original text brought about by the gender of the translator.This study makes an attempt to combine the method of sentiment analysis with the paradigm of corpus translation studies to investigate the conveyance of sentiments in translation,which will help to broaden the domain of translation studies as well as studies on sentiments.It can also raise the translators’ awareness of the sentiments in literary works so that they try to achieve the equivalence of sentiment in their translations. |