A Corpus-Based Contrastive Study Of Translational Chinese And Translational English Academic Prose

This thesis is a corpus-based study of the linguistic features of translational English and Translational Chinese academic prose. Based on four comparable corpora of English and Chinese translation and non-translation, the thesis attempts to find out the distinctive features of translational Chinese in contrast to non-translational Chinese, the distinctive features of translational English in contrast to non-translational English, and, by comparing these translational features, to testify the hypothetical universals of translation in the register of academic prose.As one of the most important conceptualizations in contemporary Translation Studies, Translation Universals started and has been extensively studied in Corpus-based Translation Studies since 1990s. Baker (1993:243) defined Translation Universals as features "which typically occur in translated texts rather than original utterances and which are not the result of interference from specific linguistic systems". That is, Translation Universals are inherent and typical common properties of all translated texts, existing in different languages, different genres and registers. These features of translation often manifest as features of Simplification, Explicitation, Normalization and Levelling out, etc. Most previous studies focus on general features of one translational language or two translational languages; few have focused on universals in specific registers. However, it is vital to validate Translation Universals hypotheses in specific genres or registers. Therefore, this thesis tries to examine the Translation Universals hypotheses in academic prose of translational languages through empirical analyses so as to testify or falsify those hypotheses. The corpora used in this study are as follows:Corpus of Translational English. Freiburg-LOB Corpus of British English, ZJU Corpus of Translational Chinese and Lancaster Corpus of Mandarin Chinese.With regard to specific linguistic features of translational languages, this study at first examines the Simplification Hypothesis by observing lexical features such as the proportion of lexical words, the proportion of function words and the ratio of lexical words over function words. In addition, this study examines the Explicitation Hypothesis by observing features such as average sentence length, average sentence segment length and specific grammatical words, etc. Furthermore, the Normalization hypothesis is examined by observing features of "Bei" constructions in Chinese academic prose of LCMC and ZCTC, and passives in English academic prose of COTE and FLOB. The study finds that common features do exist in translational Chinese and translational English academic prose:Firstly, both the proportion of lexical words and the ratio of lexical words over function words are lower in translational Chinese and translational English academic prose than those in non-translational academic prose; and the differences are statistically significant. So it shows the lower information load in translated academic prose in both languages, which supports the Simplification hypothesis. Secondly, given that Chinese is a "parataxis" language, the author compares the average sentence segment length, and analysis shows longer average sentence segment length of translated Chinese academic texts than native Chinese ones. Meanwhile, since English is a "hypotaxis" language, the author compares the average sentence, and the finding is longer average sentence length of translated English academic texts than native Chinese ones. Consequently, it can be seen that sentences are longer in both translated Chinese and translated English academic texts than in native ones, which supports the Explicitation hypothesis. Thirdly, there is no significant difference between the frequency of "Bei" constructions in translated Chinese academic texts and that in native ones, nor is there significant difference between the frequency of "By"-passives in translated English academic texts and that in native ones. Also agent-less passives are less frequently used in translational English academic texts than in native ones. Therefore, the Normalization hypothesis is supported neither in translational English nor in translational Chinese academic prose. Finally, in analyses of specific words in two translational languages, the thesis also finds that:First, particles are less frequently used in translated English academic texts and overall translational English corpora and the differences are significant. Second, modals and the word "to" (regardless of preposition or infinitive marker) are less frequently used in translated academic texts of COTE and overall COTE. And the differences are statistically significant.The empirical analyses above reveal that cross-lingual features of Translation Universals do exist in translational Chinese and translational English academic prose. These universals indicate that translation may have the same or similar cognitive process regardless of the source or target languages that translation involves. These common features of both translational Chinese and translational English in academic prose provide new evidence for the existence of Translation Universals in distant languages such as Chinese and English.