Phraseology in Corpus-Based Translation Studies: A Stylistic Study of Two Contemporary Chinese Translations of Cervantes's Don Quijote

The present work sets out to investigate the stylistic profiles of two modern Chinese versions of Cervantes’s Don Quijote (I): by Yang Jiang (1978), the first direct translation from Castilian to Chinese, and by Liu Jingsheng (1995), which is one of the most commercially successful versions of the Castilian literary classic. This thesis focuses on a detailed linguistic analysis carried out with the help of the latest textual analytical tools, natural language processing applications and statistical packages. The type of linguistic phenomenon singled out for study is four-character expressions (FCEXs), which are a very typical category of Chinese phraseology. The work opens with the creation of a descriptive framework for the annotation of linguistic data extracted from the parallel corpus of Don Quijote. Subsequently, the classified and extracted data are put through several statistical tests. The results of these tests prove to be very revealing regarding the different use of FCEXs in the two Chinese translations. The computational modelling of the linguistic data would seem to indicate that among other findings, while Liu’s use of archaic idioms has followed the general patterns of the original and also of Yang’s work in the first half of Don Quijote I, noticeable variations begin to emerge in the second half of Liu’s more recent version. Such an idiosyncratic use of archaisms by Liu, which may be defined as style shifting or style variation, is then analyzed in quantitative terms through the application of the proposed context-motivated theory (CMT). The results of applying the CMT-derived statistical models show that the detected stylistic variation may well point to the internal consistency of the translator in rendering the second half of Part I of the novel, which reflects his freer, more creative and experimental style of translation. Through the introduction and testing of quantitative research methods adapted from corpus linguistics and textual statistics, this thesis has made a major contribution to methodological innovation in the study of style within the context of corpus-based translation studies