Orpus-Aided Quality Assessment Of Graduates’ Translation: From A CLC Perspective

This paper explores the effect of part of speech, construction’s load capacity andcollocation on translations’ readability. Language features of English-majored graduates’translations are taken as subject in this paper for two reasons. On the one hand, the authorbelieves that it is not the elementary errors as much as the usage of basic expressions thataffects a translation’s readability. On the other, English-majored graduates’ translations wererarely studied before. The work presented in this paper is an attempt to contrast languagefeatures of professional translations with those of poor ones. By illustrating both global andlocal features of graduates’ translations, the author finally sums up the attributors affectingChinese translation language.Based mainly on a quantitative study of self-built corpora, with the Lancaster Corpus ofMandarin Chinese as a reference, the author also uses qualitative method to characterizegraduates’ translations. By employing empirical linguistic evidences, both grammatically andsemantically, the author demonstrates a macro and micro analysis of translations. Aftercalculation, observation and analysis, the author comes to the following conclusions.First of all, there exist differences between professional translations and poor translationsin the distribution of part of speech, particularly in content words. In terms of content words,the author mainly discusses verbs, adverbs, nouns, and adjectives, from which good translationsand bad translations are obviously distinguished. What is the point is that these four contentwords mirror the dynamic tendency of translational language (Chinese as target language); andtherefore, the author concludes that the more dynamic a translation is, the better the translationquality is.Secondly, the complexity of the embedded part in V-PN structure is related to the finalscore of a translation. The author gives a classification of the embedded part and illustratesthem respectively. The most common structure is simple parallel construction, which isfrequently found in translations of low grade. By contrast, good translations usually make itmore dynamically instead of V-PN structure. The other two structures discussed in this part areconstructions embedded with phrases and clauses. The conclusion is that the more complex theembedded part is, the less readable a translation is.Last but not least, there is a negative correlation between ill collocation and the score of atranslation. Both grammatical and semantic functions are considered when analyzing whether acollocation works appropriately or not. According to the calculation and analysis, collocationdetermines the final score of a translation. A translation of higher grade has a smaller number ofill collocations. In short, the major findings are that the above three attributors all affect translationquality. This paper characterizes translation languages, and receives tentative conclusions. Myultimate purpose is to show that, beyond part of speech, both construction’s load capacity andcollocation can be incorporated into the framework of translation quality assessment. This paper,which offers a reference to translation teaching, also suffices to say that it is necessary tooutline the above three attributors into translation quality assessment.