A Corpus-Based Study Of Keyword Use In Chinese College English Majors’ Timed Compositions

The present study aims to explore the keyword use in Chinese English learners’ writing, based on a sub-corpus of WECCL2.0(Wen et al.,2008) and the reference corpus LOCNESS (Granger et al.,2006), especially to find out college English majors’writing ability from this respect and the relationship between the keyword use and the quality of L2writing. The data in the research of the same topic consist of25,983tokens in103timed compositions in WECCL2.0and the reference corpus, LOCNESS, contains a number of288,177tokens.The major findings of this research are as follows:(1) The number of keywords in Chinese L2learners’ argumentative writings is similar to that in native writers’ writings when both compared with LOCNESS. Therefore, the average keyword or keyword unit number usually fluctuates from three to five, which can already express the theme in an acceptable way for argumentative essays of about300to500words in this sense. Thus, it reveals that the keywords are adequately used by Chinese L2learners. However, the results indicate that generally, native writers tend to employ six to nine keywords that can highlight the theme or be more exclusive to the topic more, and that they have a larger vocabulary that reflects their flexible use of keywords when expressing their ideas compared with Chinese L2learners.(2) The types in L2theme-given compositions are fewer than those in L2free compositions with similar writers’competence, composition length and tokens. While contrasted with compositions written by native writers, though with similar composition numbers, total tokens and types of L2writings are far fewer. From lexical statistics, it indicates the restriction for theme-given compositions leads to a limited range of word choice and a higher repetition. Native writers’writings provide more information than L2writing. The recurrence rate of top500tokens in L2theme-given compositions are higher than that of free ones, i.e. the lexical application of theme-given compositions are mainly distributed in top500tokens, and this rate approaches closer in top1000, which implies keywords of a given theme mainly distribute in top500tokens.(3) The choice and distribution of keywords are determined by the theme of the compositions. Moreover, the words in the theme-given compositions are greatly confined by the given topic:the more specific the theme is, the higher recurrence rate of the keywords will be. Some words indirectly correlated to the theme (e.g. pronouns, auxiliary verbs) actually betray the preference of Chinese L2learners out of cultural background, and writer’s personal style of words.(4) According to the SPSS statistics, there is no significant relationship between individual keyword number and its corresponding L2writing quality. However, there is indeed a kind of link existing among the keyword number, the quality of relevance and the quality of being focused, which closely influences composition quality of theme-prominence. Usually, there are many aspects mentioned in the composition with few keywords, however, lacking of the topic repetition or theme-prominence. While writings with six or even more keywords are related to several aspects in the passages, and at the same time, few prominent points can be grasped by the readers, so from this respect, they quite resemble the ones with one or even zero keywords.The significance of this research is to study Chinese college English majors’ compositions through keyword analysis by WordSmith Tools. It is hoped to help Chinese L2learners have a better grasp of argumentative writings in the field of theme-prominence.