A Corpus-based Study Of Noun Errors In Non-english Majors' Writings

This corpus-based study was conducted to investigate the noun errors in the writings by Chinese non-English majors. Based on EA (Error Analysis), interlanguage, and L1 transfer, this study aims to address three research questions:1. What are the categories of noun errors in non-English majors'writings?2. What is the distribution of noun error categories in non-English majors'writings?3. What is the relationship of the noun errors to the quality of non-English majors'writings?All the data in the present study were collected from ST3, a sub-corpus of CLEC (Chinese Learner English Corpus). Two control variables were taken into consideration when the data were selected: topic and length. In terms of the topic, 127 pieces with the same title Health Gains in Developing Countries were selected as subjects. In terms of length, all the data chosen from corpus must be at least over 100 words so as not to affect the writing scores. Data analysis consists of four steps: de-tagging all the errors except noun errors in the selected writings semi-automatically, counting errors with the help of MyFinder, categorizing noun errors according to Yang, Gui & Yang's categorization (2005) in CELC and calculating correlation coefficient between the noun errors and the quality of writings with the help of SPSS 11.0.The major findings from this study are as follows:Firstly, four categories of noun errors are found in students'writings, which include errors in word form, errors in noun phrases, errors in wording, and errors in collocation. These errors may be caused by four factors: incomplete mastery of the vocabulary, L1 negative transfer, overgeneralization, and the inconsistency of English words pronunciation and their spelling. This finding is generally consistent with the previous related results respectively.Secondly, in terms of distribution of noun errors, errors in word form have the highest proportion in noun errors accounting for 61.5%, followed by errors in noun phrase16%, wording 15% and collocation 7.4%. Three main reasons account for the frequency of errors. One reason is called topic-dependency. Another reason may be that the numbers of error reasons may be proportionate to the number of errors, that is the more error reasons, the more errors. The other reason may be the avoidance strategy learners employ consciously both in writing and in speaking when they face a communicative difficulty, which results in the least errors in collocation.Thirdly, there is almost no correlation between noun errors and the writing quality. This result indicates that noun errors in the compositions almost do not affect writing quality, which is consistent with Liu's (2008) finding. There are two reasons for this phenomenon. One reason is that teachers tend to employ holistic scoring method to assess the composition in the terms of its content, organization, and language. They might pay more attention to macro errors in content, organization and language, which may lead to neglecting micro errors, like noun errors. Another reason is that though they are negatively correlated with the quality of writings (Llach 2005), lexical errors are further subdivided into noun errors, verb errors, adjective errors, adverbial errors, preposition errors, article errors and so on. Noun errors are only one kind of lexical errors, which might not affect the quality of learners'writings globally.These results offer some pedagogical implications. First, in teaching English nouns, Chinese EFL teachers should pay attention to their pronunciations, forms and meanings, especially to the inconsistency between their pronunciations and their spellings. Second, Chinese EFL teachers should provide their students with adequate noun-focused exercises both in and after class to raise their awareness of learning nouns. Third, Chinese EFL teachers can employ both target language corpus and Chinese learners'corpus to facilitate vocabulary teaching and learning.The present study has three limitations. Firstly, the categorization of noun errors in this thesis is not entirely scientific, though it is based on Yang, Gui and Yang's classification of tagged errors in the CELC. However, due to her limited knowledge and academic research ability, the author could not make it any better. Secondly, the number of samples collected in this study is relatively small, because the author planned to conduct a comparative study of noun errors between the CET-4 writings and CET-6 writings at first. Then, due to a great deal of work, limited time and the author's ability, the author chose this topic from ST3 instead of from ST3 and ST4. Thirdly, subjectivity cannot be completely avoided in explaining some reasons for noun errors due to corpus-based study. If the author could interview subjects about the exact reasons for a certain type of error, explanation of these errors will be more convincing and objective.