Lexical Collocation Errors in Chinese-speaking EFL Learners'' Spoken and Written Production

• Main Authors: Ya-Pin Huang, 黃雅萍
• Other Authors: Li-szu Huang
• Format: Others
• Language: en_US
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/53471458535458458863

碩士 === 國立高雄第一科技大學 === 應用英語研究所 === 100 === In view of the importance of collocational knowledge in learners’ use of a second language, this study analyzed Chinese EFL learners’ production of lexical collocations in their speech and writing. Three research questions were addressed: (a) What categories of lexical collocations do Chinese EFL learners use in their spoken and written production? (b) What is the error rate of each lexical collocation category in learners’ spoken versus written data? (c) Based on the identified lexical collocation errors, what collocates (verbs, adjectives, nouns and adverbs) tend to cause difficulties in these learners’ spoken and written production? The two sources of data came from a course instructor’s self-compiled learner corpora at a university of science and technology in Taiwan. The spoken data were retrieved from 26 English-major juniors’monologic speech in their Advanced Listening and Conversation course. The written data consisted of another 26 students’ academic essays completed in their Advanced English Writing course. First, the researcher examined each essay and each transcript of learners’ speech to locate the six categories of lexical collocations (LC1: verb+noun; LC2: adjective+noun; LC3: noun+verb; LC4: noun1+of+noun2; LC5: adverb+adjective; LC6: verb+adverb). Then, the researcher consulted the Oxford collocations dictionary for students of English and Corpus of Contemporary American English to identify target collocations to be analyzed. Each target collocation found in the online corpus had to reach an MI (mutual information) score of 3. Any target collocation which could not be found in the above two references was marked as an error. After identifying the target collocations and errors, the researcher calculated the type-token ratio of each LC category to know its collocational variety. Then, error rates were obtained by dividing the total number of errors in each category of collocations by the total number of collocations in that category. To extract difficult collocates in each category, the researcher first corrected each erroneous collocation and then classified the corrected verb, adjective, adverb, and noun collocates into different semantic domains, functions, or relations. The classified type with the most error tokens under each LC category was regarded as most challenging for learners. The results revealed LC1 (verb+noun) and LC2 (adjective+noun) as most commonly produced in both corpora, but their collocational types as measured in the type-token ratio were least varied. By contrast, LC4 (noun1+of+noun2) was least produced in both corpora but most varied in writing whereas LC6 was most varied in speaking. Regarding the error rates, in both corpora, LC4 (noun1+of+noun2) had the highest error rate whereas LC5 (adverb+adjective) the lowest. The results of difficult collocates showed that in LC1 (verb+noun), activity verb collocates were most difficult in both language modes. For LC2 (adjective+noun), the function of determining adjectives was most challenging. For LC3 (noun+verb), the greatest difficulty occurred in activity verb collocates in speaking but causative verb collocates in writing. Regarding LC4 (noun1+of+noun2), most error tokens were found in the misuse of prepositions and compound nouns. Also challenging was the relation of reification. Finally, the difficulty in LC6 (verb+adverb) in the written data resulted from degree adverb collocates. The results of this study contribute to learners’ awareness of their collocational patterns and problematic categories of lexical collocations in their spoken and written production. Language teachers can then target the semantic domains of difficult collocates to improve learners’ production of lexical collocations.