Collocation Deficiency in a Learner Corpus of English : From an Overuse Perspective

Collocation Deficiency in a Learner Corpus of English: from an overuse perspective Rebecca Hsue-Hueh Shih Department of Foreign Languages and Literature National Sun Yat-sen University, Kaohsiung, Taiwan, R.O.C. E-mail: hsuehueh@mail.nsysu.edu.tw Abstract Collocational deficiency is a pervasive phenomenon in learner English. Language learners often fail to choose the correct combination of two or more words due to their unawareness of collocational properties in vocabulary. They are apt to adopt lexical simplification strategies such as using a synonymous or Ll-influenced expression. This paper presents a corpus-based study on the collocational deficiency of Taiwanese learners of English. The work utilizes two pre-tagged corpora, Taiwanese Learner Corpus of English and British National Corpus, to examine the learner's use of collocations over a set of synonymous words: big, large, great. The experimental findings indicate that among the three words the collocations with big are significantly overused by the learners when it is used to refer to abstract concepts. This overuse phenomenon is further investigated and it is found that the collocations of high frequency in the learner English tend to be used to express vague ideas when more specific meanings should be conveyed. It is also found that the learners are apt to apply those collocations to the cases where more concise expressions are preferred. Another finding shows that problematic collocations, pertaining to big, large and great, are produced as the result of learner's application of the Ll-transfer and synonym strategies, which the Taiwanese learners commonly adopt for lexical simplification. 1. Introduction Corpus Linguistics studies language features based on large databases of authentic language samples stored on computer. Because its automated quantitative analysis provides novel and refreshing insights into real language use [1], the corpus-based approach has rapidly spread into many language-related research. SLA (Second Language Acquisition) and EFL (English as Foreign Language) specialists, with no exception, consider large databases of learner English a useful resource for them to gain concrete evidence and a wider perspective on learners' inter-language acquired during the process of language learning. Therefore, computer learner corpora (CLCs) with various mother tongue backgrounds have been subsequently constructed [2]. With gradual availability of CLC and the awareness of its potential, a wide range of CLC- research starts to boom. The research often involves comparisons between inter-language that learners possess and native language on various linguistic features. For instance, the frequency distributions of most commonly-used words in a native and seven eastern European learner corpora are compared on different parts-of-speech categories [3]; the use of complement clauses in four learner corpora as contrasted with their native counterparts [4] is studied; the use of adverbial connectors by Swedish learners in comparison with the natives' is examined [5]. This kind of cross-language approach helps SLA and EFL specialists find out what linguistic features the language learners are apt to overuse/underuse, what are the particular areas of language behavior shared by learners from different backgrounds, and to what extent these phenomena appear in learner English. The quantitative information as such guides the researchers to carry out insightful …

Content truncated to 500 words