Acorpus-based Comparative Study Of Semantic Prosody Between Chinese English Major Learners And Native Speakers

Semantic prosody has attracted worldwide attention since corpus linguistics is introduced to the language research. The research of semantic prosody is of great significance in that lack of the semantic prosodic features of a lexical item would directly results in inappropriate collocation behaviors, which are not native and natural.From the contrastive studies already conducted at home, it can be seen that most of them are based on Chinese non-English major learners, especially on the overuse and underuse of the semantic prosody of the node word. The learner corpora is the written part of SWECCL (Spoken and Written English Corpus of Chinese Learners), and the reference corpora is a sub-corpora of BNC. And this research is supposed to compare the semantic prosodic features between the Chinese English major learners and the native speakers, while using the node word cause, commit, career and provide; and further explore the similarities and differences of collocation patterns of them. The research firstly investigates the distribution of semantic prosody; and then through specifying the minimum frequency and MI score, the significant collocates are computed to analyze the collocation behavior. The data shows that English-major learners demonstrate similar semantic prosody as the native speakers on the use of the four node words; but they differ in the deep-level collocation patterns, to be exact, typical collocation, unusual collocation and interlanguage collocation. The possible underlying reason can be explained from the perspective of negative transfer of mother tongue, learner environment, learning strategies and thought pattern and so on.This research sheds some light on the foreign language teaching and learning. Firstly, through comparing Chinese English-major and non-English major learners, it can be found that the former has acquired the semantic prosodic features much better, which means the quality and quantity of the input of the language materials are required to be improved; secondly, in order to heighten the quality of interlanguage, much more attention should be paid on the chunks and collocations of the lexical item; Last but not the least, data-driven learning method should be widely applied in the language learning, which is considered a good way to promote the self-learning ability of the learners.