Run-on Sentence Errors In Chinese EFL Learners' Compositions

In the field of Second Language Acquisition, Error Analysis Hypothesis was a hypothesis that used to be prevailing in the 1970s. Enormous researches based on Error Analysis Hypothesis at that time did make great contribution to explain the complex behavior in learning a second language. However, EA itself had its own weaknesses. A new branch of linguistics emerged in the late twenties century --- Corpus Linguistics provides a brand new perspective into the interlanguage researches. Some inevitable weaknesses in traditional error analysis are overcome in the computer-aided error analysis based on learner's corpora. By making use of comprehensive and authoritative corpora and retrieval software tools, the frequencies of errors, lists of sorts of errors and contexts of errors could be obtained accurately and quickly.The paper aims to investigate one sort of syntactic error --- run-on sentence errors produced by Chinese EFL learners in their compositions and hopes to bring about some implications in teaching English writing. The study makes use of an error-tagged learner corpus, CLEC, which is the only error-tagged corpus found so far in China. By calculating the proportion of run-on sentence errors in overall syntactic errors, the paper reveals the severity of such errors. Later on, a careful examination of run-on sentence errors in one sub-corpus reveals that most run-on sentence errors are composed of two or more independent clauses, which are of the following three semantic relations: cause-effect relation; coordination and exposition. Then run-on sentence errors are analyzed from the aspect of cohesive means. The result shows that although run-on sentences are normally caused by the omission of connectives, there are other cohesive means used in them, for instance the use of anaphora and the repetition of the same lexical unit. Those cohesive means used are found to be quite similar to the cohesive means used in a type of sentence in Chinese, running sentence (流水句). What's more, the structural features of run-on sentences and running sentences are fairly similar to each other. These findings reveal that one of the causes of run-on sentences is the influence of learner's mother tongue, i.e. Chinese. The paratactic feature of Chinese and the hypotactic feature of English are compared briefly in the paper. In consequence, another cause of run-on sentence errors is attributed to the influence of different thinking modes. Finally, the paper introduces some corresponding suggestions in language teaching according to the main findings, in the hope that they could help to improve the capability of learners to construct grammatical and logical sentences.