Corpus-Based Research On Relative Clause Distribution

As a significant part of natural language, relative clause has long been the research emphasis of linguists. In recent years, research about processing difficulty of relative clause is broadly extended and the subject preference in English is widely accepted. However, the production of relative clause is much more complex and numerous factors influence this process. The present study introduces two self-built corpora and researches about the distribution characteristics of relative clause and then discusses about some of the factors that influence relative clause production.This paper collected47pieces of writing centered on the same theme "green growth", among which25were reports and22were journal papers. Two corpora were built separately. Through using Stanford Parser and some other research method,1000sentences were randomly selected from each corpus and the relative clauses among them were coded. The distributions of these coded relative clauses were studied.Research results show that all the relative clause types mentioned in literature review are found in the two self-built corpora, including SS, SO, OS, OO, SU, DO, IO, OPREP, GEN, OCOMP, RRC and NRRC. The distribution frequency order of these relative clauses is generally in line with the processing difficulty order. Through comparisons from different angles, this paper argues that processing difficulty, animacy, genre and theme will to some extent influence the distribution and production of relative clause. This corpus-based study will offer evidence for researches about relative clause distribution and provide some references for learning and proper use of relative-clause.