1.3.2 Data analysis
Both qualitative and quantitative analyses are adopted in the current project,both of which are mutually exclusive on the surface,but are complementary in essence.Quantitative analyses rely on reliable qualitative studies so as to identify the research domain.Qualitative analyses need enough data to support the research findings.The more instances of one phenomena are found,the more persuasive are the explanations of the potential forms.Charles(2009:152)admits that corpus techniques have an important role to play in the analysis,as they provide quantitative data for identifying items which are potential markers of the interpersonal discourse features studied.
The analysis of CMs will be retrieved by the software AntConc 3.2.2 and at the same time interpreted by manual analysis.There are several reasons to combine manual analysis and software analysis together.First,the grammatical forms of some CMs are homonyms.For example,the meaning of the conjunctive adverb however is totally different from that in the clause However(No matter how)you try,you can never get that.If the analysis of the CMs completely depend on the software,the results of the data analysis will not be exact enough.Second,the relationship between the semantic framework and the grammatical realizations are not one-to-one.For one thing,various conjunctive expressions could express the same semantic meaning;for another,one CM may have various semantic meanings such as and.And could express a tremendous variety of semantic relations and one relation would have a very large set of synonyms for and.According to Deng(2012),and could convey various semantic relations such as location,simultaneity,source,succession,cause,conditionality and conclusion.
Hyland(2009:110)claims that corpora and discourse approaches are perfect bedfellows.They are complementary approaches which can inform and enrich each other,thereby leading to more insightful analyses of language use.Corpus studies require a focus on action to balance the focus on language.Therefore,this study uses corpus methods to identify frequent CMs and their associated patterns,while discourse analysis is employed to shed light on the analysis of the functions of CMs.