An interlanguage study of English discourse connectors in argumentative essays by Thai university students

Abstract

The study examined the use of English discourse connectors (DCs) in three main aspects: Orthography, Syntax, and Semantics and Pragmatics including the frequency of use and errors of DC lexis in argumentative essays written by native speakers of English (NSs), and non-native speakers of English (NNSs). For the NSs, 20 essays out of 43 essays from English native speaker undergraduate corpus, LOCNESS, was employed in this study. For the NNSs, the data were drawn from the essays written by 300 students from various universities in and around Bangkok. The 40 essays of the NNSs were selected: 20 from the top high English exposure (NNSHs), and 20 from the bottom low English exposure (NNSLs) by using English Language Exposure Questionnaire scores. Altogether, there were 60 argumentative essays randomly specified and selected for this study. For the DCs analysis, the frameworks proposed by Halliday & Hasan (1976), Biber et al (1999), and Cowan (2008) were adopted and employed. The aims of the study were (1) to describe the use of English DCs of NSs, NNSHs, and NNSLs, (2) to compare and contrast the DCs used in argumentative essays among NSs, NNSHs, and NNSLs, and identify the problems of the DCs used in the two NNS groups, and (3) to analyze the patterns and problems of DC usage in argumentative essays between NNSHs and NNSLs. The clarification was based on interlanguage study. Both descriptive statistic, and inferential statistic were used to describe the data and to test whether the differences found among the sample groups were significant or not. The following findings were found, (1) the frequency of the use of English discourse connectors among the three sample groups was significant difference in the Causal and Temporal types between the NSs and both of the NNSs., (2) in terms of the Orthography, the use of DCs were different between the NSs and the NNSs due to negative L1 transfer, overgeneralization, and insufficient knowledge in punctuation usage, (3) in terms of Syntax, the NSs and the NNSHs showed the similarity in the use of all the three sentence types and the sentential positions, whereas the NNSLs showed significant differences in the use of all three sentence types, and the sentence initial position. The differences in the use of DCs in the NSs and both groups of the NNSs could be the effect of all the five factors caused by interlanguage development i.e. Language Transfer, Transfer of training, Strategies of second language communication, Strategies of second language learning and Overgeneralized, and (4) in terms of Semantics and Pragmatics, It was found that out of the 62 DCs lexis with a total of 865 tokens that were used, the 2 DCs lexis “and” with 22 tokens, and “finally” with 3 tokens exhibited their multi pragmatic functions, i.e., there were not a one-to-one relationship between their semantic functions and the pragmatic uses. “and” was found used not only as “additive” but also as “adversative” and “causal”. “finally” in the “ordering” semantic category was found used as “ordering” as well as “causal” and “summation”. The NSs and the NNSHs had similar patterns of the use of “and” and “finally”. Conversely, the NNSLs showed the differences in the use of DCs in this aspect. Based on the findings of the study, recommendations for further research and pedagogical implications are given in order to develop the way in which discourse connectors should be investigated and taught.