A corpus-driven comparative analysis of verb phrase based bundles in L2 Malaysian polytechnic learners’ writing / Radika Subramaniam

• Main Author: Radika , Subramaniam
• Format: Thesis
• Published: 2023
• Subjects: P Philology. Linguistics
• Online Access: http://studentsrepo.um.edu.my/15286/2/Radika.pdf; http://studentsrepo.um.edu.my/15286/1/Radika_Subramaniam.pdf; http://studentsrepo.um.edu.my/15286/

This study aims to explore the similarities and differences in the use of verb phrase-based (VP-based) bundles in the writing of L2 Malaysian Engineering learners via the Malaysian Polytechnic Electronic Engineering Learner Corpus (MyPolyEELC) in comparison to their native counterparts in the British Academic Written English (BAWE) sub-corpus and experts in the Electrical and Electronic Engineering Research Articles Corpus (3ERAC). The MyPolyEELC which consists of 891,919 words is developed from 231 Final Year Project reports written by Electronic Engineering Diploma learners from 14 Malaysian polytechnics. The 374,383-word BAWE sub-corpus is compiled from the existing BAWE corpus by selecting texts written by L1 English learners from the Physical Science disciplinary group while the 3ERAC is made up of 100 research articles from the Electrical and Electronic Engineering discipline with 710,456 words. This is a mixed-method study which involves quantitative (i.e., frequency-based and log-likelihood test of significance) and qualitative approaches (i.e., text analysis) for data analysis. The structural classifications of VP-based bundles are performed using Biber et al.’s (1999) framework, while the functional categories are identified based on Hyland (2008a). Francis et al.’s (1996) verb pattern analysis is employed to postulate the lexico-grammatical patterns of shared bundle clusters across the three corpora, along with the identification of semantic preference which is exercised using Partington (2004) as well as Cortes and Hardy (2013). The analysis reveals a significant use (types and tokens) of VP-based bundles in the L2 learner writing compared to both the L1 English learner and expert writings. These clause-based structures feature more conversation-like registers contrary to academic prose in the L2 learner text. Less frequent use of ‘it be V-ed that’ and ‘it V adj’ lexico-grammatical patterns is associated with underused ‘anticipatory it’ structural category in the L2 learner corpus which accordingly affects the production of bundles signalling ‘observation’ and ‘inference’. These findings are also related to the limited use of participant-oriented functions in non-native learner writing compared to the native learner and expert writing which illustrate a less writer/reader-oriented writing style. Overall, the analysis has uncovered the gap to be addressed in academic writing, especially the Final Year Project report of Malaysian Polytechnic learners that is, to make appropriate use of lexical chunks (in this case, VP-based bundles) in delivering stance-related expressions to display writer involvement and reader engagement, and text-oriented functions to demonstrate a coherent and systematic written academic discourse as per the task requirement. The interrelation between the structural category, lexico-grammatical pattern and semantic preference of lexical bundle(s) illustrates that all these elements are important in motivating the functional category of the bundle(s) based on a larger context of its occurrence. In light of the variations discovered in the use of VP-based bundles in the non-native learner writing in comparison to the native novice and expert writing, corpus-based pedagogical practices or data-driven learning approach can be considered in EAP classrooms to be exposed to L2 learners to model more proficient writing.