A Corpus-Based Investigation of Linguistic Patterns in Academic Writing by Non-Native Speakers in Pakistan

Abstract

This study adopts a corpus-based mixed-methods approach to investigate the linguistic patterns in academic writing produced by non-native English-speaking students. Using purposive sampling, authentic academic texts specifically essays and short research papers written by postgraduate students from two Pakistani universities were collected to compile a clean, anonymized corpus of approximately 100,000-150,000 words. The texts were converted into plain-text format, checked for consistency, and organized into sub-corpora based on academic level and genre. Quantitative analysis using corpus tools such as AntConc was employed to examine frequency patterns, lexical diversity, collocational behaviour, grammatical structures, cohesive devices, and recurrent phrase formations. Complementing this, qualitative concordance analysis was conducted to interpret how these linguistic patterns reflect broader academic writing practices among non-native English speakers. The findings reveal notable deviations from native-speaker academic norms, including restricted lexical variation, overuse of general vocabulary, dependence on formulaic expressions, and challenges in producing complex syntactic constructions. These patterns highlight the influence of first-language transfer, academic socialization, and proficiency levels. The study contributes to applied linguistics and English for Academic Purposes (EAP) by offering corpus-informed insights into the linguistic difficulties faced by NNES writers and suggesting pedagogical implications for enhancing academic writing instruction.