I am pleased to nominate WONG Ka Shun Jeff for the Capstone Exemplary Award. His project offers an insightful analysis of how Hong Kong’s exam-oriented English education continues to shape university students’ long-term writing habits. Drawing on 90 essays and follow-up interviews, Jeff demonstrates that many students rely heavily on a narrow set of formulaic expressions, revealing limited awareness of the importance of lexical variety. He traces this pattern to an assessment culture that prioritizes mechanical accuracy and fixed structures, encouraging students to view writing as a procedural task rather than a flexible, meaning-making practice. A notable strength of this project is its timely examination of students’ divergent uses of AI in academic writing. Jeff identifies a spectrum ranging from full outsourcing to strategic, controlled enhancement, offering important implications for how educators should guide AI-supported writing. What makes the project especially commendable is Jeff’s thoughtful pedagogical vision. He proposes concrete, evidence-based recommendations for EAP instruction, emphasizing corpus-informed and discipline-specific approaches that foster rhetorical awareness and linguistic sophistication beyond formulaic memorization. His study also successfully connects what he has learned in our undergraduate programme with his future career as an English teacher, especially in light of the rapid emergence of AI-assisted writing.

Formulaic language use among Hong Kong university students reflects a broader pedagogical problem rooted in the Diploma of Secondary Education (DSE) examination system. Through mixed-methods research combining corpus analysis of 90 student texts and case study interviews with two contrasting participants, this study examines how exam-driven teaching shapes students’ long-term writing habits. Drawing on Biber et al.’s (2004) lexical bundle framework and recent research on formulaic competence (Paquot, 2010; Durrant, 2017), the analysis focuses on three categories of expressions: exemplification markers, stance markers, and discourse organizers. The corpus reveals striking uniformity: 531 occurrences of just three exemplification phrases appear across 90 texts, suggesting deeply ingrained patterns rather than conscious linguistic choice. Case studies of Marcus (formula-dependent) and Tom (variety-focused) demonstrate that while metacognitive awareness can enable students to transcend formulaic habits, assessment systems often fail to reward this development. Both students maintain similar GPAs despite markedly different levels of linguistic sophistication. The study further investigates how artificial intelligence disrupts traditional assumptions about formulaic instruction, revealing two contrasting approaches: complete outsourcing versus strategic enhancement. The findings call for discipline-specific, AI-literate pedagogical reform that emphasizes metalinguistic awareness over formula memorization. This research contributes empirical evidence of how DSE pedagogy shapes tertiary writing and offers practical implications for EAP curriculum design in Hong Kong and similar examination-oriented contexts.

Over the past four years, and especially through this capstone, I learned how to move from “doing assignments” to actually asking my own research questions and finding answers I care about. Working on formulaic language forced me to connect theory, methods, and my own experience as a DSE student and tutor. I struggled at first with messy data and feeling overwhelmed by corpus outputs, but learning to combine quantitative patterns with interview stories showed me what independent research really involves. More broadly, my undergraduate studies taught me to read critically, write more consciously, and question easy explanations about language and education in Hong Kong. The project also changed how I see teaching: I now think less about quick exam tricks and more about helping students develop awareness and flexibility that will matter beyond any single test.