Huang Xin Rui Joey should be nominated for an exemplary capstone due to the project’s strong analytical focus and solid methodological execution. Rather than concluding that English language policing on the LIHKG forum merely reproduces class and educational hierarchies, Joey identifies a “recursive discursive system” where users reflexively critique their own community’s policing ideologies. By synthesizing Agha’s concept of enregisterment with Bucholtz and Hall’s framework of authentication, the project presents a well-reasoned argument that the community actively enregisters evaluative testing procedures rather than just linguistic forms. Methodologically, Joey manually compiled, translated, and coded a purposive database of 240 unique threads and 8,572 posts spanning nine years of the platform’s history. Because LIHKG restricts automated web crawling and operates in a volatile digital environment where posts are frequently deleted, this intensive manual compilation provides a useful and preserved archive of enacted ideological discourse for future sociolinguistic research.

This study argues that English policing on LIHKG, Hong Kong’s largest anonymous forum, is not a series of isolated evaluative acts but a recursive discursive system that supplies the materials for its own continuation, drawing on 240 threads and 8,572 posts. English in Hong Kong carries no settled social meaning, the same feature can index authentic local identity in one thread and mark educational failure in another, making it a uniquely unstable site for the negotiation of class, belonging, and legitimacy. Using verbal hygiene, the indexical field, and enregisterment as analytical frameworks, it documents the discursive strategies through which users police English, the social meanings those acts activate, and the reflexive meta-policing through which users critique the practice itself. Findings exceed what standard accounts predict. Explicit criticism dominates over ironic or memetic forms. Policing operates through authenticity testing rather than correctness evaluation: users judge whether a speaker’s English fits a claimed identity, not merely whether it is standard BrE or AmE, grammatically correct or accurate. Reflexive meta-policing is not marginal but structural: the L1-E x L2-F cell, meta-policing crossed with meta-ideology, is the largest in the dataset, suggesting that critiquing policing is among the community’s most common moves. English on LIHKG thus functions not only as an object of judgment but as a medium through which a community negotiates its contested relationship to class, authenticity, and linguistic legitimacy, and in doing so renders that relationship available for perpetual renegotiation.

The hardest part of building this argument was learning not to settle for the version that sounded finished. Every time the theoretical claim felt sophisticated enough to stop, there was a reason to push further. What this project has taught me is that a strong argument is not one that sounds sophisticated. It is one that has survived sustained pressure from its own doubts.

Research on language policing offers a specific temptation, that is, treating the data as evidence of a phenomenon rather than as a record of people. The data records Hong Kongers’ narratives of shame, exclusion, and self-surveillance. When English is turned into a test of legitimacy, the cost is more than theoretical. It is lived by people whose accents, fluency, or projected personae become available for judgment. The analysis does not lose sight of who bears its weight.

Finally, I want to thank Wil for his support and advice. I also want to acknowledge my own effort in bringing this project to completion. This project was completed alongside everything else — another thesis and other exhaustions that do not belong in an academic submission but were present throughout. The work got done not because the conditions were right but because I refused to let them be the reason it didn’t. I showed up when it would have been easier not to.