Discourse analysis of cyberbullying among university students: An applied linguistics approach

Abstract

Cyberbullying in university contexts is often investigated as a behavioural, psychological, or institutional safety phenomenon. A few studies have comparatively conceptualized it as a discursive practice as power, identity, and group belonging are examined in complicated interactions. Adopting applied-linguistics as an approach, this paper advances an integrated framework for analyzing cyberbullying as discursive harm—a patterned configuration of stance acts, evaluative positioning, and face-attacking moves that scale across turns, participants, and semiotic modes. Theoretically, we synthesize Computer-Mediated Discourse Analysis (Herring, 2004), Critical Discourse Analysis (CDA) grounded in Fairclough’s three-dimensional logic (as operationalized by Janks, 2005), Appraisal theory (Martin & White, 2005), and facework/(im)politeness (Goffman, 1967; Brown & Levinson, 1987; Culpeper, 1996) to propose a novel model: The Stance–Face–Power (SFP) Model. Methodologically, we introduce Corpus-Assisted, Interactionally-Refined CDA (CAIR-CDA), which couples corpus-assisted discourse studies procedures (pattern discovery at scale) with sequential, turn-by-turn interactional analysis and multimodal annotation, enabling analysts to connect micro-linguistic and semiotic choices (e.g., pronouns, modality, negation, mock politeness, evaluative lexis, evidential framing, emoji/meme uptake) to macro-social dynamics (e.g., gendered norm enforcement, institutional authority, group dominance, and platform affordances). A distinctive methodological contribution is the Dialogic Harm Index (DHI)—a transparent, reproducible metric integrating evaluative polarity/intensity (Appraisal), impoliteness strategy type, alignment and piling-on dynamics, and interactional power cues (commands, gatekeeping, exposure threats). To demonstrate analytic payoffs without reproducing harmful content or revealing identities, we provide a set of ethically “sanitized” composite vignettes and show how CAIR-CDA distinguishes cyberbullying from adjacent practices (e.g., conflict, critique, and joking aggression) by tracking escalation, role shifts, and repair refusal across discourse trajectories. We conclude with implications for campus policy, reporting ecosystems, and discourse-informed digital citizenship pedagogy, and we outline an agenda for multilingual, cross-cultural university settings and ethical governance of sensitive online data.