What's Happening?
A bibliometric review of cyberbullying research from 2010 to 2025 reveals a significant increase in publications, with 218 records (63.2%) published after 2020. The study, which analyzed 345 English-language records from the Web of Science Core Collection,
found strong and well-established psychosocial and computational research strands. However, explicitly pragmatic constructs, such as implicature, stance, and speech acts, remain less consolidated in high-frequency mapped structures. The research indicates a disciplinary dispersion, with publications appearing in various fields including psychology, communication, media studies, and computer science. While foundational psychological explanations like online disinhibition and moral disengagement are prominent in cited references, their concepts are not consistently translated into lexical feature design or model evaluation in computational studies.
Why It's Important?
This research highlights a critical gap in understanding cyberbullying, particularly in the U.S. context where online interactions are pervasive. The uneven development means that while there's extensive work on the psychological impacts and computational detection of cyberbullying, the nuanced linguistic and pragmatic aspects—how language is used in context to cause harm—are less explored. This gap can hinder the development of more effective detection tools and intervention strategies, as the subtle ways cyberbullying manifests through language might be overlooked. For policymakers, educators, and technology companies in the U.S., this suggests a need for interdisciplinary approaches that integrate linguistic and pragmatic insights with existing psychological and computational frameworks to better address the complexities of online harassment.
What's Next?
The review proposes a conceptual 'language-to-harm' framework to bridge the identified gap, linking platform affordances, linguistic cues, pragmatic interpretation, psychological appraisal, behavioral and social responses, and cumulative harm. Future studies are encouraged to test these relationships using context-rich longitudinal and interactional data, pragmatic and psychological annotation, multilingual and cross-platform validation, and fairness-aware human-AI evaluation. This suggests a future direction for U.S. research and development in cyberbullying, emphasizing a more holistic understanding that incorporates the subtleties of language use. Collaboration between linguists, psychologists, and computer scientists will be crucial to developing more sophisticated detection and prevention mechanisms.
Beyond the Headlines
The findings underscore a broader challenge in addressing complex social phenomena in the digital age: the difficulty of integrating diverse disciplinary perspectives. Cyberbullying, as a media-psychological phenomenon, requires a multi-faceted approach that goes beyond isolated studies of its effects or technical detection. The limited integration of pragmatic constructs points to a potential blind spot in how online harm is conceptualized and combated. This has ethical implications for AI development, as detection algorithms might miss context-dependent forms of harassment if they are not informed by a deep understanding of human language and social interaction. Ultimately, a more integrated approach could lead to more empathetic and effective digital environments, fostering healthier online communities.
