Elamur Research⌕
☰
home / papers / 2026 / Modeling Linguistic Violence: An Ontology-Based Framework for the Computational Analysis of Violence Manifested in Language

Modeling Linguistic Violence: An Ontology-Based Framework for the Computational Analysis of Violence Manifested in Language

Paper · 2026 · Brenda Salenave Santana, Ana Marilza Pernas and Aline A. Vanin

Summary

Terms such as hate speech, toxic speech and dangerous speech are often treated as synonyms. This blurs annotation guidelines and makes automated moderation inconsistent. Detectors that rely on keywords also miss harm that is only implied. The authors first review the literature on these terms, along with verbal aggression and cyberbullying. They then build a unified ontology that separates the concepts by their defining attribute: target, intent, effect, act or pattern. Each concept is linked to recurring rhetorical hallmarks of violent discourse.

To put the ontology into practice, they design a three-stage NLP pipeline. A first stage triages texts. Semantic Role Labeling and Named Entity Recognition then break each utterance into its target, action and hallmarks. Finally, rules map these components onto ontology categories. Overlapping labels are allowed, and a text that matches no rule is labelled benign.

The proposal is demonstrated qualitatively on four Portuguese examples and has not yet been evaluated quantitatively. Its main value is a conceptual basis for annotation. Because it relies on generic semantic tasks with available multilingual models rather than large task-specific labelled datasets, it suits low-resource settings such as Portuguese.

Key points

  • The ontology distinguishes five phenomena by one primary axis each:
    • Dangerous Speech by intent (inciting violence)
    • Toxic Speech by effect (degrading the conversation)
    • Hate Speech by target (a protected group)
    • Verbal Aggression by act (a personal insult)
    • Cyberbullying by pattern (repeated harm)
  • Hallmarks are grouped into five macro sets, drawing on Leader Maynard & Benesch and de Barros:
    • Portraying people as less than human
    • Mirror accusations, which charge opponents with one's own intentions
    • Violations of social contracts concerning integrity or purity
    • Casting doubt on in-group loyalty
    • Oppositional figures
  • The pipeline has three stages: Initial Triage, then Semantic Role & Entity Deconstruction (SRL + NER), then Ontological Mapping. Classification is framed as structured, multi-label prediction.
  • The case study draws on four Portuguese datasets: the 2024 Dangerous Speech Election Dataset, ToLD-Br, HateBR2.0 and ToxSyn-PT. In one example, a sentence praising women athletes is correctly kept benign, whereas a keyword filter might flag it.
  • Subword multilingual encoders (mBERT, XLM-RoBERTa) are proposed to cope with noisy social-media text. Planned future work includes validating the ontology through corpus annotation and building explainable classifiers.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 2The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 works

🧭 Topics 1Research subjects and languages the work is about.Knowledge graphs & ontologies · 192 works

Cite

Cite as Brenda Salenave Santana, Ana Marilza Pernas and Aline A. Vanin. 2026. Modeling Linguistic Violence: An Ontology-Based Framework for the Computational Analysis of Violence Manifested in Language. Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1.

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{santana2026modeling,
  title = {{Modeling Linguistic Violence: An Ontology-Based Framework for the Computational Analysis of Violence Manifested in Language}},
  author = {Brenda Salenave Santana, Ana Marilza Pernas and Aline A. Vanin},
  booktitle = {Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1},
  year = {2026},
  url = {https://aclanthology.org/2026.propor-1.29/}
}

References

36 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Robert Arp, Barry Smith, and Andrew D Spear. 2015. Building ontologies with basic formal ontology. Mit Press. find
  2. Fabienne Baider. 2020. Pragmatics lost?: Overview, synthesis and proposition in defining online hate speech. Pragmatics and Society, 11(2):196–218. find
  3. Koray Balci and Albert Ali Salah. 2015. Automatic analysis and identification of verbal aggression and abusive behaviors for online social games. Comput. Hum. Behav., 53(C):517–526. find
  4. Michael Randall Barnes. 2025. Complex harms in online speech. Conversations Online: Explorations in Philosophy of Language, page 239. find
  5. Susan Benesch. 2023. Dangerous speech. In Christian Strippel, Sünje Paasch-Colberg, Martin Emmer, and Joachim Trebbe, editors, Challenges and perspectives of hate speech research, volume 12 of Digital Communication Research, pages 185–197. Berlin. find
  6. Iago Alves Brito, Julia Soares Dollis, Fernanda Bufon Färber, Diogo Fernandes Costa Silva, and Arlindo Rodrigues Galvão Filho. 2025. Toxsyn-pt: A large-scale synthetic dataset for hate speech detection in portuguese. Preprint, arXiv:2506.10245. — [arXiv:2506.10245](https://arxiv.org/abs/2506.10245) find
  7. Judith Butler. 2021. Discurso de ódio: Uma política do performativo. Editora Unesp, São Paulo. find
  8. Diana Luz Pessoa de Barros. 2014. O discurso intoler-ante na internet: enunciação e interação. find
  9. Robert Allan Dickinson, Cathy Buerger, Tom Cowin, Niko Shahbazian, Alexandra Filindra, Stacey Hunt, and Elizabeth Young. 2025. 2024 Dangerous Speech Election Dataset. find
  10. Mai ElSherief, Vivek Kulkarni, Dana Nguyen, William Yang Wang, and Elizabeth Belding. 2018. Hate lingo: A target-based linguistic analysis of hate speech in social media. Preprint, arXiv:1804.04257. — [arXiv:1804.04257](https://arxiv.org/abs/1804.04257) find
  11. Paula Fortuna and Sérgio Nunes. 2018. A survey on automatic detection of hate speech in text. ACM Computing Surveys (CSUR), 51(4):1–30. find
  12. Guoxiu He, Zhe Gao, Zhuoren Jiang, Yangyang Kang, Changlong Sun, Xiaozhong Liu, and Wei Lu. 2020. Think beyond the word: Understanding the implied textual meaning by digesting context, local, and noise. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Info… find
  13. Marlon Hurt. 2021. Pledging to Harm: A linguistic analysis of violent intent in threatening language. Ph.D. thesis, Aston University. find
  14. Md Saroar Jahan and Mourad Oussalah. 2021. A systematic review of hate speech automatic detection using natural language processing. arXiv preprint. find
  15. K Krishna Jayanth, G Bharathi Mohan, R Prasanna Kumar, and M Rithani. 2024. Intent recognition leveraging xlm-roberta for effective nlu. In 2024 3rd International Conference on Applied Artificial Intelligence and Computing (ICAAIC), pages 877– 882. find
  16. Sakshi Joshi, Anindita Mukherjee, Usha A. Jogalekar, Renuka Agrawal, Abhishek Anand, and Santhosh Phanitalpak Gandhala. 2025. Enhancing toxic comment classification with multi-label capabilities: Leveraging bert and roberta models. In 2025 International Conference on Emerging Trends in Industry 4.0… find
  17. Jonathan Leader Maynard and Susan Benesch. 2016. Dangerous speech and dangerous ideology: An integrated model for monitoring and prevention. Genocide Studies and Prevention. find
  18. Alyssa Lees, Vinh Q. Tran, Yi Tay, Jeffrey Sorensen, Jai Gupta, Donald Metzler, and Lucy Vasserman. 2022. A new generation of perspective api: Efficient multilingual character-level transformers. arXiv preprint. find
  19. João Augusto Leite, Diego Silva, Kalina Bontcheva, and Carolina Scarton. 2020. Toxic language detection in social media for Brazilian Portuguese: New dataset and multilingual analysis. In Proceedings of the 1st Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics a… find
  20. Barbara Lewandowska-Tomaszczyk, Anna Baczkowska, Olga Dontcheva-Navratilova, Chaya Liebeskind, Giedre Valunaite Oleskeviciene, Slavko Zitnik, Marcin Trojszczak, Renate Povolna, and Linas Selmistraitis. 2023a. Llod schema for simplified offensive language taxonomy in multilingual detection and applic… find
  21. Barbara Lewandowska-Tomaszczyk, Anna B aczkowska, Chaya Liebeskind, Giedre Valunaite Oleskeviciene, and Slavko Žitnik. 2023b. An integrated explicit and implicit offensive language taxonomy. Lodz Papers in Pragmatics, 19(1):7–48. find
  22. Barbara Lewandowska-Tomaszczyk, Slavko Zitnik, Anna Baczkowska, Chaya Liebeskind, Jelena Mitrović, and Giedre Valunaite Oleskeviciene. 2021. Lod-connected offensive language ontology and tagset enrichment. In SALLD-1: Workshop on Sentiment Analysis & Linguistic Linked Data, pages 135–150. CEUR Work… find
  23. Wenji Li, Anggeng Li, Tianqi Tang, Yue Wang, and Zejian Fang. 2022. Multilingual toxic text classification model based on deep learning. In 2022 3rd International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering (ICBAIE), pages 726–729. find
  24. Junjie Liu, Yong Yang, Xiaochao Fan, Ge Ren, Liang Yang, and Qian Ning. 2022. Offensive-language detection on multi-semantic fusion based on data augmentation. Applied System Innovation, 5(1). find
  25. Swapnil Sanjaykumar Mane, Suman Kundu, and Rajesh Sharma. 2025. A survey on online aggression: Content detection and behavioral analysis on social media. ACM Comput. Surv., 57(7). find
  26. Mary Kathryn McGowan. 2019. Just words: on speech and hidden harm. Oxford University Press. find
  27. Paul Röttger, Bertie Vidgen, Dirk Hovy, and Janet B. Pierrehumbert. 2021. Two contrasting data annotation paradigms for subjective NLP tasks. CoRR, abs/2112.07475.
  28. Semiu Salawu, Yulan He, and Joanna Lumsden. 2020. Approaches to automated detection of cyberbullying: A survey. IEEE Transactions on Affective Computing, 11(1):3–24. find
  29. Brenda S. Santana and Larissa A. de Freitas. 2024. Pln em redes sociais. In H. M. Caseli and M. G. V. Nunes, editors, Processamento de Linguagem Natural: Conceitos, Técnicas e Aplicações em Português, 3 edition, book chapter 34. BPLN. find
  30. Peter K Smith, Jess Mahdavi, Manuel Carvalho, Sonja Fisher, Shanette Russell, and Neil Tippett. 2008. Cyberbullying: Its nature and impact in secondary school pupils. Journal of child psychology and psychiatry, 49(4):376–385. find
  31. Guizhe Song, Degen Huang, and Zhifeng Xiao. 2021. A study of multilingual toxic text detection approaches under imbalanced sample distribution. Information, 12(5). find
  32. Marco A. Stranisci, Simona Frenda, Mirko Lai, Oscar Araque, Alessandra T. Cignarella, Valerio Basile, Viviana Patti, and Cristina Bosco. 2022. O-dang! the ontology of dangerous speech messages. Preprint, arXiv:2207.10652. find
  33. Lynne Tirrell. 2017. Toxic speech: Toward an epidemiology of discursive harm. philosophical topics, 45(2):139–162. find
  34. Douglas Trajano, Rafael H. Bordini, and Renata Vieira. 2024. Olid-br: offensive language identification dataset for brazilian portuguese. Language Resources and Evaluation, 58(4):1263–1289. find
  35. Francielle Vargas, Isabelle Carvalho, Thiago A. S. Pardo, and Fabrício Benevenuto. 2025. Context-aware and expert data resources for brazilian portuguese hate speech detection. Natural Language Processing, 31(2):435–456. find
  36. Mengfan Yao, Charalampos Chelmis, and Daphney-Stavroula Zois. 2019. Cyberbullying ends here: Towards robust detection of cyberbullying in social media. In The World Wide Web Conference, pages 3427– 3433. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/santana-2026-modeling-linguistic-violence-ontology-based/