Elamur Research⌕
☰
home / papers / 2018 / Before Name-Calling: Dynamics and Triggers of Ad Hominem Fallacies in Web Argumentation

Before Name-Calling: Dynamics and Triggers of Ad Hominem Fallacies in Web Argumentation

Paper · 2018 · Ivan Habernal, Henning Wachsmuth, Iryna Gurevych et al.

Summary

Ad hominem attacks are among the best-known fallacies, yet empirical evidence about them is scarce. Theoretical treatments usually rely on a handful of invented examples. This paper studies personal attacks in real dialogical argumentation at three levels of context: isolated comments, direct replies to an original post, and longer multi-party exchanges. The authors worked with Reddit to scrape the full Change My View (CMV) forum, including comments that moderators deleted. They treated the moderators' "rude or hostile" rule violations as ad hominem labels and checked those labels through crowdsourced annotation studies. They then trained neural classifiers to address three questions: how much context recognition needs, whether controversy or reasonableness relates to attacks, and whether a thread is heading toward an attack. To make the models interpretable, they re-implemented a self-attentive network (SSAE-NN) and read its attention weights to identify linguistic triggers. Attacks proved rare in this moderated forum and often end the discussion. The work offers the first NLP study of multi-faceted ad hominem in genuine debates, and the data and code are publicly released.

Key points

  • Dataset: The complete CMV dump holds roughly 2M posts spread over 32k submissions, which together make up 780k distinct threads. It contains 3,866 moderator-flagged ad hominem comments, only about 0.2% of all posts.
  • Thread dynamics: In 48.6% of single-attack threads, the attack is the final comment. Only 15% of attacks are answered with another attack, and about 66% of attackers strike without any prior involvement in the thread.
  • Label verification: Amazon Mechanical Turk annotation used MACE aggregation with two groups of six workers and a 0.9 threshold, reaching 0.79 Cohen's κ.
  • Controversial topics: Submissions with the most attacks dealt with religion, gender/sexuality, U.S. politics (mostly Trump), U.S. racism and veganism.
  • Identified triggers: Attention analysis of SSAE-NN surfaced cues such as vulgar intensifiers, direct imperatives, accusations of fallacies or of not reading arguments, demands for evidence, sarcasm, uppercase text, mentions of trolling, and loaded words like "Nazi."

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

 Institutions

🎓Technische Universität DarmstadtDarmstadt · Germany76 works · 35,008 citations🎓Bauhaus-Universität WeimarWeimar · Germany9 works · 3,111 citations🎓Paderborn UniversityPaderborn · Germany9 works · 3,554 citations

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 9,086 worksArtificial Intelligence · 8,889 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 worksPhilosophy · 211 works

Cite

Cite as Ivan Habernal, Henning Wachsmuth, Iryna Gurevych and Benno Stein. 2018. Before Name-Calling: Dynamics and Triggers of Ad Hominem Fallacies in Web Argumentation. Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/n18-1036

@inproceedings{habernal2018before,
  title = {{Before Name-Calling: Dynamics and Triggers of Ad Hominem Fallacies in Web Argumentation}},
  author = {Ivan Habernal, Henning Wachsmuth, Iryna Gurevych and Benno Stein},
  booktitle = {Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers)},
  year = {2018},
  doi = {10.18653/v1/n18-1036},
  url = {https://doi.org/10.18653/v1/n18-1036}
}

References

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

  1. Aristotle and George Kennedy (translator). 1991. On Rhetoric: A Theory of Civil Discourse. Oxford University Press, USA. find
  2. False negatives were caused basically by presence of many ‘informative’ content words (980 unemployment, quarterly publication, inflation data, 474 actual publications, this experiment, biological ailments, medical doctorate, 1214 graduate degree, education, health insurance) and misinterpreted sarc… find
  3. Lora Aroyo and Chris Welty. 2015. Truth is a lie: Crowd truth and the seven myths of human annotation. AI Magazine 36(1):15–24. https://doi.org/10.1609/aimag.v36i1.2564. ↗
  4. Leila Arras, Grégoire Montavon, Klaus-Robert Müller, and Wojciech Samek. 2017. Explaining Recurrent Neural Network Predictions in Sentiment Analysis. In Proceedings of the 8th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis. Association for Computational L… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/habernal-2018-before-name-calling-dynamics-triggers/