Elamur Research
☰
home / papers / 2018 / Conversations Gone Awry: Detecting Early Signs of Conversational Failure

Conversations Gone Awry: Detecting Early Signs of Conversational Failure

Written for the Research library (claude-opus-5-5) from the paper text. Original — aclanthology.org/P18-1125.pdf

Summary

Most work on online toxicity flags harmful comments only after they appear. This paper instead asks whether the opening of a seemingly civil conversation can reveal that it will later turn into a personal attack, early enough for intervention. The authors rebuilt about 50 million conversations from English Wikipedia talk page histories. They found candidate derailments with a toxicity classifier and confirmed them through crowdsourcing, producing a labeled set of conversations that either stay civil or end in an attack. Using ideas from causal inference, they paired derailing and healthy conversations from similar topical contexts, so that subject matter alone could not explain the outcome. From the first comment-reply exchange they extracted pragmatic features: politeness strategies plus rhetorical prompt types discovered without supervision. These features carry real predictive signal. Hedged openings tend to stay civil, while blunt questioning and direct second-person address signal risk. Humans reach 72% on the same task, so it is possible but far from easy. The work introduces the task, a dataset and baselines for research on early-warning moderation.

Key points

  • Source corpus: about 50 million conversations across 16 million English Wikipedia talk pages. Roughly 1% of comments there are estimated to be antisocial, which is why classifier-based pre-filtering was needed.
  • Crowdsourced vetting included a job that asked 367 annotators to judge whether the final comment of 4,022 conversations was a personal attack.
  • Baselines on the paired, balanced task:
    • Interlocutor features (edit counts, anonymity) score at chance.
    • The Perspective API toxicity score reaches 60.5%, rising to 64.9% when combined with the pragmatic features.
    • Volunteer humans reach 72% on 100 pairs.
  • The pragmatic-feature classifier is correct on 80% of the pairs that humans also got right. When the first attack comes after the fourth comment (282 pairs), it still reaches 67.4%, while the sentiment and bag-of-words baselines fall to chance.
  • Data and code are released in the Cornell Conversational Analysis Toolkit (ConvoKit).

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: Linguistics & languages · Natural language processing

Cited by

75 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 73 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/p18-1125

@inproceedings{zhang2018conversations,
  title = {{Conversations Gone Awry: Detecting Early Signs of Conversational Failure}},
  author = {Zhang, Justine and Chang, Jonathan and Danescu-Niculescu-Mizil, Cristian and Dixon, Lucas and Hua, Yiqing and Taraborelli, Dario and Thain, Nithum},
  booktitle = {Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2018},
  doi = {10.18653/v1/p18-1125},
  url = {https://doi.org/10.18653/v1/p18-1125}
}

References

64 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Yavuz Akbulut, Yusuf Levent Sahin, and Bahadir Eristi. 2010. Cyberbullying victimization among Turkish online social utility members. Journal of Educational Technology & Society. find
  2. Kelsey Allen, Giuseppe Carenini, and Raymond T Ng. 2014. Detecting disagreement in conversations using pseudo-monologic rhetorical structure. In Proceedings of EMNLP. find
  3. Herbert H Clark and Dale H Schunk. 1980. Polite responses to polite requests. Cognition. find
  4. Benjamin Collier and Julia Bear. 2012. Conflict, criticism, or confidence: An empirical examination of the gender gap in Wikipedia contributions. In Proceedings of CSCW. find
  5. Lewis A Coser. 1956. The Functions of Social Conflict. Routledge. find
  6. Tim Althoff, Cristian Danescu-Niculescu-Mizil, and Dan Jurafsky. 2014. How to ask for a favor: A case study on the success of altruistic requests. In Proceedings of ICWSM. find
  7. Cristian Danescu-Niculescu-Mizil, Lillian Lee, Bo Pang, and Jon Kleinberg. 2012. Echoes of power: Language effects and power differences in social interaction. In Proceedings of WWW. find
  8. Ofer Arazy, Lisa Yeo, and Oded Nov. 2013. Stay on the Wikipedia task: When task-related disagreements slip into personal and procedural conflicts. Journal of the Association for Information Science and Technology. find
  9. Cristian Danescu-Niculescu-Mizil, Moritz Sudhof, Dan Jurafsky, Jure Leskovec, and Christopher Potts. 2013. A computational approach to politeness with application to social factors. In Proceedings of ACL. find our card
  10. Malika Aubakirova and Mohit Bansal. 2016. Interpreting neural networks to improve politeness comprehension. In Proceedings of EMNLP. find
  11. Thomas Davidson, Dana Warmsley, Michael Macy, and Ingmar Weber. 2017. Automated hate speech detection and the problem of offensive language. In Proceedings of ICWSM. find our card
  12. Lars Backstrom, Jon Kleinberg, Lillian Lee, and Cristian Danescu-Niculescu-Mizil. 2013. Characterizing and curating conversation threads: Expansion, focus, volume, re-entry. In Proceedings of WSDM. find
  13. Carsten K De Dreu and Laurie R Weingart. 2003. Task versus relationship conflict, team performance, and team member satisfaction: A meta-analysis. Journal of Applied Psychology. find
  14. Penelope Brown and Stephen Levinson. 1987. Politeness: Some universals in language usage. Cambridge University Press. find our card
  15. Bruce Fraser. 1980. Conversational mitigation. Journal of Pragmatics. find
  16. Moira Burke and Robert Kraut. 2008. Mind your Ps and Qs: The impact of politeness and rudeness in online communities. In Proceedings of CSCW. find
  17. Liye Fu, Lillian Lee, and Cristian Danescu-Niculescu-Mizil. 2017. When confidence and competence collide: Effects on online decision-making discussions. In Proceedings of WWW. find
  18. Eshwar Chandrasekharan, Umashanthi Pavalanathan, Anirudh Srinivasan, Adam Glynn, Jacob Eisenstein, and Eric Gilbert. 2017. You can’t stay here: The efficacy of Reddit’s 2015 ban examined through hate speech. In Proceedings of CSCW. find
  19. Björn Gambäck and Utpal Kumar Sikdar. 2017. Using convolutional neural networks to classify hate-speech. In Proceedings of the Workshop on Abusive Language Online. find
  20. Despoina Chatzakou, Nicolas Kourtellis, Jeremy Blackburn, Emiliano De Cristofaro, Gianluca Stringhini, and Athena Vakali. 2017. Measuring #GamerGate: A tale of hate, sexism, and bullying. In Proceedings of WWW. find
  21. Justin Cheng, Michael Bernstein, Cristian Danescu-Niculescu-Mizil, and Jure Leskovec. 2017. Anyone can become a troll: Causes of trolling behavior in online discussions. In Proceedings of CSCW. find
  22. Justin Cheng, Cristian Danescu-Niculescu-Mizil, and Jure Leskovec. 2015. Antisocial behavior in online discussion communities. In Proceedings of ICWSM. find
  23. Ali Gheitasy, José Abdelnour-Nocera, and Bonnie Nardi. 2015. Socio-technical gaps in online collaborative consumption (OCC): An example of the Etsy community. In Proceedings of ICDC. find
  24. Erving Goffman. 1955. On face-work: An analysis of ritual elements in social interaction. Psychiatry. find
  25. Christophe Henner and Maria Sefidari. 2016. Wikimedia Foundation Board on healthy Wikimedia community culture, inclusivity, and safe spaces. Wikimedia Blog. find
  26. Elizabeth F Churchill and Sara Bly. 2000. Culture vultures: Considering culture and communication in virtual environments. SIGGroup Bulletin. find
  27. Pamela J Hinds and Mark Mortensen. 2005. Understanding conflict in geographically distributed teams: The moderating effects of shared identity, shared context, and spontaneous communication. Organization Science. find
  28. Herbert H Clark. 1979. Responding to indirect speech acts. Cognitive psychology. find
  29. Axel Hübler. 1983. Understatements and Hedges in English. John Benjamins Publishing. find
  30. Joseph M Kayany. 1998. Contexts of uninhibited online behavior: Flaming in social newsgroups on usenet. Journal of the Association for Information Science and Technology. find
  31. John Pavlopoulos, Prodromos Malakasiotis, and Ion Androutsopoulos. 2017b. Deeper attention to abusive user content moderation. In Proceedings of EMNLP. find
  32. Aniket Kittur, Ed H Chi, and Bongwon Suh. 2009. What’s in Wikipedia?: Mapping topics and conflict using socially annotated category structure. In Proceedings of CHI. find
  33. Alan Ritter, Colin Cherry, and Bill Dolan. 2010. Unsupervised modeling of Twitter conversations. In Proceedings of NAACL. find
  34. Aniket Kittur and Robert E Kraut. 2008. Harnessing the wisdom of crowds in Wikipedia: Quality through coordination. In Proceedings of CSCW. find
  35. Aniket Kittur, Bongwon Suh, Bryan A Pendleton, and Ed H Chi. 2007. He says, she says: Conflict and coordination in Wikipedia. In Proceedings of CHI. find
  36. Vinodh Krishnan and Jacob Eisenstein. 2015. “You’re Mr. Lebowski, I’m the Dude”: Inducing address term formality in signed social networks. In Proceedings of NAACL. find
  37. Ravi Kumar, Mohammad Mahdian, and Mary McGlohon. 2010. Dynamics of conversations. In Proceedings of KDD. find
  38. Haewoon Kwak, Jeremy Blackburn, and Seungyeop Han. 2015. Exploring cyberbullying and other toxic behavior in team competition online games. In Proceedings of CHI. find
  39. Robin T Lakoff. 1973. The logic of politeness: Minding your P’s and Q’s. In Proceedings of the Chicago Linguistic Society. find
  40. Bing Liu, Minqing Hu, and Junsheng Cheng. 2005. Opinion observer: Analyzing and comparing opinions on the web. In Proceedings of WWW. find
  41. Vlad Niculae and Cristian Danescu-Niculescu-Mizil. 2016. Conversational markers of constructive discussions. In Proceedings of NAACL. find
  42. Vlad Niculae, Srijan Kumar, Jordan Boyd-Graber, and Cristian Danescu-Niculescu-Mizil. 2015. Linguistic harbingers of betrayal: A case study on an online strategy game. In Proceedings of ACL. find
  43. Chikashi Nobata, Joel Tetreault, Achint Thomas, Yashar Mehdad, and Yi Chang. 2016. Abusive language detection in online user content. In Proceedings of WWW. find
  44. Marco Ortu, Bram Adams, Giuseppe Destefanis, Parastou Tourani, Michele Marchesi, and Roberto Tonelli. 2015. Are bullies more productive? Empirical study of affectiveness vs. issue fixing time. In Proceedings of MSR. find
  45. John Pavlopoulos, Prodromos Malakasiotis, and Ion Androutsopoulos. 2017a. Deep learning for user comment moderation. In Proceedings of the Workshop on Abusive Language Online. find
  46. Paul R Rosenbaum. 2010. Design of observational studies. Springer. find
  47. Donald B Rubin. 2007. The design versus the analysis of observational studies for causal effects: Parallels with the design of randomized trials. Statistics in Medicine. find
  48. Tamara Shepherd, Alison Harvey, Tim Jordan, Sam Srauy, and Kate Miltner. 2015. Histories of hating. Social Media + Society. find
  49. Vivek K Singh, Marie L Radford, Qianjia Huang, and Susan Furrer. 2017. “They basically like destroyed the school one day”: On newer app features and cyberbullying in schools. In Proceedings of CSCW. find
  50. Sara Owsley Sood, Elizabeth F Churchill, and Judd Antin. 2012. Automatic identification of personal insults on social news sites. Journal of the American Society for Information Science and Technology. find
  51. Chenhao Tan, Vlad Niculae, Cristian Danescu-Niculescu-Mizil, and Lillian Lee. 2016. Winning arguments: Interaction dynamics and persuasion strategies in good-faith online discussions. In Proceedings of WWW. find
  52. Yla R Tausczik and James W Pennebaker. 2010. The psychological meaning of words: LIWC and computerized text analysis methods. Journal of Language and Social Psychology. find
  53. Kal Turnbull. 2018. “Thats Bullshit” – Rude Enough for Removal? A Multi-Mod Perspective. Change My View Blog. find
  54. Jessica Vitak, Kalyani Chadha, Linda Steiner, and Zahra Ashktorab. 2017. Identifying women’s experiences with and strategies for mitigating negative effects of online harassment. In Proceedings of CSCW. find
  55. Lu Wang and Claire Cardie. 2014. A piece of my mind: A sentiment analysis approach for online dispute detection. In Proceedings of ACL. find
  56. William Warner and Julia Hirschberg. 2012. Detecting hate speech on the World Wide Web. In Proceedings of the Workshop on Language in Social Media. find
  57. Wikimedia Support and Safety Team. 2015. Harassment survey. Wikimedia Foundation. find
  58. Ellery Wulczyn, Nithum Thain, and Lucas Dixon. 2016. Wikipedia talk labels: Toxicity. find
  59. Ellery Wulczyn, Nithum Thain, and Lucas Dixon. 2017. Ex machina: Personal attacks seen at scale. In Proceedings of WWW. find
  60. Naomi Yamashita and Toru Ishida. 2006. Automatic prediction of misconceptions in multilingual computer-mediated communication. In Proceedings of IUI. find
  61. Dawei Yin, Zhenzhen Xue, Liangjie Hong, Brian D Davison, April Kontostathis, and Lynne Edwards. 2009. Detection of harassment on Web 2.0. In Proceedings of the Workshop on Content Analysis in the Web 2.0. find
  62. Amy X Zhang, Bryan Culbertson, and Praveen Paritosh. 2017a. Characterizing online discussion using coarse discourse sequences. In Proceedings of ICWSM. find
  63. Justine Zhang, Ravi Kumar, Sujith Ravi, and Cristian Danescu-Niculescu-Mizil. 2016. Conversational flow in Oxford-style debates. In Proceedings of NAACL. find
  64. Justine Zhang, Arthur Spirling, and Cristian Danescu-Niculescu-Mizil. 2017b. Asking too much? The rhetorical role of questions in political discourse. In Proceedings of EMNLP. find