Elamur Research
☰
home / papers / 2022 / Investigating Reasons for Disagreement in Natural Language Inference

Investigating Reasons for Disagreement in Natural Language Inference

Written for the Research library (claude-opus-5-5) from the paper text. Original — aclanthology.org/2022.tacl-1.78.pdf

Summary

NLI benchmarks usually assume that each premise–hypothesis pair has one correct label, chosen by majority vote. In practice, annotators often disagree in systematic ways. Jiang and de Marneffe ask where this disagreement comes from and how models should handle it. They built a taxonomy of 10 disagreement sources, grouped into 3 classes following Aroyo and Welty's "Triangle of Reference": uncertainty in sentence meaning, underspecified guidelines, and annotator behavior. They then hand-annotated part of MNLI with these categories, drawing on ChaosNLI's larger label counts. The analysis shows that disagreement items vary widely, so the NLI label distribution cannot be read the same way for every item. The authors then compared two categorical alternatives to predicting distributions: a 4-way classifier with an extra "Complicated" class, and a multilabel classifier that can output several NLI labels. The multilabel model proved more expressive and interpretable, and it recalled more of the plausible readings. This matters because downstream uses such as fact-checking need categorical decisions that still signal genuine ambiguity.

Key points

  • Taxonomy categories by class:
    • Sentence meaning: Lexical, Implicature, Presupposition, Probabilistic Enrichment, Imperfection.
    • Guidelines: Coreference, Temporal Reference, Interrogative Hypothesis.
    • Annotator behavior: Accommodating Minimally Added Content, High Overlap.
  • Data: a subset of MNLI, including items that have no majority label, which are normally dropped from evaluation. ChaosNLI label counts were used alongside it, and a label counted as present when at least 20 annotators chose it.
  • Probabilistic Enrichment: the multilabel model gave more than one label to 68% of these items. Of these items, 36% were predicted as EN (Entailment + Neutral) and 27% as NC (Neutral + Contradiction).
  • Minimally Added Content: 44% of these items were predicted as EN, and 76% received at least the Neutral label. The 4-way model tended to output "Complicated" for Coreference items whose sentences share no structure.
  • Release: annotations, guidelines and code are available at github.com/njjiang/NLI_disagreement_taxonomy.

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

72 papers of our library cite it (of the 140673 whose bibliographies we read); 4 of them have a page here, 68 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.1162/tacl_a_00523

@article{jiang2022investigating,
  title = {{Investigating Reasons for Disagreement in Natural Language Inference}},
  author = {Jiang, Nan-Jiang and Marneffe, Marie-Catherine de},
  journal = {Transactions of the Association for Computational Linguistics},
  year = {2022},
  doi = {10.1162/tacl\_a\_00523},
  url = {https://doi.org/10.1162/tacl_a_00523}
}

References

52 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Lora Aroyo and Chris Welty. 2015. Truth is a lie: Crowd truth and the seven myths of human annotation. AI Magazine, 36(1):15. https:// doi.org/10.1609/aimag.v36i1.2564 ↗ our card
  2. Samuel R. Bowman, Gabor Angeli, Christopher Potts, and Christopher D. Manning. 2015. A large annotated corpus for learning natural language inference. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 632–642, Lisbon, Portugal. Association for Computational Linguistics. https://doi.org/10 .18653/v1/D15-1075 ↗
  3. Aida Mostafazadeh Davani, Mark Dı́az, and Vinodkumar Prabhakaran. 2022. Dealing with disagreements: Looking beyond the majority vote in subjective annotations. Transactions of the Association for Computational Linguistics, 10:92–110. https://doi.org/10.1162/tacl_a_00449 ↗ our card
  4. Ondřej Dušek and Zdeněk Kasner. 2020. Evaluating semantic accuracy of data-to-text generation with natural language inference. In Proceedings of the 13th International Conference on Natural Language Generation, pages 131–137, Dublin, Ireland. Association for Computational Linguistics. find
  5. Nouha Dziri, Ehsan Kamalloo, Kory Mathewson, and Osmar Zaiane. 2019. Evaluating coherence in dialogue systems using entailment. In Proceedings of the 2019 Workshop on Widening NLP, pages 146–148, Florence, Italy. Association for Computational Linguistics. https://doi.org/10.18653/v1/N19 -1381 ↗
  6. Katrin Erk and Diana McCarthy. 2009. Graded word sense assignment. In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, pages 440–449, Singapore. Association for Computational Linguistics. find
  7. Elisa Ferracane, Greg Durrett, Junyi Jessy Li, and Katrin Erk. 2021. Did they answer? Subjective acts and intents in conversational discourse. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1626–1644, Online. Association for Computational Linguistics. https://doi.org/10.18653/v1/2021 .naacl-main.129 ↗
  8. Tommaso Fornaciari, Alexandra Uma, Silviu Paun, Barbara Plank, Dirk Hovy, and Massimo Poesio. 2021. Beyond black & white: Leveraging annotator disagreement via soft-label multi-task learning. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2591–2597, Online. Association for Computational Linguistics. https://doi.org/10.18653/v1/2021.naacl-main.204 ↗
  9. Oren Glickman and Ido Dagan. 2005. A probabilistic setting and lexical coocurrence model for textual entailment. In Proceedings of the ACL Workshop on Empirical Modeling of Semantic Equivalence and Entailment, pages 43–48, Ann Arbor, Michigan. Association for Computational Linguistics. https:// doi.org/10.3115/1631862.1631870 ↗
  10. Mitchell L. Gordon, Kaitlyn Zhou, Kayur Patel, Tatsunori Hashimoto, and Michael S. Bernstein. 2021. The disagreement deconvolution: Bringing machine learning performance metrics in line with reality. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, CHI ’21, New York, NY, USA. Association for Computing Machinery. https://doi.org/10.1145/3411764.3445423 ↗
  11. Herbert P. Grice. 1975. Logic and conversation, Speech Acts, pages 41–58. Brill. https:// doi.org/10.1163/9789004368811 003 ↗
  12. Jeroen Antonius Gerardus Groenendijk and Martin Johan Bastiaan Stokhof. 1984. Studies on the Semantics of Questions and the Pragmatics of Answers. Ph.D. thesis, University of Amsterdam. find
  13. Chuan Guo, Geoff Pleiss, Yu Sun, and Kilian Q. Weinberger. 2017. On calibration of modern neural networks. In Proceedings of the 34th International Conference on Machine Learning, volume 70 of Proceedings of Machine Learning Research, pages 1321–1330. PMLR. find
  14. Nanjiang Jiang and Marie-Catherine de Marneffe. 2019. Evaluating BERT for natural language inference: A case study on the Commitment-Bank. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 6086–6091, Hong Kong, China. Association for Computational Linguistics. https://doi.org/10.18653/v1/D19-1630 ↗
  15. Aikaterini-Lida Kalouli, Annebeth Buis, Livy Real, Martha Palmer, and Valeria de Paiva. 2019. Explaining simple natural language inference. In Proceedings of the 13th Linguistic Annotation Workshop, pages 132–143, Florence, Italy. Association for Computational Linguistics. https://doi.org/10 .18653/v1/W19-4016 Kian Kenyon-Dean, Eisha Ahmed, Scott Fujimoto, Jeremy Georges-Filteau, Christopher Glasz, Barleen Kaur, Auguste Lalande, Shruti Bhanderi, Robert Belfer, Nirmal Kanagasabai, Roman Sarrazingendron, Rohit Verma, and Derek Ruths. 2018. Sentiment analysis: It’s complicated! In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 1886–1895, New Orleans, Louisiana. Association for Computational Linguistics. https://doi.org/10.18653/v1/N18-1171 ↗
  16. Henry E. Kyburg. 1968. Bets and beliefs. American Philosophical Quarterly, 5(1):54–63. find
  17. John P. Lalor, Hao Wu, and Hong Yu. 2017. Soft label memorization-generalization for natural language inference. find
  18. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. RoBERTa: A robustly optimized BERT pretraining approach. find
  19. Peter LoBue and Alexander Yates. 2011. Types of common-sense knowledge needed for recognizing textual entailment. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, pages 329–334, Portland, Oregon, USA. Association for Computational Linguistics. find
  20. Jackson Luken, Nanjiang Jiang, and Marie-Catherine de Marneffe. 2018. QED: A fact verification system for the FEVER shared task. In Proceedings of the First Workshop on Fact Extraction and VERification (FEVER), pages 156–160, Brussels, Belgium. Association for Computational Linguistics. https:// doi.org/10.18653/v1/W18-5526 ↗
  21. Marco Marelli, Stefano Menini, Marco Baroni, Luisa Bentivogli, Raffaella Bernardi, and Roberto Zamparelli. 2014. A SICK cure for the evaluation of compositional distributional semantic models. In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pages 216–223, Reykjavik, Iceland. European Language Resources Association (ELRA). ↗
  22. Marie-Catherine de Marneffe, Christopher D. Manning, and Christopher Potts. 2012. Did it happen? The pragmatic complexity of veridicality assessment. Computational Linguistics, 38:301–333. https://doi.org/10.1162/COLI a 00097 and Morphology, pages 80–86, Florence, Italy. Association for Computational Linguistics. ↗ our card
  23. Marie-Catherine de Marneffe, Anna N. Rafferty, and Christopher D. Manning. 2008. Finding contradictions in text. In Proceedings of ACL-08: HLT, pages 1039–1047, Columbus, Ohio. Association for Computational Linguistics. find
  24. Rebecca Passonneau. 2006. Measuring agreement on set-valued items (MASI) for semantic and pragmatic annotation. In Proceedings of the Fifth International Conference on Language Resources and Evaluation (LREC’06). Genoa, Italy. European Language Resources Association (ELRA). find
  25. Marie-Catherine de Marneffe, Mandy Simons, and Judith Tonhauser. 2019. The Commitment-Bank: Investigating projection in naturally occurring discourse. Proceedings of Sinn und Bedeutung, 23(2):107–124. find
  26. Tom McCoy, Ellie Pavlick, and Tal Linzen. 2019. Right for the wrong reasons: Diagnosing syntactic heuristics in natural language inference. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3428–3448, Florence, Italy. Association for Computational Linguistics. https://doi.org/10.18653/v1/P19 -1334 ↗ our card
  27. Louise McNally. 2016. Modification, Maria Aloni and Paul Dekker, editors, The Cambridge Handbook of Formal Semantics, Cambridge Handbooks in Language and Linguistics, chapter 15, pages 442–464. Cambridge University Press. find
  28. Shachar Mirkin, Ido Dagan, and Sebastian Padó. 2010. Assessing the role of discourse references in entailment inference. In Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics, pages 1209–1219, Uppsala, Sweden. Association for Computational Linguistics. find
  29. Yixin Nie, Xiang Zhou, and Mohit Bansal. 2020. What can we learn from collective human opinions on natural language inference data? In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 9131–9143, Online. Association for Computational Linguistics. https://doi.org/10.18653/v1/2020 .emnlp-main.734 ↗ our card
  30. Byung-Doh Oh, Pranav Maneriker, and Nanjiang Jiang. 2019. THOMAS: The hegemonic OSU morphological analyzer using seq2seq. In Proceedings of the 16th Workshop on Computational Research in Phonetics, Phonology, Rebecca J. Passonneau, Vikas Bhardwaj, Ansaf Salleb-Aouissi, and Nancy Ide. 2012. Multiplicity and word sense: Evaluating and learning from multiply labeled word sense annotations. Language Resources and Evaluation, 46(2):219–252. https://doi.org/10.1007/s10579-012-9188-x Ellie Pavlick and Tom Kwiatkowski. 2019. Inherent disagreements in human textual inferences. Transactions of the Association for Computational Linguistics, 7:677–694. https://doi .org/10.1162/tacl_a_00293 ↗ our card
  31. Barbara Plank, Dirk Hovy, and Anders Søgaard. 2014a. Learning part-of-speech taggers with inter-annotator agreement loss. In Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 742–751, Gothenburg, Sweden. Association for Computational Linguistics. https://doi.org/10.3115/v1/E14-1078 ↗
  32. Barbara Plank, Dirk Hovy, and Anders Søgaard. 2014b. Linguistically debatable or just plain wrong? In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 507–511, Baltimore, Maryland. Association for Computational Linguistics. https:// doi.org/10.3115/v1/P14-2083 ↗
  33. Massimo Poesio and Ron Artstein. 2005. The reliability of anaphoric annotation, reconsidered: Taking ambiguity into account. In Proceedings of the Workshop on Frontiers in Corpus Annotations II: Pie in the Sky, pages 76–83, Ann Arbor, Michigan. Association for Computational Linguistics. https://doi.org/10 .3115/1608829.1608840 ↗
  34. Massimo Poesio, Jon Chamberlain, Silviu Paun, Juntao Yu, Alexandra Uma, and Udo Kruschwitz. 2019. A crowdsourced corpus of multiple judgments and disagreement on anaphoric interpretation. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 1778–1789, Minneapolis, Minnesota. Association for Computational Linguistics. https://doi.org/10.18653/v1/N19 -1176 ↗ our card
  35. Christopher Potts. 2005. The Logic of Conventional Implicatures. Oxford Studies in Theoretical Linguistics. Oxford University Press, Oxford. find
  36. Marta Recasens, Eduard Hovy, and M. Antònia Martı́. 2011. Identity, non-identity, and near-identity: Addressing the complexity of coreference. Lingua, 121(6):1138–1152. https:// doi.org/10.1016/j.lingua.2011.02.004 ↗
  37. Craige Roberts. 2012. Information structure in discourse: Towards an integrated formal theory of pragmatics. Semantics and Pragmatics, 5. https://doi.org/10.3765/sp.5.6 ↗ our card
  38. Mark Sammons, V. G. Vinod Vydiswaran, and Dan Roth. 2010. ‘‘Ask not what textual entailment can do for you...’’. In Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics, pages 1199–1208, Uppsala, Sweden. Association for Computational Linguistics. find
  39. Satoshi Sekine, Kentaro Inui, Ido Dagan, Bill Dolan, Danilo Giampiccolo, and Bernardo Magnini, editors. 2007. Proceedings of the ACL-PASCAL Workshop on Textual Entailment and Paraphrasing. Prague. Association for Computational Linguistics. find
  40. Thibault Sellam, Dipanjan Das, and Ankur Parikh. 2020. BLEURT: Learning robust metrics for text generation. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7881–7892, Online. Association for Computational Linguistics. https://doi.org/10.18653/v1/2020.acl-main.704 ↗
  41. Mandy Simons, Judith Tonhauser, David Beaver, and Craige Roberts. 2010. What projects and why. In Semantics and linguistic theory, volume 20, pages 309–327. https://doi.org/10.3765/salt.v20i0.2584 ↗
  42. James Thorne, Andreas Vlachos, Christos Christodoulopoulos, and Arpit Mittal. 2018. FEVER: A large-scale dataset for fact extraction and VERification. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 809–819, New Orleans, Louisiana. Association for Computational Linguistics. https://doi.org/10.18653/v1/2020.acl-main.704 ↗
  43. Alexandra Uma, Dina Almanea, and Massimo Poesio. 2022. Scaling and disagreements: Bias, noise, and ambiguity. Frontiers in Artificial Intelligence, 5. https://doi.org/10.3389/frai.2022.818451, PubMed: 35434607 ↗
  44. Alexandra N. Uma, Tommaso Fornaciari, Dirk Hovy, Silviu Paun, Barbara Plank, and Massimo Poesio. 2021. Learning from disagreement: A survey. Journal of Artificial Intelligence Research, 72:1385–1470. https://doi.org/10.1613/jair.1.12752 ↗
  45. Yannick Versley. 2008. Vagueness and referential ambiguity in a large-scale annotated corpus. Research on Language and Computation, 6(3):333–353. https://doi.org/10 .1007/s11168-008-9059-1 ↗
  46. Alex Wang, Amanpreet Singh, Julian Michael, Felix Hill, Omer Levy, and Samuel R. Bowman. 2019a. GLUE: A multi-task benchmark and analysis platform for natural language understanding. In International Conference on Learning Representations. https://doi.org/10.18653/v1/W18-5446 ↗ our card
  47. Alex Wang, Ian F. Tenney, Yada Pruksachatkun, Phil Yeres, Jason Phang, Haokun Liu, Phu Mon Htut, Katherin Yu, Jan Hula, Patrick Xia, Raghu Pappagari, Shuning Jin, R. Thomas McCoy, Roma Patel, Yinghui Huang, Edouard Grave, Najoung Kim, Thibault Févry, Berlin Chen, Nikita Nangia, Anhad Mohananey, Katharina Kann, Shikha Bordia, Nicolas Patry, David Benton, Ellie Pavlick, and Samuel R. Bowman. 2019b. jiant 1.3: A software toolkit for research on general-purpose text understanding models. http://jiant.info/. ↗
  48. Adina Williams, Nikita Nangia, and Samuel Bowman. 2018. A broad-coverage challenge corpus for sentence understanding through inference. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 1112–1122, New Orleans, Louisiana, Association for Computational Linguistics. https://doi.org/10.18653/v1/N18-1101 ↗
  49. Adina Williams, Tristan Thrush, and Douwe Kiela. 2022. ANLIzing the adversarial natural language inference dataset. In Proceedings of the Society for Computation in Linguistics, volume 5. University of Massachusetts Amherst. find
  50. Shujian Zhang, Chengyue Gong, and Eunsol Choi. 2021. Learning with different amounts of annotation: From zero to many labels. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 7620–7632, Online and Punta Cana, Dominican Republic, Association for Computational Linguistics. https://doi.org/10 .18653/v1/2021.emnlp-main.601 ↗
  51. Xinliang Frederick Zhang and Marie-Catherine de Marneffe. 2021. Identifying inherent disagreement in natural language inference. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4908–4915, Online. Association for Computational Linguistics. find
  52. Xiang Zhou, Yixin Nie, and Mohit Bansal. 2021. Distributed NLI: learning to predict human opinion distributions for language reasoning. CoRR, abs/2104.08676. https:// doi.org/10.18653/v1/2022.findings -acl.79 ↗