Collocational bootstrapping: A hypothesis about the learning of subject-verb agreement in humans and neural networks

Summary

The paper asks how statistical regularities in linguistic input could help a learner acquire syntax. It proposes "collocational bootstrapping": which words tend to occur together hints at which words are syntactically related. The test case is English subject-verb number agreement. In most natural sentences the correct rule (agree with the subject) and a wrong rule (agree with the nearest preceding noun) give the same answer, so the input is ambiguous between them. The authors train neural language models on synthetic sentences that cannot tell the two rules apart. The subject paired with each verb is drawn from a Zipfian distribution whose exponent controls how predictable the pairing is. Evaluation uses sentences where the two rules disagree. Models generalise correctly only when the variability is moderate (exponent around 1.4), and child-directed speech in CHILDES has almost exactly that value (about 1.43). The conclusion is that the frequency structure of real language may itself push both children and models toward the correct syntactic generalisation.

How Mova uses it

• Singular subject with VBP: cited in the seed for the warning rules en.errors.singular-subject-vbp and en.errors.singular-subject-vbp-auxcop (a singular noun subject before a plural-present VBP verb, copula or auxiliary). The paper's "agree with the subject" versus "agree with the nearest noun" contrast frames why these rules key on the nsubj relation in the tree and not on the closest noun.
• The rules themselves come from error-annotation schemes and grammars (VERB:SVA, Fowler, Poutsma, Curme); this paper is supporting background.

Effectiveness in Mova

Background reading, one supporting citation in one seed. Not measured separately. For reference, the seed reports the rules' behaviour on gold data: EWT 2.18, verb 54 checks / 22 hits, copula or auxiliary 129/37, mostly real text errors, with the rest quantity and collective nouns.

👨‍🔬💥

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.conll-main.7 · arXiv 2605.20529

@inproceedings{hobbs2026collocational,
  title = {{Collocational bootstrapping: A hypothesis about the learning of subject-verb agreement in humans and neural networks}},
  author = {Hobbs, Claire and McCoy, R. Thomas},
  booktitle = {Proceedings of the 30th Conference on Computational Natural Language Learning},
  year = {2026},
  doi = {10.18653/v1/2026.conll-main.7},
  eprint = {2605.20529},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.18653/v1/2026.conll-main.7}
}

References

48 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Omri Abend, Tom Kwiatkowski, Nathaniel J. Smith, Sharon Goldwater, and Mark Steedman. 2017. Bootstrapping language acquisition. Cognition, 164:116– 143. find
2. Diederik Kingma and Jimmy Ba. 2015. Adam: A method for stochastic optimization. In International Conference on Learning Representations. find
3. Adhiguna Kuncoro, Chris Dyer, John Hale, Dani Yogatama, Stephen Clark, and Phil Blunsom. 2018. LSTMs can learn syntax-sensitive dependencies well, but modeling structure makes them better. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Pape… find
4. Kathryn Bock and Carol A. Miller. 1991. Broken agreement. Cognitive Psychology, 23(1):45–93. find
5. Noam Chomsky. 1965. Aspects of the Theory of Syntax. MIT Press, Cambridge, MA. find
6. Jeffrey L. Elman. 1990. Finding structure in time. Cognitive Science, 14(2):179–211. find
7. Thomas K. Landauer and Susan T. Dumais. 1997. A solution to Plato’s problem: The latent semantic analysis theory of acquisition, induction, and representation of knowledge. Psychological Review, 104(2):211. find
8. Jeffrey L. Elman. 1991. Distributed representations, simple recurrent networks, and grammatical structure. Machine Learning, 7(2):195–225. find
9. Steven Finch and Nick Chater. 1992. Bootstrapping syntactic categories. In Proceedings of the Annual Meeting of the Cognitive Science Society, volume 14. find
10. Ori Lavi-Rotbain and Inbal Arnon. 2023. Zipfian distributions in child-directed speech. Open Mind: Discoveries in Cognitive Science, 7:1–30. find
11. Robert Frank, Donald Mathis, and William Badecker. 2013. The acquisition of anaphora by simple recurrent networks. Language Acquisition, 20(3):181– 227. find
12. Cara Su-Yi Leong and Tal Linzen. 2026. Manipulating language models’ training data to study syntactic constraint learning: The case of English passivization. Journal of Memory and Language, 149:104751. find
13. Yoav Goldberg. 2019. Assessing BERT’s syntactic abilities. arXiv preprint arXiv:1901.05287. find
14. Tal Linzen, Emmanuel Dupoux, and Yoav Goldberg. 2016. Assessing the ability of LSTMs to learn syntax-sensitive dependencies. Transactions of the Association for Computational Linguistics, 4:521–535. find
15. Rebecca L. Gómez. 2002. Variability and detection of invariant structure. Psychological Science, 13(5):431–436. find
16. Kristina Gulordava, Piotr Bojanowski, Edouard Grave, Tal Linzen, and Marco Baroni. 2018. Colorless green recurrent networks dream hierarchically. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1… find
17. Ilya Loshchilov and Frank Hutter. 2019. Decoupled weight decay regularization. In International Conference on Learning Representations. find
18. Brian MacWhinney. 2000. The CHILDES Project: Tools for Analyzing Talk. Lawrence Erlbaum, Mahwah, NJ. find
19. Michael Maratsos and Mary Anne Chalkley. 1980. The internal language of children’s syntax: The ontogenesis and representation of syntactic categories. Children’s Language. find
20. Rebecca Marvin and Tal Linzen. 2018. Targeted syntactic evaluation of language models. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 1192–1202. find
21. R. Thomas McCoy, Robert Frank, and Tal Linzen. 2020. Does syntax need to grow on trees? Sources of hierarchical inductive bias in sequence-to-sequence networks. Transactions of the Association for Computational Linguistics, 8:125–140. find
22. Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S. Corrado, and Jeff Dean. 2013. Distributed representations of words and phrases and their compositionality. Advances in Neural Information Processing Systems, 26. find
23. Toben H. Mintz. 2003. Frequent frames as a cue for grammatical categories in child directed speech. Cognition, 90(1):91–117. find
24. Kanishka Misra and Najoung Kim. 2024. Generating novel experimental hypotheses from language models: A case study on cross-dative generalization. arXiv preprint arXiv:2408.05086. — [arXiv:2408.05086](https://arxiv.org/abs/2408.05086) find
25. Kanishka Misra and Kyle Mahowald. 2024. Language models learn rare phenomena from less rare phenomena: The case of the missing AANNs. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 913–929. find
26. James L. Morgan and Katherine Demuth. 1996. Signal to syntax: Bootstrapping from speech to grammar in early acquisition. Psychology Press. find
27. Francesca Padovani, Jaap Jumelet, Yevgen Matusevych, and Arianna Bisazza. 2025. Child-directed language does not consistently boost syntax learning in language models. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, page 19746–19767. Association for Computa… find
28. Abhinav Patil, Jaap Jumelet, Yu Ying Chiu, Andy La-pastora, Peter Shen, Lexie Wang, Clevis Willrich, and Shane Steinert-Threlkeld. 2024. Filtered corpus training (FiCT) shows that language models can generalize from indirect evidence. Transactions of the Association for Computational Linguistics, 12… find
29. Lisa S. Pearl and Benjamin Mis. 2016. The role of indirect positive evidence in syntactic acquisition: A look at anaphoric one. Language, 92(1):1–30. find
30. Steven Pinker. 1984. Language Learnability and Language Development. Harvard University Press. find
31. Geoffrey K. Pullum and Barbara C. Scholz. 2002. Empirical assessment of stimulus poverty arguments. The Linguistic Review, 19(1-2):9–50. find
32. Tian Qin, Naomi Saphra, and David Alvarez-Melis. 2025. Data drives unstable hierarchical generalization in LMs. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 11722–11740, Suzhou, China. Association for Computational Linguistics. find
33. Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. OpenAI. find
34. Limor Raviv, Gary Lupyan, and Shawn C. Green. 2022. How variability shapes learning and generalization. Trends in Cognitive Sciences, 26(6):462–483. find
35. Aaron Mueller, Robert Frank, Tal Linzen, Luheng Wang, and Sebastian Schuster. 2022. Coloring the blank slate: Pre-training imparts a hierarchical inductive bias to sequence-to-sequence models. In Findings of the Association for Computational Linguistics: ACL 2022, pages 1352–1368. find
36. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in Neural Information Processing Systems, volume 30. find
37. Karl Mulligan, Robert Frank, and Tal Linzen. 2021. Structure here, bias there: Hierarchical generalization by jointly learning syntactic transformations. In Proceedings of the Society for Computation in Linguistics 2021, pages 125–135. find
38. Jason Wei, Dan Garrette, Tal Linzen, and Ellie Pavlick. 2021. Frequency effects on syntactic rule learning in transformers. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 932–948. find
39. Nazbanou Nozari and Akira Omaki. 2022. Revisiting agreement: Do children and adults compute subject-verb agreement differently? In Proceedings of the Annual Meeting of the Cognitive Science Society, volume 44. find
40. Kenneth Wexler and Peter W. Culicover. 1980. Formal Principles of Language Acquisition. MIT Press. find
41. Luca Onnis, Padraic Monaghan, Morten H. Christiansen, and Nick Chater. 2004. Variability is the spice of learning, and a crucial ingredient for detecting and generalizing in nonadjacent dependencies. In Proceedings of the Annual Meeting of the Cognitive Science Society, volume 26. find
42. Ethan Gotlieb Wilcox, Richard Futrell, and Roger Levy. 2024. Using computational models to test syntactic learnability. Linguistic Inquiry, 55(4):805–848. find
43. Elizabeth Wonnacott, Elissa L. Newport, and Michael K. Tanenhaus. 2008. Acquiring and processing verb argument structure: Distributional learning in a miniature language. Cognitive Psychology, 56(3):165– 209. find
44. Charles Yang. 2016. The Price of Linguistic Productivity: How Children Learn to Break the Rules of Language. MIT press. find
45. Xiulin Yang, Arianna Bisazza, Nathan Schneider, and Ethan Gotlieb Wilcox. 2026a. A unified assessment of the poverty of the stimulus argument for neural language models. Preprint, arXiv:2602.09992. find
46. Xiulin Yang, Heidi Getz, and Ethan Gotlieb Wilcox. 2026b. From linear input to hierarchical structure: Function words as statistical cues for language learning. Preprint, arXiv:2601.21191. find
47. Aditya Yedetore and Najoung Kim. 2024. Semantic training signals promote hierarchical syntactic generalization in transformers. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 4059–4073, Miami, Florida, USA. Association for Computational Linguistics. find
48. George Kingsley Zipf. 1949. Human Behavior and the Principle of Least Effort: An Introduction to Human Ecology. Addison-Wesley Press, Cambridge, MA. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/hobbs-2026-collocational-bootstrapping-hypothesis-about-learning/