Elamur Research⌕
☰
home / papers / 2026 / Is Cross-Lingual Transfer in Bilingual Models Human-Like? A Study with Overlapping Word Forms in Dutch and English

Is Cross-Lingual Transfer in Bilingual Models Human-Like? A Study with Overlapping Word Forms in Dutch and English

Summary

Bilingual readers process words that look the same in both languages differently. Cognates such as winter tend to speed up reading. Interlingual homographs such as Dutch brand ("fire") either slow reading down or have no effect. This paper asks whether bilingual language models show the same pattern.

The authors trained Dutch-English causal Transformers on roughly 400 million tokens from mixed genres. Dutch made up most of the data and was shown first in each epoch, to mimic an unbalanced late bilingual. Across four vocabulary conditions, they varied whether cognates, false friends, both or neither received a single shared embedding. They then tested the models on stimuli from psycholinguistic reading experiments, using surprisal and embedding-similarity analyses.

The models mostly kept the two languages apart. Cross-lingual effects appeared only where embeddings were shared, and in those cases both word types were easier to predict than controls. Regression analyses traced this mainly to word frequency, not to whether form and meaning matched across languages. Only the cognate-only sharing setup reproduced the human pattern. The authors conclude that the way lexical overlap is encoded limits how well such models explain bilingual reading.

Key points

  • Four vocabulary conditions: Full Overlap, Friends Overlap, False Friends Overlap and Minimal Overlap. Punctuation and named entities, detected with spaCy's xx_ent_wiki_sm, are shared in every condition. Vocabulary size ranges from 77,369 to 144,681 entries, and model size from about 145M to 196M parameters.
  • Training corpus: 75% Dutch and 25% English. By genre, it contains 49% Wikipedia, 26% OpenSubtitles and TedTalks, and 25% CC100. Models were trained for 6 epochs.
  • Word lists: the authors' own annotations were merged with existing lists, giving 2,806 friends (cognates and loan words) and 511 false friends. Evaluation stimuli come from studies such as Bultena et al. (2014) and Huisman (2025).
  • Main finding: shared embeddings lowered surprisal for both cognates and false friends. Frequency, not form-meaning consistency, drove this effect. Only Friends Overlap matched the human facilitation-versus-interference pattern.
  • Code: available on GitHub (izaskr/cross_lingual_transfer_dutch_english_forms).

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

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/29mhg9umxszj

@inproceedings{skrjanec2026is,
  title = {{Is Cross-Lingual Transfer in Bilingual Models Human-Like? A Study with Overlapping Word Forms in Dutch and English}},
  author = {Škrjanec, Iza and Winther, Irene Elisabeth and Demberg, Vera and Frank, Stefan L.},
  booktitle = {Proceedings of the 15th Workshop on Cognitive Modeling and Computational Linguistics},
  year = {2026},
  doi = {10.63317/29mhg9umxszj},
  url = {https://doi.org/10.63317/29mhg9umxszj}
}

References

43 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Tatsuya Aoyama and Nathan Schneider. 2024. Modeling nonnative sentence processing with L2 language models. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 4927–4940, Miami, Florida, USA. Association for Computational Linguistics. find
  2. Douglas Bates, Martin Mächler, Ben Bolker, and Steve Walker. 2015. Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1):1–48. ↗ our card
  3. Sybrine Bultena, Ton Dijkstra, and Janet G. van Hell. 2014. Cognate effects in sentence context depend on word class, L2 proficiency, and task. Quarterly Journal of Experimental Psychology, 67:1214–1241. ↗
  4. Tyler A. Chang and Benjamin K. Bergen. 2022. Word acquisition in neural language models. Transactions of the Association for Computational Linguistics, 10:1–16. ↗
  5. Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2020. Unsupervised cross-lingual representation learning at scale. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8440–8451, Online. Association for Computational Linguistics. ↗
  6. Ionut Constantinescu, Tiago Pimentel, Ryan Cotterell, and Alex Warstadt. 2025. Investigating critical period effects in language acquisition through neural language models. Transactions of the Association for Computational Linguistics, 13:96– 120. ↗
  7. Ton Dijkstra and Walter van Heuven. 2002. The architecture of the bilingual word recognition system: From identification to decision. Bilingualism: Language and Cognition, 5:175 – 197. ↗
  8. Kawin Ethayarajh. 2019. How contextual are contextualized word representations? Comparing the geometry of BERT, ELMo, and GPT-2 embeddings. 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 55–65, Hong Kong, China. Association for Computational Linguistics. find
  9. Stefan L. Frank. 2021. Toward computational models of multilingual sentence processing. Language Learning, 71:193–218. ↗
  10. John Hale. 2001. A probabilistic Earley parser as a psycholinguistic model. In Second Meeting of the North American Chapter of the Association for Computational Linguistics. find our card
  11. Liv J. Hoversten and Matthew J. Traxler. 2016. A time course analysis of interlingual homograph processing: Evidence from eye movements*. Bilingualism: Language and Cognition, 19(2):347 – 360. ↗
  12. Merit Huisman. 2025. The role of phonology on recognising interlingual homographs in L1 sentence context. Master’s thesis, Radboud University, Nijmegen, the Netherlands. find
  13. Arthur M. Jacobs and Jonathan Grainger. 1994. Models of visual word recognition: Sampling the state of the art. Journal of Experimental Psychology: Human Perception and Performance, 20(6):1311–1334. ↗
  14. Karthikeyan K, Zihan Wang, Stephen Mayhew, and Dan Roth. 2020. Cross-Lingual Ability of Multilingual BERT: An Empirical Study. In Proceedings of International Conference on Learning Representations (ICLR). ↗
  15. Julie Kallini, Dan Jurafsky, Christopher Potts, and Martijn Bartelds. 2025. False Friends are not foes: Investigating vocabulary overlap in multilingual language models. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 21138–21154, Suzhou, China. Association for Computational Linguistics. find
  16. Judith F. Kroll, Paola E. Dussias, Cari A. Bogul-ski, and Jorge R. Valdes Kroff. 2012. Juggling two languages in one mind: What bilinguals tell us about language processing and its consequences for cognition. volume 56 of Psychology of Learning and Motivation, pages 229–262. Academic Press. find
  17. Tatsuki Kuribayashi, Yohei Oseki, Souhaib Ben Taieb, Kentaro Inui, and Timothy Baldwin. 2025. Large language models are human-like internally. Transactions of the Association for Computational Linguistics, 13:1743–1766. ↗
  18. Justin Lauro and Ana I. Schwartz. 2017. Bilingual non-selective lexical access in sentence contexts: A meta-analytic review. Journal of Memory and Language, 92:217–233. ↗
  19. Els Lefever, Sofie Labat, and Pranaydeep Singh. 2020. Identifying cognates in English-Dutch and French-Dutch by means of orthographic information and cross-lingual word embeddings. In Proceedings of the Twelfth Language Resources and Evaluation Conference, pages 4096–4101, Marseille, France. European Language Resources Association. find
  20. Kristin Lemhöfer and Ton Dijkstra. 2004. Recognizing cognates and interlingual homographs: Effects of code similarity in language-specific and generalized lexical decision. Memory & Cognition, 32:533–550. ↗
  21. Pietro Lesci, Clara Meister, Thomas Hofmann, Andreas Vlachos, and Tiago Pimentel. 2025. Causal estimation of tokenisation bias. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 28325–28340, Vienna, Austria. Association for Computational Linguistics. ↗
  22. Roger Levy. 2008. Expectation-based syntactic comprehension. Cognition, 106:1126–1177. ↗
  23. Maya Libben and Debra A. Titone. 2009. Bilingual lexical access in context: evidence from eye movements during reading. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(2):381–90. ↗
  24. Katherine J. Midgley, Phillip J. Holcomb, and Jonathan Grainger. 2011. Effects of cognate status on word comprehension in second language learners: An erp investigation. Journal of Cognitive Neuroscience, 23(7):1634–1647. ↗
  25. Miyu Oba, Tatsuki Kuribayashi, Hiroki Ouchi, and Taro Watanabe. 2023. Second language acquisition of neural language models. In Findings of the Association for Computational Linguistics: ACL 2023, pages 13557–13572, Toronto, Canada. Association for Computational Linguistics. ↗
  26. Byung-Doh Oh and William Schuler. 2022. Entropy-and distance-based predictors from GPT-2 attention patterns predict reading times over and above GPT-2 surprisal. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 9324–9334, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. ↗
  27. Mohammad Taher Pilehvar and Jose Camacho-Collados. 2019. WiC: the word-in-context dataset for evaluating context-sensitive meaning representations. 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 1267–1273, Minneapolis, Minnesota. Association for Computational Linguistics. ↗ our card
  28. Telmo Pires, Eva Schlinger, and Dan Garrette. 2019. How multilingual is multilingual BERT? In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4996–5001, Florence, Italy. Association for Computational Linguistics. ↗ our card
  29. Irina Pivneva, Julie Mercier, and Debra Titone. 2014. Executive control modulates cross-language lexical activation during l2 reading: Evidence from eye movements. Journal of Experimental Psychology: Learning, Memory, and Cognition, 40(3):787–796. ↗
  30. Eva D. Poort and Jennifer M. Rodd. 2019. A Database of Dutch-English Cognates, Interlingual Homographs and Translation Equivalents. Journal of Cognition, 2(1):15. ↗
  31. Alessandro Raganato, Tommaso Pasini, Jose Camacho-Collados, and Mohammad Taher Pilehvar. 2020. XL-WiC: A multilingual benchmark for evaluating semantic contextualization. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 7193–7206, Online. Association for Computational Linguistics. find
  32. Yasaman Razeghi, Robert L Logan IV, Matt Gardner, and Sameer Singh. 2022. Impact of pre-training term frequencies on few-shot numerical reasoning. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 840–854, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. find
  33. Rasmus Roslund and Yevgen Matusevych. 2022. Modeling sentence processing effects in bilingual speakers: A comparison of neural architectures. In Annual Meeting of the Cognitive Science Society. find
  34. Job Schepens, Ton Dijkstra, Franc Grootjen, and Walter J. B. van Heuven. 2013. Cross-language distributions of high frequency and phonetically similar cognates. PLOS ONE, 8(5):1–15. ↗
  35. Cory Shain, Clara Meister, Tiago Pimentel, Ryan Cotterell, and Roger Levy. 2024. Large-scale evidence for logarithmic effects of word predictability on reading time. Proceedings of the National Academy of Sciences, 121(10):e2307876121. ↗
  36. Kristof Strijkers, Albert Costa, and Guillaume Thierry. 2010. Tracking lexical access in speech production: Electrophysiological correlates of word frequency and cognate effects. Cerebral Cortex, 20(4):912–928. ↗
  37. William Timkey and Marten van Schijndel. 2021. All bark and no bite: Rogue dimensions in transformer language models obscure representational quality. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 4527–4546, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics. find
  38. Debra A. Titone, Maya Libben, Julien Mercier, Veronica Whitford, and Irina Pivneva. 2011. Bilingual lexical access during l1 sentence reading: The effects of l2 knowledge, semantic constraint, and l1-l2 intermixing. Journal of experimental psychology. Learning, memory, and cognition, 37 6:1412–31. ↗
  39. Madeleine Voga and Jonathan Grainger. 2007. Cognate status and cross-script translation priming. Memory & Cognition, 35(5):938–952. ↗
  40. Alex Warstadt and Samuel R. Bowman. 2022. What artificial neural networks can tell us about human language acquisition. In Algebraic Structures in Natural Language. ↗
  41. Irene Elisabeth Winther, Yevgen Matusevych, and Martin J. Pickering. 2021. Cumulative frequency can explain cognate facilitation in language models. In Annual Meeting of the Cognitive Science Society. ↗
  42. Shijie Wu and Mark Dredze. 2019. Beto, bentz, becas: The surprising cross-lingual effectiveness of BERT. 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 833–844, Hong Kong, China. Association for Computational Linguistics. ↗ our card
  43. Aditya Yadavalli, Alekhya Yadavalli, and Vera Tobin. 2023. SLABERT talk pretty one day: Modeling second language acquisition with BERT. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 11763–11777, Toronto, Canada. Association for Computational Linguistics. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/skrjanec-2026-is-cross-lingual-transfer-bilingual/