Hostname: page-component-76d6cb85b7-dqfph Total loading time: 0 Render date: 2026-07-11T13:29:17.086Z Has data issue: false hasContentIssue false
  • English
  • Français

Across the levels of analysis: Explaining predictive processing in humans requires more than machine-estimated probabilities

Published online by Cambridge University Press:  01 July 2026

Sathvik Nair Open the ORCID record for Sathvik Nair [Opens in a new window]  and
Colin Phillips Open the ORCID record for Colin Phillips [Opens in a new window]
Sathvik Nair*
Affiliation:
University of Maryland, College Park: University of Maryland, USA sathvik@umd.edu
Colin Phillips
Affiliation:
University of Maryland, College Park: University of Maryland, USA sathvik@umd.edu Oxford University: University of Oxford, UK colin.phillips@ling-phil.ox.ac.uk
*
*Corresponding author.
Get access Rights & Permissions [Opens in a new window]

Abstract

Under the lens of Marr’s (levels of analysis, we critique and extend the authors’ two points about language models (LMs) and language processing: first, predicting upcoming linguistic information based on context is key to language processing, and second, that many advances in psycholinguistics would be impossible without LLMs. We also outline directions combining LLMs’ strengths with psycholinguistic models.

Information

Type
Open Peer Commentary
Information
Behavioral and Brain Sciences , Volume 49 , 2026 , e218
Check for updates
Copyright
© The Author(s), 2026. Published by Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Article purchase

Temporarily unavailable

References

Schrimpf, M., Blank, I. A., Tuckute, G., Kauf, C., Hosseini, E. A., Kanwisher, N., Tenenbauam, J.B, & Fedorenko, E. (2021). The neural architecture of language: Integrative modeling converges on predictive processing. Proceedings of the National Academy of Sciences, 118(45), e2105646118.10.1073/pnas.2105646118CrossRefGoogle ScholarPubMed
Aurnhammer, C., Delogu, F., Schulz, M., Brouwer, H., & Crocker, M. W. (2021). Retrieval (N400) and integration (P600) in expectation-based comprehension. Plos one, 16(9), e0257430.10.1371/journal.pone.0257430CrossRefGoogle ScholarPubMed
Brothers, T., & Kuperberg, G. R. (2021). Word predictability effects are linear, not logarithmic: Implications for probabilistic models of sentence comprehension. Journal of Memory and Language, 116, 104174.CrossRefGoogle Scholar
Chow, W. Y., Lau, E., Wang, S., & Phillips, C. (2018). Wait a second! Delayed impact of argument roles on on-line verb prediction. Language, Cognition and Neuroscience, 33(7), 803–828.10.1080/23273798.2018.1427878CrossRefGoogle Scholar
Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36(3), 181–204.10.1017/S0140525X12000477CrossRefGoogle ScholarPubMed
De Santo, A. (2025). Capturing online SRC/ORC effort with memory measures from a minimalist parser. In Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics (pp. 24–35).10.18653/v1/2025.cmcl-1.5CrossRefGoogle Scholar
Elman, J. L. (1991). Distributed representations, simple recurrent networks, and grammatical structure. Machine learning, 7(2), 195–225.10.1023/A:1022699029236CrossRefGoogle Scholar
Fitz, H., & Chang, F. (2019). Language ERPs reflect learning through prediction error propagation. Cognitive Psychology, 111, 15–52.10.1016/j.cogpsych.2019.03.002CrossRefGoogle ScholarPubMed
Frisson, S., Harvey, D. R., & Staub, A. (2017). No prediction error cost in reading: Evidence from eye movements. Journal of Memory and Language, 95, 200–214.10.1016/j.jml.2017.04.007CrossRefGoogle Scholar
Friston, K. (2009). The free-energy principle: A rough guide to the brain?. Trends in Cognitive Sciences, 13(7), 293–301.10.1016/j.tics.2009.04.005CrossRefGoogle ScholarPubMed
Giulianelli, M., Wallbridge, S., Cotterell, R., & Fernández, R. (2024). Incremental alternative sampling as a lens into the temporal and representational resolution of linguistic prediction. PsyArXiv.Google Scholar
Giulianelli, M., Wallbridge, S., & Fernández, R. (2023). Information value: Measuring utterance predictability as distance from plausible alternatives. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (pp. 5633–5653).10.18653/v1/2023.emnlp-main.343CrossRefGoogle Scholar
Goldstein, A., Zada, Z., Buchnik, E., Schain, M., Price, A., Aubrey, B., et al. (2022). Shared computational principles for language processing in humans and deep language models. Nature Neuroscience, 25(3), 369–380.10.1038/s41593-022-01026-4CrossRefGoogle ScholarPubMed
Huang, K. J., Arehalli, S., Kugemoto, M., Muxica, C., Prasad, G., Dillon, B., & Linzen, T. (2024). Large-scale benchmark yields no evidence that language model surprisal explains syntactic disambiguation difficulty. Journal of Memory and Language, 137, 104510.10.1016/j.jml.2024.104510CrossRefGoogle Scholar
Jacobs, C. L., Hubbard, R. J., Grobol, L., & Federmeier, K. D. (2025). Uncovering patterns of semantic predictability in sentence processing. Journal of Memory and Language, 144, 104653.10.1016/j.jml.2025.104653CrossRefGoogle Scholar
Kandel, M., Wyatt, C. R., & Phillips, C. (2024). Number attraction in pronoun production. Open Mind, 8, 1247–1290.10.1162/opmi_a_00167CrossRefGoogle ScholarPubMed
Krieger, B., Brouwer, H., Aurnhammer, C., & Crocker, M. W. (2025). On the limits of LLM surprisal as a functional explanation of the N400 and P600. Brain Research, 1865, 149841.CrossRefGoogle ScholarPubMed
Kuperberg, G. R., Brothers, T., & Wlotko, E. W. (2020). A tale of two positivities and the N400: Distinct neural signatures are evoked by confirmed and violated predictions at different levels of representation. Journal of Cognitive Neuroscience, 32(1), 12–35.10.1162/jocn_a_01465CrossRefGoogle ScholarPubMed
Lee, E. K., Nair, S., & Feldman, N. (2024). A psycholinguistic evaluation of language models’ sensitivity to argument roles. In Findings of the Association for Computational Linguistics: EMNLP 2024 (pp. 3262–3274).Google Scholar
Lee, E. K. R., & Phillips, C. (2025). Argument role sensitivity in real-time sentence processing: Evidence from a hybrid comprehension and production task. Cognition, 264, 106255.10.1016/j.cognition.2025.106255CrossRefGoogle ScholarPubMed
Levy, R. (2008). Expectation-based syntactic comprehension. Cognition, 106(3), 1126–1177.CrossRefGoogle ScholarPubMed
Marr, D. (1982). Vision: A computational investigation into the human representation and processing of visual information. W.H. Freeman.Google Scholar
McClelland, J. L., & Rumelhart, D. E. (1981). An interactive activation model of context effects in letter perception: I. An account of basic findings. Psychological review, 88(5), 375.CrossRefGoogle Scholar
Miller, G. A., & McKean, K. O. (1964). A chronometric study of some relations between sentences. Quarterly Journal of Experimental Psychology, 16(4), 297–308.10.1080/17470216408416385CrossRefGoogle Scholar
Nakamura, M., Momma, S., Sakai, H., & Phillips, C. (2024). Task and timing effects in argument role sensitivity: Evidence from production, EEG, and computational modeling. Cognitive Science, 48(12), e70023.10.1111/cogs.70023CrossRefGoogle ScholarPubMed
Nour Eddine, S. N., Brothers, T., Wang, L., Spratling, M., & Kuperberg, G. R. (2024). A predictive coding model of the N400. Cognition, 246, 105755.10.1016/j.cognition.2024.105755CrossRefGoogle ScholarPubMed
Ohams, C., Nair, S., Bhattasali, S., & Resnik, P. (2026). A predictive coding model for online sentence processing. Journal of Memory and Language, 146, 104705.10.1016/j.jml.2025.104705CrossRefGoogle Scholar
Paape, D. (2024). How do linguistic illusions arise? Rational inference and good-enough processing as competing latent processes within individuals. Language, Cognition and Neuroscience, 39(10), 1334–1365.10.1080/23273798.2024.2387226CrossRefGoogle Scholar
Phillips, C., Wagers, M. W., & Lau, E. F. (2011). Grammatical illusions and selective fallibility in real-time language comprehension. Experiments at the Interfaces, 37, 147–180.CrossRefGoogle Scholar
Rao, R. P., & Ballard, D. H. (1999). Predictive coding in the visual cortex: A functional interpretation of some extra-classical receptive-field effects. Nature Neuroscience, 2(1), 79–87.CrossRefGoogle ScholarPubMed
Ryskin, R., & Nieuwland, M. S. (2023). Prediction during language comprehension: What is next?. Trends in Cognitive Sciences, 27(11), 1032–1052.10.1016/j.tics.2023.08.003CrossRefGoogle ScholarPubMed
Shain, C., Meister, C., Pimentel, T., Cotterell, R., & Levy, R. (2024). Large-scale evidence for logarithmic effects of word predictability on reading time. Proceedings of the National Academy of Sciences, 121(10), e2307876121.10.1073/pnas.2307876121CrossRefGoogle ScholarPubMed
Slaats, S., Meyer, A. S., & Martin, A. E. (2024). Lexical surprisal shapes the time course of syntactic structure building. Neurobiology of Language, 5(4), 942–980.10.1162/nol_a_00155CrossRefGoogle ScholarPubMed
Smith, N. J., & Levy, R. (2013). The effect of word predictability on reading time is logarithmic. Cognition, 128(3), 302–319.CrossRefGoogle ScholarPubMed
Stanojević, M., Brennan, J. R., Dunagan, D., Steedman, M., & Hale, J. T. (2023). Modeling structure-building in the brain with CCG parsing and large language models. Cognitive science, 47(7), e13312.10.1111/cogs.13312CrossRefGoogle ScholarPubMed
Staub, A. (2025). Predictability in language comprehension: Prospects and problems for surprisal. Annual Review of Linguistics, 11(1), 17–34.10.1146/annurev-linguistics-011724-121517CrossRefGoogle Scholar
Szewczyk, J. M., & Federmeier, K. D. (2022). Context-based facilitation of semantic access follows both logarithmic and linear functions of stimulus probability. Journal of memory and language, 123, 104311.10.1016/j.jml.2021.104311CrossRefGoogle ScholarPubMed
Tabor, W., & Hutchins, S. (2004). Evidence for self-organized sentence processing: Digging-in effects. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30(2), 431.Google ScholarPubMed
Timkey, W., Huang, K. J., Oh, B. D., Prasad, G., Arehalli, S., Linzen, T., & Dillon, B. (2025). Eye movements reveal a dissociation between prediction and structural processing in language comprehension. Preprint. 10.31234/osf.io/eq2ra_v1CrossRefGoogle Scholar
Vigly, J., Qian, P., Sonderegger, M., & O’Donnell, T. (2025). Comprehension effort as the cost of inference. PsyArxiv.CrossRefGoogle Scholar
Wilcox, E. G., Pimentel, T., Meister, C., Cotterell, R., & Levy, R. P. (2023). Testing the predictions of surprisal theory in 11 languages. Transactions of the Association for Computational Linguistics, 11, 1451–1470.10.1162/tacl_a_00612CrossRefGoogle Scholar
Xu, W., Chon, J., Liu, T., & Futrell, R. (2023). The linearity of the effect of surprisal on reading times across languages. In Findings of the Association for Computational Linguistics: EMNLP 2023 (pp. 15711–15721).Google Scholar
Zhang, Y., Gibson, E., & Davis, F. (2023). Can language models be tricked by language illusions? Easier with syntax, harder with semantics. In Proceedings of the 27th Conference on Computational Natural Language Learning (CoNLL) (pp. 1–14).10.18653/v1/2023.conll-1.1CrossRefGoogle Scholar