Hostname: page-component-76d6cb85b7-s74w7 Total loading time: 0 Render date: 2026-07-11T07:34:51.743Z Has data issue: false hasContentIssue false
  • English
  • Français

Machine yearning: LLMs do not capture formal linguistic structure and obscure neuroscientific inquiry

Published online by Cambridge University Press:  01 July 2026

Elliot Murphy Open the ORCID record for Elliot Murphy [Opens in a new window] ,
Paolo Morosi ,
Evelina Leivada Open the ORCID record for Evelina Leivada [Opens in a new window]  and
Andrew Nevins
Elliot Murphy*
Affiliation:
Vivian L. Smith Department of Neurosurgery, University of Texas Health Science Center at Houston, Houston, USA elliot.murphy@uth.tmc.edu
Paolo Morosi
Affiliation:
Universitat Autonoma de Barcelona, Spain paolo.morosi@uab.cat
Evelina Leivada
Affiliation:
Universitat Autonoma de Barcelona, Spain paolo.morosi@uab.cat Institució Catalana de Recerca i Estudis Avançats (ICREA), Spain evelina.leivada@uab.cat
Andrew Nevins
Affiliation:
Department of Linguistics, University College London, UK a.nevins@ucl.ac.uk
*
*Corresponding author.
Get access Rights & Permissions [Opens in a new window]

Abstract

Futrell and Mahowald argue that neural networks have learned non-trivial aspects of language. We argue that these systems have not in fact demonstrated “mastery” of syntax, marshalling recent evidence, and that they further obscure explanatory insights with respect to topics in the cognitive neuroscience of language.

Information

Type
Open Peer Commentary
Information
Behavioral and Brain Sciences , Volume 49 , 2026 , e217
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

Baggio, G. (2018). Meaning in the brain. MIT Press.10.7551/mitpress/11265.001.0001CrossRefGoogle Scholar
Dentella, V., Günther, F., & Leivada, E. (2025a). Language in vivo vs. in silico: Size matters but larger language models still do not comprehend language on a par with humans due to impenetrable semantic reference. PLoS One, 20(7), e0327794.10.1371/journal.pone.0327794CrossRefGoogle Scholar
Dentella, V., Günther, F., Murphy, E., Marcus, G., & Leivada, E. (2024). Testing AI on language comprehension tasks reveals insensitivity to underlying meaning. Scientific Reports, 14, 28083.10.1038/s41598-024-79531-8CrossRefGoogle ScholarPubMed
Dentella, V., Huang, W., Mansi, S. A., Grieve, J., & Leivada, E. (2025b). ChatGPT–generated texts show authorship traits that identify them as non-human. arXiv:2508.16385.Google Scholar
Diego-Simón, P., D’Ascoli, S., Chemla, E., Lakretz, Y., & King, J. R. (2024). A polar coordinate system represents syntax in large language models. In 38th Conference on Neural Information Processing Systems (NeurIPS 2024), Vancouver, Canada.10.52202/079017-3344CrossRefGoogle Scholar
Diego-Simón, P. J., Chemla, E., King, J. R., & Lakretz, Y. (2025). Probing syntax in large language models: Successes and remaining challenges. arXiv:2508.03211.Google Scholar
Fodor, J., Murawski, C., & Suzuki, S. (2025). When word order matters: Human brains represent sentence meaning differently from large language models. eLife, 14, RP108442.Google Scholar
Guest, O., & Martin, A. E. (2023). On logical inference over brains, behaviour, and artificial neural networks. Computational Brain & Behavior, 6, 213–227.10.1007/s42113-022-00166-xCrossRefGoogle Scholar
Hadidi, N., Feghhi, E., Song, B. H., Blank, I. A., & Kao, J. C. (2025). Illusions of alignment between large language models and brains emerge from fragile methods and overlooked confounds. bioRxiv https://doi.org/10.1101/2025.03.09.642245.CrossRefGoogle Scholar
Hunter, T. (2024). Kallini et al. (2024) do not compare impossible languages with constituency-based ones. arXiv:2410.12271.Google Scholar
Jackendoff, R., & Erk, K. E. (2025). Toward a deeper lexical semantics. Topics in Cognitive Science, 17 (4), 962–972. https://doi.org/10.1111/tops.70013.CrossRefGoogle Scholar
Kallini, J., Papadimitriou, I., Futrell, R., Mahowald, K., & Potts, C. (2024). Mission: Impossible language models. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics, 14691–14714. https://doi.org/10.18653/v1/2024.acl-long.787 CrossRefGoogle Scholar
Kobzeva, A., Arehalli, S., Linzen, T., & Kush, D. (2025). Learning filler-gap dependencies with neural language models: Testing island sensitivity in Norwegian and English. Journal of Memory and Language, 144, 104663.10.1016/j.jml.2025.104663CrossRefGoogle Scholar
Krakauer, D. C., Krakauer, J. W., & Mitchell, M. (2025). Large language models and emergence: A complex systems perspective. arXiv:2506.11135.Google Scholar
Leivada, E., Marcus, G., Günther, F., & Murphy, E. (2025b). A sentence is worth a thousand pictures: Can Large Language Models understand hum4n l4ngu4ge and the w0rld behind w0rds? Philosophical Transactions of the Royal Society A (Forthcoming).10.1098/rsta.2025.0008CrossRefGoogle Scholar
Leivada, E., Montero, R., Morosi, P., Moskvina, N., Serrano, T., Aguilar, M., & Guenther, F. (2025a). Large language model probabilities cannot distinguish between possible and impossible languages. arXiv:2509.15114.Google Scholar
Leivada, E., Murphy, E., & Marcus, G. (2023). DALL·E 2 fails to reliably capture common syntactic processes. Social Sciences & Humanities Open, 8(1), 100648.10.1016/j.ssaho.2023.100648CrossRefGoogle Scholar
Marcolli, M., Chomsky, N., & Berwick, R. C. (2025). Mathematical structure of syntactic merge: An algebraic model for generative linguistics. MIT Press.10.7551/mitpress/15811.001.0001CrossRefGoogle Scholar
Marcus, G. F. (2001). The algebraic mind: Integrating connectionism and cognitive science. MIT Press.10.7551/mitpress/1187.001.0001CrossRefGoogle Scholar
McDermott, D. (1981). Artificial Intelligence meets natural stupidity. ACM SIGART Bulletin, 57, 4–9.Google Scholar
Meijer, G. (2021). Neurons in the mouse brain correlate with cryptocurrency price: A cautionary tale. Peer Community Journal, 1, e29.10.24072/pcjournal.30CrossRefGoogle Scholar
Mitchell, M. (2024). The metaphors of artificial intelligence. Science, 386(6723), eadt6140.10.1126/science.adt6140CrossRefGoogle ScholarPubMed
Momma, S., Richards, N., & Ferreira, V. S. (2025). Speakers encode silent structures: Evidence from complementizer priming in English. Journal of Memory and Language, 145, 104671.10.1016/j.jml.2025.104671CrossRefGoogle Scholar
Murphy, E. (2015). The brain dynamics of linguistic computation. Frontiers in Psychology, 6, 1515.10.3389/fpsyg.2015.01515CrossRefGoogle ScholarPubMed
Murphy, E. (2020). The oscillatory nature of language. Cambridge University Press.10.1017/9781108864466CrossRefGoogle Scholar
Murphy, E. (2024a). Predicate order and coherence in copredication. Inquiry, 67(6), 1744–1780.10.1080/0020174X.2021.1958054CrossRefGoogle Scholar
Murphy, E. (2024b). ROSE: A neurocomputational architecture for syntax. Journal of Neurolinguistics, 70, 101180.10.1016/j.jneuroling.2023.101180CrossRefGoogle Scholar
Murphy, E. (2025). ROSE: A universal neural grammar. Cognitive Neuroscience, 16(1–4), 49–80.10.1080/17588928.2025.2523875CrossRefGoogle Scholar
Murphy, E., de Villiers, J., & Morales, S. L. (2025). A comparative investigation of compositional syntax and semantics in DALL⋅E and young children. Social Sciences & Humanities Open, 11, 101332.10.1016/j.ssaho.2025.101332CrossRefGoogle Scholar
Murphy, E., Leivada, E., Dentella, V., Günther, F., & Marcus, G. (2025). Fundamental principles of linguistic structure are not represented by ChatGPT. Biolinguistics, 19, 1–55.10.5964/bioling.19021CrossRefGoogle Scholar
Murphy, E., Woolnough, O., Rollo, P. S., Roccaforte, Z., Segaert, K., Hagoort, P., & Tandon, N. (2022). Minimal phrase composition revealed by intracranial recordings. Journal of Neuroscience, 42(15), 3216–3227.10.1523/JNEUROSCI.1575-21.2022CrossRefGoogle ScholarPubMed
Shojaee, P., Mirzadeh, I., Alizadeh, K., Horton, M., Bengio, S., & Farajtabar, M. (2025). The illusion of thinking: Understanding the strengths and limitations of reasoning models via the lens of problem complexity. Apple Machine Learning Research report. June.10.70777/si.v2i6.15919CrossRefGoogle Scholar
Slaats, S., & Martin, A. E. (2025). What’s surprising about surprisal. Computational Brain & Behavior, 8, 233–248.10.1007/s42113-025-00237-9CrossRefGoogle ScholarPubMed
Yang, X., Aoyama, T., Yao, Y., & Wilcox, E. (2025). Anything goes? A crosslinguistic study of (im)possible language learning in LMs. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) 26058–26077.Google Scholar