Skip to main content

Advertisement

SpringerLink
Log in

Embodied human language models vs. Large Language Models, or why Artificial Intelligence cannot explain the modal be able to

  • Research
  • Published:
Biosemiotics Aims and scope Submit manuscript

Abstract

This paper explores the challenges posed by the rapid advancement of artificial intelligence specifically Large Language Models (LLMs). I show that traditional linguistic theories and corpus studies are being outpaced by LLMs’ computational sophistication and low perplexity levels. In order to address these challenges, I suggest a focus on language as a cognitive tool shaped by embodied-environmental imperatives in the context of Agentive Cognitive Construction Grammar. To that end, I introduce an Embodied Human Language Model (EHLM), inspired by Active Inference research, as a promising alternative that integrates sensory input, embodied representations, and adaptive strategies for contextualized analysis and conceptual utility maximization. By incorporating Active Inference, which sees perception as inferring the world's state from sensory data, the findings reveal that the characterization of the English modal be able to, as a triadic construction encoding biological intelligent agency, introduces a more plausible theoretical basis for the positing of linguistic constructions. This emphasizes the crucial role of embodied human language models in the comprehension of how humans construct preferred futures through language.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Data availability

There is no data in this manuscript.

Code availability

No code has been created for the study in the paper.

Notes

  1. Wittgenstein’s Philosophical Investigations (1984 [1953]) provided baffling ideas regarding the nature of language and how normative use is subject to the rules of specific language games and forms of life (a way of doing and living within a community). I argue that Wittgenstein’s idea that rule interpretation is akin to rule postulation is false, since the reasons to do something according to a rule are not dependent on goals existing prior to the accomplishment of a “given task”. The idea I am advancing here, then, is that an expected consequence of endorsing the rule-following paradox is that any institutional use of language, in the context of a form of life or language game, is always open to a hypothesis-driven reconfiguration of the content of the rule due to the intentional character of actions. More concretely put, the content of rules is defined, not by epistemic blindness, and the systematic adherence to conventions (as Wittgenstein suggested), but the predictive character of human cognition. My interpretation posits that both the content of the rule and one's grasp of it fall under the broader category of reasons, thereby eluding a reduction of human action to a strictly "correct" or "incorrect" application of a rule. The reason is that rules do not contain behavior (as some wrongly claim, e.g., Ez-zizi et al., 2023), as reasons do not contain a norm to act in ways that privilege correctness over goals and purpose. Perhaps better put, agents engaging in a social use of language will always have the possibility to modify a given course of action in order to accrue epistemic capital. It is interesting in this connection to note that the possibility of accepting the components of a rule also means that my engagement with the rule itself is defined by self-interest, that is, in Parfit’s words (1984, p. 10), that “[f]or each person, there is one supremely rational ultimate aim: that his life go, for him, as well as possible.” Indeed, the idea that linguistic interpretations are rules for grasping the world according to a fixed form of life contradicts the cognitive imperative pushing us to strive for a form of preferred future in line with an individual’s own intentions and goals. In other words, “closed thoughts” are simply not to be had (see Berto, 2022, pp. 3-4).

  2. Intelligent Agency refers to actions in which agent-conceptualizers engage in a specific behavior during an event that is the product of both perceptual, hypothesis-driven states, and intentional doxic (objectifying) lived experiences (see Husserl, 1989, p. 5), and that are defined by the agent’s intention to attain a goal.

  3. An affordance (Gibson, 1966, 1977) is a potentiality that can be used by an agent in a given way according to its phylogenetic configuration. Objects can thus be used in manners defined by criteria such as weight, size, form, etc., that are perceptually accessible, which results in the emergence of specific action profiles.

References

  • Barbieri, M. (2007). The codes of life: The rules of macroevolution. Springer.

    Google Scholar 

  • Berto, F. (2022). Topics of thought: The logic of knowledge, belief, imagination. Oxford University Press.

    Book  Google Scholar 

  • Boghossian, P., & Williamson, T. (2020). Debating the A priori. Oxford University Press.

    Book  Google Scholar 

  • Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, S., ... Leskovec, J. (2022). On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258

  • Cappelle, B., & Depraetere, I. (2016). Short-circuited interpretations of modal verb constructions: Some evidence from The Simpsons. Constructions and Frames, 8(1), 7–39. https://doi.org/10.1075/cf.8.1.02cap

    Article  Google Scholar 

  • Cappelle, B., Depraetere, I., & Lesuisse, M. (2019). The necessity modals have to, must, need to and should: Using n-grams to help identify common and distinct semantic and pragmatic aspects. Constructions and Frames, 11(2), 220–243. https://doi.org/10.1075/cf.00028.cap

    Article  Google Scholar 

  • Cappelle, B., De Cuypere, L., Depraetere, I., Grandin, C., & Leclercq, B. (2023). Possibility modals: Which conditions make them possible? In S. M. Fitzmaurice & B. Kortmann (Eds.), Models of modals: from pragmatics and corpus linguistics to machine learning (pp. 93–117). De Gruyter Mouton.

    Chapter  Google Scholar 

  • Chomsky, N., Gallego, Á. J., & Ott, D. (2019). Generative grammar and the faculty of language: insights, questions, and challenges. Catalan Journal of Linguistics, Special Issue, 229–261. https://doi.org/10.5565/rev/catjl.288

  • Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36(3), 181–204. https://doi.org/10.1017/S0140525X12000477

    Article  PubMed  Google Scholar 

  • Coghlan, S., & Parker, C. (2023). Harm to non-human animals from AI: A systematic account and framework. Philosophy and Technology, 36. https://doi.org/10.1007/s13347-023-00627-6

  • Davies, M. (2008). The Corpus of Contemporary American English (COCA): 520 Million Words, 1990-Present. Available online at http://corpus.byu.edu/coca/

  • Depraetere, I., Cappelle, B., & Hilpert, M. (2023). Introduction. In S. M. Fitzmaurice & B. Kortmann (Eds.), Models of modals: from pragmatics and corpus linguistics to machine learning (pp. 1–13). De Gruyter Mouton.

    Google Scholar 

  • Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American chapter of the association for computational linguistics: human language technologies, Vol. 1 (Long and Short Papers) (pp. 4171–4186). Association for Computational Linguistics.

  • Divjak, D., Romain, L., & Milin, P. (2023). From their point of view: The article category as a hierarchically structured referent tracking system. Linguistics. https://doi.org/10.1515/ling-2022-0186

    Article  Google Scholar 

  • Ez-zizi, A., Divjak, D., & Milin, P. (2023). Error-Correction Mechanisms in Language Learning: Modeling Individuals. Language Learning, 1–37. https://doi.org/10.1111/lang.12569

  • Favareau, D. (2009). Essential readings in biosemiotics: Anthology and commentary. Springer.

    Book  Google Scholar 

  • Flach, S., Cappelle, B., & Hilpert, M. (2023). You must/have to choose: Experimenting with choices between near-synonymous modals. In S. M. Fitzmaurice & B. Kortmann (Eds.), Models of modals: from pragmatics and corpus linguistics to machine learning (pp. 149–176). De Gruyter Mouton.

    Chapter  Google Scholar 

  • Gibson, J. J. (1966). The senses considered as perceptual systems. Houghton Mifflin.

    Google Scholar 

  • Gibson, J. J. (1977). The ecological approach to visual perception. Houghton Mifflin.

    Google Scholar 

  • Goldberg, A. E. (1995). Constructions: A Construction Grammar Approach to Argument Structure. University of Chicago Press.

    Google Scholar 

  • Goldberg, A. E. (2006). Constructions at Work: The Nature of generalization in language. Oxford University Press.

    Google Scholar 

  • Friston, K. (2005). A theory of cortical responses. Philosophical Transactions of the Royal Society of London, 360, 815–836. https://doi.org/10.1098/rstb.2005.1622

    Article  PubMed  PubMed Central  Google Scholar 

  • Friston, K. (2009). The Free-Energy Principle: A rough guide to the brain? Trends in Cognitive Science, 13, 293–301. https://doi.org/10.1016/j.tics.2009.04.005

    Article  Google Scholar 

  • Grabar, Hamon, & Leclercq, T. L. (2023). Modals as a predictive factor for L2 proficiency level. In S. M. Fitzmaurice & B. Kortmann (Eds.), Models of modals: from pragmatics and corpus linguistics to machine learning (pp. 199–224). DeGruyter Mouton.

    Chapter  Google Scholar 

  • Hilpert, M. (2013). Die englischen Modalverben im Daumenkino: Zur dynamischen Visualisierung von Phänomenen des Sprachwandels. Zeitschrift Für Literaturwissenschaft Und Linguistik, 42, 67–82.

    Google Scholar 

  • Hilpert, M., & Perek, F. (2022). You don’t get to see that every day: On the development of permissive get. Constructions and Frames, 14(1), 13–40. https://doi.org/10.1075/cf.20011.hil

    Article  Google Scholar 

  • Hilpert, M., & Flach, S. (2023). Modals in the network model of construction grammar. In S. M. Fitzmaurice & B. Kortmann (Eds.), Models of modals: from pragmatics and corpus linguistics to machine learning (pp. 254–269). DeGruyter Mouton.

    Chapter  Google Scholar 

  • Hoffemeyer, J. (2008). Biosemiotics: An examination into the signs of life and life of signs. University of Scranton Press.

    Google Scholar 

  • Huber, E., Sauppe, S., Isasi-Isasmendia, A., Bornkessel-Schlesewsky, I., Merlo, P., & Bickela, B. (2023). Surprisal from language models can predict ERPs in processing predicate-argument structures only if enriched by an Agent Preference principle. Neurobiology of Language. Advance publication. https://doi.org/10.1162/nol_a_00121

  • Hufeld, C., & Schmid, H. J. (2023). Does the intersubjectivity of modal verbs boost inter-individual differences? In S. M. Fitzmaurice & B. Kortmann (Eds.), Models of modals: from pragmatics and corpus linguistics to machine learning (pp. 177–196). DeGruyter Mouton.

    Chapter  Google Scholar 

  • Jackendoff, R. (2012). A user’s guide to thought and meaning. Oxford University Press.

    Google Scholar 

  • Jary, M. (2022). Nothing is said: Utterance and interpretation. Oxford University Press.

    Book  Google Scholar 

  • Jiang, N., & Nekrasova, T. M. (2007). The processing of formulaic sequences by second language speakers. The Modern Language Journal, 91(3), 433–445. https://doi.org/10.1111/j.1540-4781.2007.00621.x

    Article  Google Scholar 

  • Leclercq, B., & Depraetere, I. (2021). Making meaning with be able to: Modality and actualisation. English Language and Linguistics, 26(1), 27–48. https://doi.org/10.1017/S1360674320000341

    Article  Google Scholar 

  • Leone, M. (2023). The main tasks of a semiotics of artificial Intelligence. Language and Semiotic Studies. https://doi.org/10.1515/lass-2022-0006

    Article  PubMed  PubMed Central  Google Scholar 

  • Lewis, D. (2002). Convention: A philosophical study. Blackwell Publishers.

    Book  Google Scholar 

  • Linzen, T., & Baroni, M. (2021). Syntactic Structure from Deep Learning. Annual Review of Linguistics, 7(1), 195–212. https://doi.org/10.1146/annurev-linguistics-011619-030303

    Article  Google Scholar 

  • Ludlow, P., & Živanović, S. (2022). Language, form, and logic: In pursuit of Natural Logic’s Holy Grail. Oxford University Press.

    Book  Google Scholar 

  • Magnani, L. (2018). Eco-Cognitive Computationalism: From mimetic minds to morphology-based enhancement of mimetic bodies. Entropy, 20, 430. https://doi.org/10.3390/e20060430

    Article  PubMed  PubMed Central  Google Scholar 

  • Nekrasova, T. M. (2009). English L1 and L2 speakers’ knowledge of lexical bundles. Language Learning, 59(3), 647–686. https://doi.org/10.1111/j.1467-9922.2009.00524.x

    Article  Google Scholar 

  • Nowakowski, P. (2017). Bodily processing: The role of morphological computation. Entropy, 19, 295. https://doi.org/10.3390/e19070295

    Article  Google Scholar 

  • Odling-Smee, F. J., Laland, K. N., & Feldman, M. W. (2003). Niche construction: The neglected process in evolution. Princeton University Press.

    Google Scholar 

  • Parfit, D. (1984). Reasons and persons. Oxford University Press.

    Google Scholar 

  • Peng, Y., Wang, Z., Zhang, Q., Du, S., Zhao, Y., Yang, L., Liu, J., Cheng, Y., Wang, A., & Liu, Y. (2019). Basic research on wireless remote control rabbit animal robot movement. In H. Yu, J. Liu, L. Liu, Z. Ju, Y. Liu, & D. Zhou (Eds.), Intelligent Robotics and Applications. ICIRA 2019. Lecture Notes in Computer Science, vol. 11741. Springer, Cham. https://doi.org/10.1007/978-3-030-27532-7_4

  • Petrilli, S., & Ponzio, A. (2015). Language as primary modeling and natural languages: A biosemiotic perspective. In E. Velmezova, K. Kull, & S. J. Cowley (Eds.), Biosemiotic perspectives on language and linguistics (pp. 47–76). Springer.

    Chapter  Google Scholar 

  • Piantadosi, S. (2023). Modern language models refute Chomsky’s approach to language. LingBuzz. Retrieved from https://lingbuzz.net/lingbuzz/007180

  • Pietarinen, A-V, & Beni, M. (2021). Active Inference and abduction. Biosemiotics, 499–517. https://doi.org/10.1007/s12304-021-09432-0

  • Romanini, V, & Lacková, L. (2023). Morphoesthetics in artificial intelligence: proteins versus machines. Transdisciplinary Journal of Linguistics. : https://doi.org/10.53987/2178-5368-2023-12-08

  • Romano, D., Donati, E., Benelli, G., & Stefanini, C. (2018). A review on animal–robot interaction: From bio-hybrid organisms to mixed societies. Biological Cybernetics. https://doi.org/10.1007/s00422-018-0787-5

    Article  PubMed  PubMed Central  Google Scholar 

  • Sarker, I. H. (2021). Deep learning: A comprehensive overview on techniques, taxonomy, applications and research directions. SN Computer Science, 2, Article 420. https://doi.org/10.1007/s42979-021-00711-3

  • Schmid, H.-J. 2020. The dynamics of the linguistic system. Usage, conventionalization, and entrenchment. Oxford University Press. https://doi.org/10.1093/oso/9780198814771.001.0001

  • Sharov, A. (2018). Mind, agency, and biosemiotics. Journal of Cognitive Science, 19(2), 195–228.

    Article  Google Scholar 

  • Sharov, A., & Tønnessen, M. (2021). Semiotic agency: Science beyond mechanism. Springer Nature.

    Book  Google Scholar 

  • Silvennoinen, O. (2023). Is construction grammar cognitive? Constructions, 15 (1). https://doi.org/10.24338/cons-544

  • Tønnessen, M. (2015). Introduction: The relevance of Uexküll’s Umwelt theory today. In C. Brentari (Ed.), Jakob von Uexküll: The discovery of the Umwelt between Biosemiotics and theoretical biology (pp. 1–19). Springer.

    Google Scholar 

  • Torres-Martínez, S. (2018). Constructions as triads of form, function and agency: An agentive cognitive construction grammar analysis of English modals. Cognitive Semantic, 4(1), 1–38.

    Article  Google Scholar 

  • Torres-Martínez, S. (2021). Complexes, rule-following, and language games: Wittgenstein’s philosophical method and its relevance to semiotics. Semiotica, 242, 63–100.

    Article  Google Scholar 

  • Torres-Martínez, S. (2023a). A radical embodied characterization of German Modals. Cognitive Semantics, 9(1), 132–168.

    Article  Google Scholar 

  • Torres-Martínez, S. (2023b). An Integrated Bayesian-Heuristic semiotic model for understanding human and SARS-CoV-2 representational structures. Biosemiotics. https://doi.org/10.1007/s12304-023-09546-7

    Article  Google Scholar 

  • Torres-Martínez, S. (2024). Agentive Cognitive Construction Grammar: A predictive semiotic theory of mind and language. Semiotica (Accepted manuscript). https://doi.org/10.1515/sem-2018-0138

  • Wang, J., Chen, W., Xiao, X., Xu, Y., Li, C., Jia, X., & Meng, M.Q.-H. (2021). A survey of the development of biomimetic intelligence and robotics. Biomimetic Intelligence and Robotics. https://doi.org/10.1016/j.birob.2021.100001

    Article  Google Scholar 

  • Wilson, M. (2022). Imitation of Rigor: An alternative history of analytic philosophy. Oxford University Press.

    Google Scholar 

  • Winter, B., Fischer, M. H., Scheepers, C., & Myachykov, A. (2023). More is better: English language statistics are biased toward addition. Cognitive Science, 47, e13254. https://doi.org/10.1111/cogs.13254

    Article  PubMed  Google Scholar 

  • Wittgenstein, L. (1984). Werkausgabe, Band 1: Tractatus Logico-Philosophicus/Tagebücher 1914–1916 /Philosophische Untersuchungen. Suhrkamp Verlag

  • Woodin, G., Winter, B., Littlemore, J., Perlman, M., & Grieve, J. (2023). Large-scale patterns of number use in spoken and written English. Corpus Linguistics and Linguistic Theory. https://doi.org/10.1515/cllt-2022-0082

    Article  PubMed  PubMed Central  Google Scholar 

  • Zhou, Z., Mei, H., Li, R., Wang, C., Fang, K., Wang, W., Tang, Y., & Dai, Z. (2022). Progresses of animal robots: A historical review and perspectiveness. Heliyon, 8, e11499. https://doi.org/10.1016/j.heliyon.2022.e11499

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

I want to thank the Editors and two anonymous reviewers for the insightful comments and suggestions on an earlier version of the manuscript. The remaining errors are mine.

Funding

This research was not funded by any institution or agency.

Author information

Authors and Affiliations

  1. Universidad de Antioquia, Cll. 67 #53 - 108, Medellin, Medellin, Antioquia, Colombia

    Sergio Torres-Martínez

Authors
  1. Sergio Torres-Martínez
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Sergio Torres-Martínez wrote the manuscript, and devised the concepts, equations and formulas.

Corresponding author

Correspondence to Sergio Torres-Martínez.

Ethics declarations

Competing interests

The authors declare no competing interests.

Use of AI tools

No AI tools were used in the construction of the paper.

Ethics approval

Not applicable.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Torres-Martínez, S. Embodied human language models vs. Large Language Models, or why Artificial Intelligence cannot explain the modal be able to. Biosemiotics 17, 185–209 (2024). https://doi.org/10.1007/s12304-024-09553-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s12304-024-09553-2

Keywords

Search

Navigation