Artificial Intelligence in the Chinese Room: Exploring Meaning Understanding with Large Language Models

Document Type : Original Article

Author
Hamid Mohseni
Assistant Professor, Department of Cyberspace and Artificial Intelligence Studies, Research Institute for Islamic Culture and Thought (IICT).
10.22034/zehn.2026.2090523.2212
Abstract
The main issue of this research is to investigate the possibility of "understanding meaning" in large language models; an issue that has gained increasing importance with the expansion of the use of these models and the attribution of cognitive concepts to them. The aim of the research is to critically analyze this claim by relying on the distinction between "form" and "concept" and to analyze it with the focus on John Searle's "Chinese Room" argument. The research method is analytical-conceptual and based on the examination of philosophical texts and computational linguistics studies. The findings show that language models, despite their high ability to process and produce language, operate solely on the basis of syntactic and statistical patterns and lack access to communicative intent and reference to the outside world. Therefore, what is considered as "understanding" remains at a formal level. The result is that attributing understanding meaning to these models, in the strict philosophical sense, is not justified and the realization of meaning requires some kind of connection with the real world.
Keywords
Artificial intelligence
large language models
semantic understanding
Chinese room reasoning
Subjects
1.    Bender, Emily M. and Koller, Alexander (2020). “Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics”. Association for Computational Linguistics, Online, 5185–5198.
2.    Bender, Emily M. et al (2021). On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?, Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, Canada (pp. 610–623), New York, NY: ACM.
3.    Brown, Tom B., et al, (2020), Language models are few-shot learners, in Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020, December 6-12, virtual, H. Larochelle, M. Ranzato, R. Hadsell, M. Balcan, and H. Lin, Eds.
4.    Carter, Matt (2007). Minds and Computers An Introduction to the Philosophy of Artificial Intelligence, Edinburgh: Edinburgh University Press Ltd 22 George Square.
5.    Chowdhery, Aakanksha, et al (2023). “Palm: Scaling language modeling with pathways,”. https://doi.org/10.48550/arXiv.2204.02311.
6.    Devlin, Jacob, et al (2019). BERT: Pre-training of deep bidirectional transformers for language under standing. 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), pp 4171–4186, Minneapolis, Minnesota. Associ ation for Computational Linguistics.
7.    Douglas Heaven, Will (2020). “OpenAI’s new language generator GPT-3 is shockingly good—and completely mindless”, MIT Technology Review.
8.    Dwivedi, Yogesh K. et al (2023). “So what if ChatGPT wrote it? Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy”, International Journal of Information Management ,71, 102642.
9.    Fierro, Constanza, et al (2024). “Defining Knowledge: Bridging Epistemology and Large Language Models”, Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pp 16096–16111.
10.  Floridi , Luciano & Chiriatti, Massimo (2020). “GPT‑3: Its Nature, Scope, Limits, and Consequences”, Minds and Machines, 30:681–694.
11.  Hadi, Muhammad Usman, et al (2023). “Large Language Models: A Comprehensive Survey of its Applications, Challenges, Limitations, and Future Prospects”, DOI:10.36227/techrxiv.23589741.v1.
12. https://blog.google/products/search/search-language-understanding-bert/.
13.  https://linguistics.washington.edu/people/emily-m-bender.
14. https://www.nytimes.com/2018/11/18/technology/artificial-intelligence-language.html
15.  LeCun, Yann (2022). “A path towards autonomous machine intelligence version 0.9. 2, 2022-06-27”, Open Review, vol. 62.
16. McCarthy, John W. (1959). “Programs with common sense. In Proc. Tedding Conf”, on the Mechanization of Thought Processes, pp 75–91.
17.  Ohsugi, Yasuhito, et al (2019). A simple but effective method to incorporate multi-turn con text with BERT for conversational machine compre hension. In Proceedings of the First Workshop on NLP for Conversational AI, pp 11–17, Florence, Italy. Association for Computational Linguistics.
18.  Rae, Jack W., et al (2021). “Scaling language models: Methods, analysis & insights from training Gopher”. arXiv preprint arXiv:2112.11446, 2021.
19.  Saussure, Ferdinand de (1959). Course in General Linguistics. The Philosophical Society, Translated by Wade Baskin, New York.
20.  Searle, John (1980). “Minds, brains, and programs”. Be havioral and Brain Sciences, 3(3): 417–457.
21.  Shanahan, Murray (2022). “Talking about large language models”, CoRR, vol. abs/2212.03551
22.  Stevan Harnad (1990). “The symbol grounding problem”. Physica D, 42:335–346.
23.  Touvron, Hugo, et al (2023).“Llama: Open and efficient foundation, language models”, CoRR, arXiv:2302.13971v1 [cs.CL] 27 Feb.