Content Summary

The recent success of Large Language Models, which now reach human-like performance in responding to written or oral language input, raises a number of empirical, theoretical and philosophical questions which will be discussed in the seminar. We will start with classical papers on language, consciousness and AI, and will then focus on current discussions about relevant questions regarding the language skills of AI systems, including meaning and semantic grounding. Second, we will discuss methodological problems arising if human-like cognitive abilities, such as mind-reading or even consciousness, are being ascribed to these systems. Third, we will discuss problems coming up when AI systems are used as models for the human mind. A major issue to discuss is whether the remarkable abilities of tools such as ChatGPT still show specific limitations compared with human language skills and whether there are any principal reasons to adopt or reject them as putative models of the human cognitive mind and brain.

You can find the pdf version here.

Papers

Bender, E. M., & Koller, A. (2020). Climbing towards NLU: On meaning, form, and understanding in the age of data. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 5185–5198. http://dx.doi.org/10.18653/v1/2020.acl-main.463

Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big?? Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 610–623. https://doi.org/10.1145/3442188.3445922

Biever, C. (2023). ChatGPT broke the Turing test—the race is on for new ways to assess AI. Nature, 619(7971), 686–689. http://dx.doi.org/10.1038/d41586-023-02361-7

Butlin, P. (2023). Sharing our concepts with machines. Erkenntnis, 88(7), 3079–3095. http://dx.doi.org/10.1007/s10670-021-00491-w

Coelho Mollo, D. C., & Millière, R. (2023). The vector grounding problem. arXiv preprint arXiv:2304.01481. https://arxiv.org/abs/2304.01481 Elman, J. L. (1990).

Finding structure in time. Cognitive Science, 14, 179–211. http://dx.doi.org/10.1016/0364-0213(90)90002-e Harnad, S. (1990). The symbol grounding problem. Physica D: Nonlinear Phenomena, 42(1–3), 335– 346. http://dx.doi.org/10.1016/0167-2789(90)90087-6

Harnad, S. (2024). Language writ large: LLMs, ChatGPT, grounding, meaning and understanding. arXiv preprint arXiv:2402.02243. http://dx.doi.org/10.31234/osf.io/ch2wx

Moro, A., Greco, M., & Cappa, S. F. (2024). Large languages, impossible languages and human brains. Cortex, in press. http://dx.doi.org/10.1016/j.cortex.2023.07.003

Piantadosi, S. T. (2023). Modern language models refute Chomsky’s approach to language. In From fieldwork to linguistic theory: A tribute to Dan Everett (pp. 353–414). https://langscipress.org/catalog/book/434

Pulvermüller, F. (2023). Neurobiological mechanisms for language, symbols and concepts: Clues from brain-constrained deep neural networks. Progress in Neurobiology, 102511. https://doi.org/10.1016/j.pneurobio.2023.102511

Søgaard, A. (2023). Grounding the vector space of an octopus: Word meaning from raw text. Minds and Machines, 33(1), 33–54. http://dx.doi.org/10.1007/s11023-023-09622-4

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30. https://arxiv.org/abs/1706.03762