Language, Thought and Turing: Davidson's Views on the Limits of Machine Intelligence

Öz

This article examines the Turing Test, initially proposed by Alan Turing in 1950, to assess machine intelligence by its ability to mimic human behavior in linguistic communication. While Turing's Test has been foundational in artificial intelligence research, it has faced substantial philosophical scrutiny. The focus of this paper is on Donald Davidson’s critique, which argues that the Turing Test fails to account for the complexities of thought and propositional attitudes that are fundamentally social and communicative. The discussion includes an analysis of how some AI technologies, such as sophisticated language models, challenge and yet fail to overcome these philosophical objections. The paper concludes that while Turing's Test provides a pragmatic measure for machine "intelligence," it does not fully capture the nuanced capabilities of human thought as envisioned by Davidson, thereby necessitating broader criteria for evaluating machine thinking. The implications for future AI development and testing are considered, underscoring the need for an approach that integrates interactional and contextual dynamics in assessing machine cognition.

Anahtar Kelimeler

• Turing’s Test
• Propositional Attitudes
• Donald Davidson
• Human-Machine Interaction
• Triangulation

Dil, Düşünce ve Turing: Davidson'ın Makine Zekasının Sınırlarına İlişkin Görüşleri

Öz

Bu makale, ilk olarak 1950 yılında Alan Turing tarafından önerilen ve makine zekasını dilsel iletişimde insan davranışını taklit etme becerisine göre değerlendiren Turing Testi'ni incelemektedir. Turing Testi yapay zeka araştırmalarının temelini oluştururken, önemli felsefi itirazlarla da karşı karşıya kalmıştır. Bu makalenin odak noktası, Turing Testi'nin, temelde sosyal ve iletişimsel olan düşünce ve önermesel tutumların karmaşıklığını açıklamakta başarısız olduğunu savunan Donald Davidson'ın eleştirisidir. Tartışma, sofistike dil modelleri gibi bazı yapay zeka teknolojilerinin bu felsefi itirazlara ne şekilde yenik düştüğünün bir analizini içermektedir. Makalede, Turing Testi'nin makine "zekası" için pragmatik bir ölçüt sunmasına rağmen, Davidson tarafından öngörülen insan düşüncesinin nüanslı yeteneklerini tam olarak yakalayamadığı, dolayısıyla makine düşüncesini değerlendirmek için daha geniş kriterlere ihtiyaç duyulduğu sonucuna varılmakta ve makine bilişini değerlendirirken etkileşimsel ve bağlamsal dinamikleri göz önünde bulunduran bir yaklaşıma duyulan ihtiyacın altı çizilmektedir.

Anahtar Kelimeler

• Turing Testi
• Önermesel Tutumlar
• Donald Davidson
• İnsan-Makine Etkileşimi
• Üçgenleme

Kaynakça

1. Adamopoulou, E., & Moussiades, L. (2020). Chatbots: History, technology, and applications. Machine Learning with Applications. https://doi.org/10.1016/j.mlwa.2020.100006
2. Bubeck et al., S. (2023). Sparks of Artificial General Intelligence: Early experiments with GPT-4. digital preprint, arXiv:2303.12712
3. Colby, K. M., Weber, S., & Hif, F. D. (1971). Artificial Paranoia. Artificial Intelligence, 1-25.
4. Copeland, B. J. (2000). The Turing Test. Minds and Machines , 519-539. https://doi.org/10.1023/A:1011285919106
5. Crane, T. (2001). Elements of Mind: An Introduction to the Philosophy of Mind. Oxford: Oxford University Press.
6. Davidson, D. (1982). Rational Animals. Dialectica, 317-327. https://doi.org/10.1111/j.1746-8361.1982.tb01546.x
7. Davidson, D. (1989). The Conditions of Thought. Le Cahier (Collège international de philosophie), 165-171. https://doi.org/10.5840/gps19893627
8. Davidson, D. (1991). Three Varieties of Knowledge. Royal Institute of Philosophy Supplement, 153-166. https://doi.org/10.1017/s1358246100007748

9. Davidson, D. (1999). The Emergence of Thought. Erkenntnis, 51(1), 7-17. http://www.jstor.org/stable/20012936 Davidson, D. (2004). Problems of Rationality. New York: Oxford University Press.
10. Dennett, D. C. (2003). Can Machines Think? C. Teuscher (Ed.), Alan Turing: Life and Legacy of a Great Thinker (ss. 295-316). Heidelberg: Springer Berlin.
11. Floridi, L., & Chiriatti, M. (2020). GPT‐3: Its Nature, Scope, Limits, and Consequences. Minds and Machines, 681–694. https://doi.org/10.1007/s11023-020-09548-1
12. Harnad, S. (2003). Minds, Machines and Turing. H. J. Moor (ed.), The Turing Test: The Elusive Standard of Artificial Intelligence (ss. 253-273). Springer Dordrecht.
13. Jones, C. R., & Bergen, B. K. (2024). Does GPT-4 pass the Turing test? arXiv:2310.20216 .
14. Mahowald, K. I. (2023). Dissociating language and thought in large language models: a cognitive perspective. arXiv preprint arXiv:2301.06627.
15. Myers, R. H., & Verheggen, C. (2016). Donald Davidson’s Triangulation Argument: A Philosophical Inquiry. New York: Routledge.
16. Nachiappan, A. (2024, 04 18). Technology behind ChatGPT better with eye problem advice than non-specialist doctors, study test finds. https://news.sky.com/: https://news.sky.com/story/technology-behind-chatgpt-better-with-eye-problem-advice-than-non-specialist-doctors-study-test-finds-13117259
17. Natale, S. (2021). Deceitful Media: Artificial Intelligence and Social Life After the Turing Test. Oxford: Oxford University Press.
18. Proudfoot, D. (2020). Rethinking Turing’s Test and the Philosophical Implications. Minds and Machines, 487-512. https://doi.org/10.1007/s11023-020-09534-7
19. Sanderson, K. (2023, 12 20). This GPT-powered robot chemist designs reactions and makes drugs — on its own. nature.com: https://www.nature.com/articles/d41586-023-04073-4
20. Searle, J. (1980). Minds, Brains, and Programs. Behavioral and Brain Sciences, 417-457. https://doi.org/10.1017/S0140525X00005756
21. Turing, A. (1950). Computing Machinery and Intelligence. Mind, 443-460. https://doi.org/10.1093/mind/LIX.236.433
22. Wilford, J. N. (1968, Haziran 15). Computer Is Being Taught to Understand English. New York Times, p. 58.