出刊年月/Date of Publishing
2020.06
所屬卷期/Vol. & No. 第50卷第2期 Vol. 50, No. 2
類型/Type 研究論文 Research Article
出刊年月/Date of Publishing
2020.06
所屬卷期/Vol. & No. 第50卷第2期 Vol. 50, No. 2
類型/Type 研究論文 Research Article
篇名/Title
戴維森論圖靈測試
Davidson on the Turing Test
作者/Author
趙之振 Chi-Chun Chiu
頁碼/Pagination
pp. 147-178
摘要
圖靈提出以模仿遊戲來測試機器能否思想。戴維森論證這測試是無法決定受測對象之語意學,故而無法決定機器能否思想;但庫辛思基卻反駁戴維森的論證。針對此中爭論,本文首先勾勒並釐清戴維森對模仿遊戲所作之評述,然後展示他對圖靈測試之批評;其次是論證庫辛思基批評戴維森之理據或是不成立,或是奠基於其對戴維森哲學的誤解之上;最後則是展示戴維森有關思想之條件的觀點與AI的關係:對於我們可以如何修訂圖靈測試,以此判定受測對象是否具有思想,戴維森的看法可以給我們怎樣的啟發。
Abstract
Davidson argues that the well-known Turing Test, being a method of determining whether a machine can think, fails to tell us anything about the semantics of the tested object and thus is inadequate to discover whether can think or not. However, against Davidson, Kuczynski claims that his reasoning is entirely fallacious and has little force in attacking the Turing Test. In this paper I will first delineate and clarify Davidson’s comments on Turing’s imitation game and his reasons for reject it as a proper test of machine thinking. Second, I will object to Kuczynski’s criticisms by showing that his arguments are either ill-founded, inconclusive, or based on his misinterpretation of Davidson’s thought. Finally, I will show how Davidson proposes to modify the Test in accordance with his own theory of interpretation, and the significance of this modified version for the attribution of thought to AI.
關鍵字/Key Word
戴維森、圖靈測試、思想歸屬、徹底翻譯、人工智慧
Davidson, Turing Test, thought attribution, radical interpretation, artificial intelligence
DOI
https://doi.org/10.7015/JEAS.202006_50(2)0002
學門分類/Subject
哲學 Philosophy