人工智能的内在表征何以可能
|
摘要
塞尔等人认为,符号的语义派生于使用者的赋予和解释,而非内在于事物本身,没有任何对象凭借自身就是符号;这一批评带来了人工智能哲学的"符号接地问题"。计算主义认为外显符号的语义派生性并不意味着一切符号都是语义外在的。那么,非派生、固有语义的符号或者说内在表征何以可能?通过分析皮尔士等人的表征元理论可以发现,表征本质上呈现为目的论的多元功能系统。我们尝试藉以表征系统的多元结构对内在表征与外在表征的机制给以分析,得出内在表征在机制上依赖于自组织的内在目的,而外在表征则依赖于外在目的;人工智能若要将内在表征内置其中,需建构出基于内在目的性的自组织计算系统。
Searle and some others believe that symbols' meanings are derived from user's assigning and interpretation, but not intrinsic to itself, and no object is a symbol by virtue of its physics. This criticism has brought about the "symbol grounding problem" in the philosophy of artificial intelligence. Computationalism believe that explicit symbols' semantic contents are derived does not mean all the other kinds of symbols are also extrinsic. So, how is it possible that non-derivative, intrinsic representations are possible? By analyzing Peirce and some others' meta-theories of representation, we can find representation is essentially a multifunctional system of teleology. We try to analyze the mechanisms and forms of intrinsic representation and extrinsic representation by means of the structure of the system, benefit from which it can be drawn that intrinsic representations depend on self-organization, and extrinsic representations depend on the extrinsic purpose. It is necessary for intrinsic representational AI to construct a self-organizing computing system with intrinsic purpose.
Searle and some others believe that symbols' meanings are derived from user's assigning and interpretation, but not intrinsic to itself, and no object is a symbol by virtue of its physics. This criticism has brought about the "symbol grounding problem" in the philosophy of artificial intelligence. Computationalism believe that explicit symbols' semantic contents are derived does not mean all the other kinds of symbols are also extrinsic. So, how is it possible that non-derivative, intrinsic representations are possible? By analyzing Peirce and some others' meta-theories of representation, we can find representation is essentially a multifunctional system of teleology. We try to analyze the mechanisms and forms of intrinsic representation and extrinsic representation by means of the structure of the system, benefit from which it can be drawn that intrinsic representations depend on self-organization, and extrinsic representations depend on the extrinsic purpose. It is necessary for intrinsic representational AI to construct a self-organizing computing system with intrinsic purpose.
引文
[1]保罗·萨伽德.认知科学导论[M].朱菁、陈梦雅译,上海:上海辞书出版社,1995,11.
[2]Horst,S.'Symbols and Computation A Critique of the Computational Theory of Mind'[J].Minds&Machines,1999,9(3):347-381.
[3]Hinton,G.E.,McClelland,J.L.,Rumelhart,D.E.'Distributed Representations'[A],Rumelhart,D.,McClelland,J.(Eds)The PDP Researcch Group,Parallel Distributed Processing[C],Cambridge,MA:MIT Press,1986,77-109.
[4]Chalmers,D.J.'Subsymbolic Computation and the Chinese Room'[A],Dinsmore,J.(Ed)The Symbolic and Connectionist Paradigms:Closing the Gap[C],Lawrence Erlbaum,1992,25-48.
[5]Robert,C.Meaning and Mental Representation[M].Cambridge,MA:MIT Press,1991,13.
[6]Tom,Z.'Rethinking Grounding'[A],Riegler,A.,Peschl,M.,von Steinin,A.(Eds)Understanding Representation in the Cognitive Sciences[C],Boston,MA:Springer US,1999,177-90.
[7]Haugeland,J.Artificial Intelligence:The Very Idea[M].MIT Press,1985,106.
[8]Searle,J.R.'Who Is Computing with the Brain?'Author's Response'[J].Behavioral and Brain Sciences,1990,13(4):635-637.
[9]Harnad,S.'The Symbol Grounding Problem'[J].Physica D,1990,(42):335-346.
[10]Jerry,F.Representations:Philosophical Essays on the Foundations of Cognitive Science[M].The Harvester Press,1981,31.
[11]Peirce,C.S.'Division of Signs'[A],Hartshorne,C.,Weiss,P.(Eds)The Collected Papers of Charles Sanders Peirece[C],Harvard University Press,1956,228-253.
[12]Ronald,G.'How Models Are Used to Represent Reality'[J].Philosophy of Science,2004,(12):742-752.
[13]Van Fraassen.Scientific Representation:Paradoxes of Perspective[M].Oxford University Press,2008,8.
[14]Dretske,F.Naturalizing the Mind[M].The MIT Press,1997,2.
[15]Dretske,F.Explaining Behavior:Reasons in a World of Causes[M].The MIT Press,1988,61.
[16]Gualtiero,P.'Computational Explanation and Mechanistic Explanation of Mind'[A],Marraffa,M.,De Caro,M.,Ferretti,F.(Eds)Cartographies of the Mind[C],New York,Springer,2007,24-25.
[17]Kamppinen,M.Consciousness,Cognitive Schemata,and Relativism[M].New York,Springer,1993,14.
[18]Ramsey,W.M.Representation Reconsidered[M].New York,Cambridge University Press,2007,3.
[19]亚里士多德.物理学[M].张竹明译,北京:商务印书馆,2004,60.
[20]康德.判断力批判[M].邓晓芒译,北京:人民出版社,2002,210.
[21]Haken,H.'Self-organization,Scholarpedia'[EB/OL].http://www.scholarpedia.org/article/Self-organization.2008-6-23.
[22]钱学森.系统工程与系统科学的体系[M].北京:中国宇航出版社,2011,129.
[23]艾什比.大脑设[M].乐秀成译,北京:商务印书馆,1991,66-79.
[24]贝塔朗菲.一般系统论:基础、发展、应用[M].秋同、袁嘉新译,北京:社会科学文献出版社,1987,37.
[25]桂起权.目的论的科学-哲学分析[M].北京:人民出版社,2010,33.
[26]安东尼奥·达马西奥.感受发生的一切[M].教育科学出版社,2007,123-132.
[27]Paul,S.'On the Proper Treatment of Connectionism'[J].Behavioral and Brain Sciences,1988,11(1):14-15.
[28]高新民.联结主义的意向性缺失难题及其化解[J].自然辩证法通讯,2018,40(8):119-124.
[29]村田智、黑河治久.自组织机器人[M].李原译,北京:北京理工大学出版社,2015.
[2]Horst,S.'Symbols and Computation A Critique of the Computational Theory of Mind'[J].Minds&Machines,1999,9(3):347-381.
[3]Hinton,G.E.,McClelland,J.L.,Rumelhart,D.E.'Distributed Representations'[A],Rumelhart,D.,McClelland,J.(Eds)The PDP Researcch Group,Parallel Distributed Processing[C],Cambridge,MA:MIT Press,1986,77-109.
[4]Chalmers,D.J.'Subsymbolic Computation and the Chinese Room'[A],Dinsmore,J.(Ed)The Symbolic and Connectionist Paradigms:Closing the Gap[C],Lawrence Erlbaum,1992,25-48.
[5]Robert,C.Meaning and Mental Representation[M].Cambridge,MA:MIT Press,1991,13.
[6]Tom,Z.'Rethinking Grounding'[A],Riegler,A.,Peschl,M.,von Steinin,A.(Eds)Understanding Representation in the Cognitive Sciences[C],Boston,MA:Springer US,1999,177-90.
[7]Haugeland,J.Artificial Intelligence:The Very Idea[M].MIT Press,1985,106.
[8]Searle,J.R.'Who Is Computing with the Brain?'Author's Response'[J].Behavioral and Brain Sciences,1990,13(4):635-637.
[9]Harnad,S.'The Symbol Grounding Problem'[J].Physica D,1990,(42):335-346.
[10]Jerry,F.Representations:Philosophical Essays on the Foundations of Cognitive Science[M].The Harvester Press,1981,31.
[11]Peirce,C.S.'Division of Signs'[A],Hartshorne,C.,Weiss,P.(Eds)The Collected Papers of Charles Sanders Peirece[C],Harvard University Press,1956,228-253.
[12]Ronald,G.'How Models Are Used to Represent Reality'[J].Philosophy of Science,2004,(12):742-752.
[13]Van Fraassen.Scientific Representation:Paradoxes of Perspective[M].Oxford University Press,2008,8.
[14]Dretske,F.Naturalizing the Mind[M].The MIT Press,1997,2.
[15]Dretske,F.Explaining Behavior:Reasons in a World of Causes[M].The MIT Press,1988,61.
[16]Gualtiero,P.'Computational Explanation and Mechanistic Explanation of Mind'[A],Marraffa,M.,De Caro,M.,Ferretti,F.(Eds)Cartographies of the Mind[C],New York,Springer,2007,24-25.
[17]Kamppinen,M.Consciousness,Cognitive Schemata,and Relativism[M].New York,Springer,1993,14.
[18]Ramsey,W.M.Representation Reconsidered[M].New York,Cambridge University Press,2007,3.
[19]亚里士多德.物理学[M].张竹明译,北京:商务印书馆,2004,60.
[20]康德.判断力批判[M].邓晓芒译,北京:人民出版社,2002,210.
[21]Haken,H.'Self-organization,Scholarpedia'[EB/OL].http://www.scholarpedia.org/article/Self-organization.2008-6-23.
[22]钱学森.系统工程与系统科学的体系[M].北京:中国宇航出版社,2011,129.
[23]艾什比.大脑设[M].乐秀成译,北京:商务印书馆,1991,66-79.
[24]贝塔朗菲.一般系统论:基础、发展、应用[M].秋同、袁嘉新译,北京:社会科学文献出版社,1987,37.
[25]桂起权.目的论的科学-哲学分析[M].北京:人民出版社,2010,33.
[26]安东尼奥·达马西奥.感受发生的一切[M].教育科学出版社,2007,123-132.
[27]Paul,S.'On the Proper Treatment of Connectionism'[J].Behavioral and Brain Sciences,1988,11(1):14-15.
[28]高新民.联结主义的意向性缺失难题及其化解[J].自然辩证法通讯,2018,40(8):119-124.
[29]村田智、黑河治久.自组织机器人[M].李原译,北京:北京理工大学出版社,2015.