论文部分内容阅读
[摘要]
葛梯尔问题是如何解决知识论上存在的悖论的问题,并非仅仅是寻找第四条件的问题;葛梯尔问题的实质在于可错论的合理证明与运气的结合,而非仅仅是推理的前提有误,更非语词的误用。可错论的合理证明始终会为信念之碰巧为真保留一定的可能性,运气使这种可能性成为现实;可错论的合理证明原则不可废,否则,怀疑论将不可避免,因此,消除运气的影响就成了解决葛梯尔问题的关键。
[关键词] 可错论; 合理证明; 运气; 葛梯尔问题; 敏感性; 安全性; 知识论
Fallibilist Justification and Veritic Epistemic Luck: The Essence
of Gettier Problem and Misunderstandings in It
Wen Xueping
(School of Marxism, Southwest University of Political Science & Law, Chongqing 401120, China)
Abstract:
More than half a century ago, the American philosopher Edmund Gettier proposed two classic counterexamples against the traditional tripartite analysis of knowledge. These counterexamples have brought a great problem for the definition of knowledge, which is called ″the Gettier problem.″ There has been endless debate among epistemologists about how to understand and solve it and a lot of misunderstandings of that problem have thus arisen.
Firstly, the Gettier problem has been narrowly interpreted by some epistemologists as the problem of searching for a fourth condition that has to add to the justifiedtruebelief definition of propositional knowledge. In fact, the the Gettier problem aims to solve the paradox in epistemology which can be shown in the following three propositions: (1) Knowledge is justified true belief; (2) S has a justified true belief p; (3) S does not know that p. These three propositions can not be true at the same time because they are inconsistent. To solve this paradox, you can partly or completely deny any one or two of those three propositions. Searching for the fourth condition is but one of the many proposed approaches to the Gettier problem.
Secondly, some epistemologists think that the Gettier problem results from the false premise of deduction, which directly leads to the nofalse premise solution (NFPS) to the problem. NFPS can either be strict or moderate. As a strong advocate of strict NFPS, the Chinese epistemologist Chen Jiaming interprets epistemic justification as ″verification,″ and then views Gettiers counterexample of the tencoin case not as an epistemic case but as the case of misusage of words. In fact, the strict NFPS is a disguised form of infallibilism which directly results in skepticism while the proponents of the moderate NFPS fail to make clear the meaning of ″essentially″ in the condition that ″Ss justification for p does not essentially depend on any false premise.″ The failure of NFPS arises from the misconception about the fundamental structure of the Gettier problem, which is a combination of fallibilist justification and veritic epistemic luck. Fallibilist justification may cause the justifiers of proposition to separate from the facts asserted by the proposition. That is to say, the justifiers of proposition are inconsistent with its truthmaker, which will usually induce a falsehood. But the good luck called ″veritic epistemic luck″ makes you magically avoid the falsehood and get the truth. In order to solve the Gettier problem, we should eliminate either the inconsistency between justifier and truthmaker or the effect of veritic epistemic luck. But you cannot eliminate the inconsistency completely; otherwise you will get into infallibilism. So it is a preponderant solution to the Gettier problem to eliminate the effect of veritic epistemic luck by means of the condition of ″safety.″ Finally, some epistemologists think that knowledge can be lucky, which means that the epistemic agents in Gettiers counterexamples actually know the target propositions instead of not knowing. The claim is based on the idea that knowledge varies in degree. Lucky knowledge is at the edge of ″knowing,″ but it is still ″knowing″ instead of ″not knowing.″ This argument hardly holds water because it begs the question. It is the case that knowledge is incompatible with veritic epistemic luck, but it is not the case that knowledge is incompatible with all kinds of luck. Furthermore, it is our universal intuition and daytoday practice that knowledge excludes veritic epistemic luck.
Key words: fallibilism; justification; luck; the Gettier Problem; sensitivity; safety; epistemology
葛梯尔1963年发表了经典论文《有合理证明的真信念就是知识吗》见E.Gettier, ″Is Justified True Belief Knowledge?″ Analysis, Vol.23, No.6(1963), pp.12123。“Justified”一词国内主要有三种译法,即“证实了的”(胡军译)、“有确证的”(陈嘉明译)和“得到辩护的”(陈真译);“justification”也就相应地译为“证实”、“确证”、“辩护”。“证实”和“确证”的译法看似很专业,实则欠妥。其原因有三个:(1)认知上的“justification”表明的是知识的评价性要素、价值要素,而非仅仅是描述性要素或事实要素,而“确证”或“证实”不容易被理解为价值评价,因为“命题p已经被证实了还是没有被证实(已经被确证了还是没有被确证)”,这更像是一个事实描述的问题,而非价值评价的问题。(2)认知上的“justification”有“可错论”(fallibilism)和“不可错论”(infallibilism)之分,而“确证”或“证实”容易让人想到“justification”一定是不可错的,“命题p是被证实了的(或确证了的),但p有可能是错的”,这话听起来是自相矛盾的,其原因就在于人们习惯于自觉地将“证实”或“确证”作不可错论的理解:p既然是被证实了的(或确证了的),那么它就不可能是错的。“命题p是得到合理证明的,但它仍然有可能是错的”,这话听起来并不自相矛盾,原因在于“合理证明”与可错论是可以相容的,当然也可以对“合理证明”作不可错论的理解。(3)认知上的“justification”是要表明信念的合理性,而非行为的合理性,人们习惯于将“辩护”理解成为行为提供合理性证明,而且“辩护”也不大容易被理解为知识/知道所具有的价值因素。我国台湾地区学者习惯于将“justification”译为“证成”,这同样患有(1)、(2)两项毛病。因此,笔者将“justification”一词译为“合理证明”;与此相对应,“justified”译为“有合理证明的”或“得到合理证明的”,“justify”译为“证明……是合理的”,“justifier”译为“理据”,即理由或证据。这种译法似乎有些笨拙,但有助于避免错误的理解,而且比较生活化,有助于知识论研究的成果被大众所接受。,提出了针对传统知识概念分析的两个反例,这立即引起了西方哲学界经久不衰的研究兴趣,相关的论著可谓汗牛充栋,却又争议不断对于葛梯尔反例所引发的知识论问题,国内学者虽有引介和探讨,但不够细致,更未形成百家争鸣的局面。。之所以如此,是因为葛梯尔揭示了人类认知可能面临的普遍困境。
一、 何谓“葛梯尔问题”
日常意义上的“知识”(knowledge)一词至少有两层意思:一是指作为过程的内在状态,即“知道”(knowing)的状态;二是指认识活动的真理性成果。“知道”的状态必定蕴含着真理性的认知成果;获得真理性认知成果必然蕴含着“知道”的状态。因此,哲学家们对“知识”概念的分析通常就是对“知道”这一意向状态的分析。
知识概念的传统定义是“有合理证明的真信念”(justified true belief)“Belief”一词与“knowledge”一样,有相同类型的过程/结果歧义,既可指一种作为过程的内在的意向状态,即“相信”的状态,也可指作为“相信”之结果的状态,即“信念”。这两层含义也是相互蕴含的,因而经常相互替换。,即S知道p,当且仅当:
(1)p是真的;
(2)S相信p;
(3)S的信念p是有合理证明的。参见E.Gettier, ″Is Justified True Belief Knowledge?″ Analysis, Vol.23, No.6(1963), p.121。S代表认知主体,p代表命题内容。
这种理解通常被简称为“JTB理论”或“三要素分析”,其来源可追溯到柏拉图的对话参见《美诺篇》(Meno)97e98a(《柏拉图全集》第1卷,王晓朝译,人民出版社2002年版,第532533页)和《泰阿泰德篇》(Theaetetus)201c202d(《柏拉图全集》第2卷,王晓朝译,人民出版社2002年版,第737738页)。。 “JTB理论”并不要求各要素之间具有相互的蕴含关系,也就是说,信念并非都是真的,真的东西也不一定已成信念;信念不一定都是有合理证明的,有合理证明的东西也不一定都已成信念;有合理证明的东西不一定都是真的,真的东西不一定都得到合理证明。因此,我们可以用相互交错的三个圆圈(图1)来表示它们之间的关系。
图1关于知识的JTB理论
只有三者交汇的区域7才可能代表“知识”,其背后的基本逻辑是很符合人们的常识或直觉的:
(1)人们可能错误地相信p,但不可能错误地知道p,“错误地知道”(knowing falsely)是一个自相矛盾的说法,“知道”只能是真的知道或正确地知道。因此,知道p蕴含着p是真的,此即“真理要素”。
(2)如果你确实知道p是真的,但不相信p,这显得很荒谬,甚至不可能,因此知道p蕴含着相信p,此即“信念要素”。
(3)前面两个要素意味着知识至少是“真信念”,是否所有的“真信念”都是知识呢?直觉告诉我们,有些“真信念”并非知识,比如凭空瞎猜而碰巧获得的真信念、仅因痴心妄想而凑巧获得的真信念,即区域2所代表的东西。因此,我们还需要加上另一个要素,无论这个要素的具体内容是什么,其作用都是给“真信念”赋予“知识”的资格,或者说将“真信念”转化为“知识”,该要素通常的名称是“合理证明”(justification)。
因此,知识等于有合理证明的真信念。这曾是哲学家们难得的共识,但葛梯尔却令人惊奇地表明这个等式极有可能是错误的。
葛梯尔原初的两个反例在葛梯尔提出其著名的反例之前,罗素曾于1912年提出了“首相姓氏首字母”的例子(见[英]罗素《哲学问题》,何兆武译,商务印书馆2000年版,第109110页),他又于1948年提出了“停止的时钟”的例子(见[英]罗素《人类的知识:其范围与限度》,张金言译,商务印书馆1983年版,第120、191页)。这两个例子与葛梯尔提出的反例结构相同,但罗素给出这两个例子的目的与葛梯尔是不一样的,罗素的直接目的是要说明知识并不等于“真信念”,葛梯尔是要挑战“JTB理论”。在葛梯尔反例引起重大反响之前,哲学家们并没有太在意罗素的这两个十分富有独创性的例子。虽在细节上有所不同,但其基本预设和内在结构一样,两者之间的区别并没有多大的实质意义,故在此仅举出“十枚硬币”的反例。史密斯和琼斯在申请同一份工作,史密斯有很好的理由和证据相信命题q,即“琼斯是将要得到那份工作的人,并且他口袋里有十枚硬币”;史密斯相信q的理由是r,即“公司总裁向史密斯保证琼斯最终会被选中,并且史密斯十分钟前才数了琼斯口袋里的硬币”;史密斯从q正确地推导出并相信命题p,即“将得到那份工作的人口袋里有十枚硬币”。然而,实际上是“史密斯将得到那份工作,并且他口袋里也碰巧有十枚硬币”,对此史密斯本人并不知道;在这种情况下,“命题p仍然为真,史密斯也相信p,而且史密斯的信念p是有合理证明的”E.Gettier, ″Is Justified True Belief Knowledge?″ Analysis, Vol.23, No.6(1963), p.122。如果觉得葛梯尔提供的理由或证据不足以支撑命题q,还可以增加证据。。这样一来,就出现了一个悖论:
(1)知识=“有合理证明的真信念”(即知识=JTB);
(2)S拥有一个“有合理证明的真信念p”(即S拥有一个JTB);
(3)S却不知道p(S却没有知识)。
这三者似乎每一个都是正确的,但放在一起却构成了一个矛盾,因而不可能同时为真,至少有一个是错误的。究竟该如何解答这个悖论,这就是所谓的葛梯尔问题有学者将葛梯尔问题与罗素悖论做对比,并认为,“在很多关键的地方,葛梯尔关于知识的论文应该被视为与罗素悖论相似。葛梯尔对知识论所做的事情正是罗素悖论对数学之基础所做的事情”。这种看法是比较有道理的。参见J.Bigelow, ″Gettiers Theorem,″ in S.Hetherington(ed.), Aspects of Knowing: Epistemological Essays, Oxford: Elsevier Ltd., 2006, p.207。。解决这个悖论的通常策略可以概括为两类:一是捍卫传统分析,否认葛梯尔的(或葛梯尔式的)例子葛梯尔的原初例子虽只有两个,但类似的例子可以举出无限多。肖普(Robert Shope)在其《分析知道》一书中曾举出过学者们提出的98个葛梯尔式的反例。参见R.Shope, The Analysis of Knowing, New Jersey: Princeton University Press, 1983; J.S.Crumley Ⅱ, An Introduction to Epistemology, Antario: Broadview Press, 2009, p.63。构成了传统知识概念分析的反例,即否定前提(2)或结论(3);二是修正传统分析,承认葛梯尔的(或葛梯尔式的)例子构成了传统分析的反例,传统的知识概念分析需要修正,因为传统分析是不充分的,需要增加“第四个条件”,即增加一个条件以缩小图1中区域7的范围,或者传统的分析既不充分又不必要,应以外在主义的合理证明代替内在主义的合理证明,即部分地否定前提(1)。
有一种流行的观点认为:“对葛梯尔反例的普遍反应是承认‘S知道P’的分析必须加入第四个条件。寻求这第四个条件就成为众所周知的‘葛梯尔问题’。”[1]17即“通过增加知识的第四个条件等方法, 来达到‘修补篱笆’ 的目的”陈嘉明《专名、摹状词与葛梯尔问题》,载《世界哲学》2008年第6期,第13页。相似的看法还可参见陈真《盖梯尔问题的来龙去脉》,载《哲学研究》2005年第11期,第42页;胡军《关于知识定义的分析》,载《华中科技大学学报(社会科学版)》2008年第4期,第22页。。这种广为流传的观点显然是不正确的,已尝试过的大部分解答方案都不是寻求“第四条件”的问题。只有当我们承认“信念”、“真理”和“合理证明”这三个要件是知识的必要而非充分条件时,才存在寻求“第四个条件”的问题;以捍卫传统分析的方式解答葛梯尔问题,或者以外在主义的合理证明概念代替内在主义的合理证明概念的方式来解答葛梯尔问题,都与“第四个条件”无关。 [2] N.Rescher, Epistemology: An Introduction to the Theory of Knowledge, Albany: State University of New York Press, 2003.
[3] [美]路易斯·P. 波伊曼: 《知识论导论:我们能知道什么》,洪汉鼎译,北京:中国人民大学出版社,2008年。[L.P.Pojman, What Can We Know?An Introduction to the Theory of Knowledge, trans. by Hong Handing, Beijing: China Renmin University Press, 2008.]
[4] M.Clark, ″Knowledge and Grounds: A Comment on Mr. Gettiers Paper,″ Analysis, Vol.24, No.2(1963), pp.4648.
[5] J.S.Crumley Ⅱ, An Introduction to Epistemology, Toronto: Broadview Press, 2009.
[6] R.Meyers, The Likelihood of Knowledge, Dordrecht: Kluwer Academic Publishers, 1988.
[7] R.Feldman, Epistemology, New Jersey: Prentice Hall, 2003.
[8] 陈嘉明: 《葛梯尔问题与知识的条件(下)》,《哲学动态》2001年第1期,第4245页。[Chen Jiaming, ″Gettier Problem and the Conditions of Knowledge(Ⅱ),″ Philosophical Trends, No.1(2001), pp.4245.]
[9] 陈嘉明: 《知识论的“葛梯尔问题”及其解决方式》,《东南学术》2004年增刊,第161163页。[Chen Jiaming, ″Gettier Problem and Its Solution in Epistemology,″ Southeast Academic Research, S1(2004), pp.161163.]
[10] L.P.Pojman, What Can We Know: An Introduction to the Theory of Knowledge, California: Wadsworth Publishing Company, 1995.
[11] R.Feldman, ″An Alleged Defect in Gettier CounterExamples,″ Australasian Journal of Philosophy, Vol.52, No.1(1974), pp.6869.
[12] A.Goldman, ″Discrimination and Perceptual Knowledge,″ The Journal of Philosophy, Vol.73, No.20(1976), pp.771791.
[13] N.Lemos, An Introduction to the Theory of Knowledge, Cambridge: Cambridge University Press, 2007.
[14] E.Gettier, ″Is Justified True Belief Knowledge?″ Analysis, Vol.23, No.6(1963), pp.121123.
[15] L.Zagzebski, On Epistemology, Belmont, CA: Wadsworth, 2009.
[16] O.Johnson, ″The Standard Definition,″ Midwest Studies in Philosophy, Vol.5, No.1(1980), pp.113126.
[17] A.Heathcote, ″Truthmaking and the Gettier Problem,″ in S.Hetherington(ed.), Aspects of Knowing: Epistemological Essays, Oxford: Elsevier Ltd., 2006,pp.151168.
[18] D.Statman, ″Moral and Epistemic Luck,″ Ratio, Vol.4, No.2(1991), pp.146156.
[19] L.Zagzebski, Virtues of the Mind, Cambridge: Cambridge University Press, 1996.
[20] 陈嘉明: 《专名、摹状词与葛梯尔问题》,《世界哲学》2008年第6期,第1218页。[Chen Jiaming, ″Proper Name, Descriptions, and Gettier Problem,″ World Philosophy, No.6(2008), pp.1218.]
[21] D.Armstrong, Belief, Truth, and Knowledge, Cambridge: Cambridge University Press, 1973.
[22] K.Lehrer & T.Paxson, ″Knowledge: Undefeated Justified True Belief,″ The Journal of Philosophy, Vol.66, No.8(1969), pp.225237. [23] A.Goldman, ″A Causal Theory of Knowing,″ The Journal of Philosophy, Vol.64, No.12(1967) , pp.357372.
[24] F.Dretske, ″Conclusive Reasons,″ Australasian Journal of Philosophy, Vol.49, No.1(1971), pp.122.
[25] R.Shope, The Analysis of Knowing, New Jersey: Princeton University Press, 1983.
[26] S.Hetherington, ″The Gettier Problem,″ in S.Bernecker & D.Pritchard(eds.), The Routledge Companion to Epistemology, New York: Rutledge, 2011, pp.119130.
[27] S.Hetherington, ″Knowing Failably,″ The Journal of Philosophy, Vol.96, No.11(1999), pp.565587.
[28] D.Pritchard, Epistemic Luck, Oxford: Oxford University Press, 2005.
[29]D.Pritchard, ″Knowledge Cannot Be Lucky,″ in M.Steup & E.Sosa(eds.), Contemporary Debates in Epistemology, Oxford: WileyBlackwell, 2014, pp.154161.
[30]R.Nozick, Philosophical Explanations, Cambridge, MA: Harvard University Press, 1981.
[31] E.Sosa, ″How to Defeat Opposition to Moore,″ Nos, Vol.33, No.S13(1999), pp.141153.
葛梯尔问题是如何解决知识论上存在的悖论的问题,并非仅仅是寻找第四条件的问题;葛梯尔问题的实质在于可错论的合理证明与运气的结合,而非仅仅是推理的前提有误,更非语词的误用。可错论的合理证明始终会为信念之碰巧为真保留一定的可能性,运气使这种可能性成为现实;可错论的合理证明原则不可废,否则,怀疑论将不可避免,因此,消除运气的影响就成了解决葛梯尔问题的关键。
[关键词] 可错论; 合理证明; 运气; 葛梯尔问题; 敏感性; 安全性; 知识论
Fallibilist Justification and Veritic Epistemic Luck: The Essence
of Gettier Problem and Misunderstandings in It
Wen Xueping
(School of Marxism, Southwest University of Political Science & Law, Chongqing 401120, China)
Abstract:
More than half a century ago, the American philosopher Edmund Gettier proposed two classic counterexamples against the traditional tripartite analysis of knowledge. These counterexamples have brought a great problem for the definition of knowledge, which is called ″the Gettier problem.″ There has been endless debate among epistemologists about how to understand and solve it and a lot of misunderstandings of that problem have thus arisen.
Firstly, the Gettier problem has been narrowly interpreted by some epistemologists as the problem of searching for a fourth condition that has to add to the justifiedtruebelief definition of propositional knowledge. In fact, the the Gettier problem aims to solve the paradox in epistemology which can be shown in the following three propositions: (1) Knowledge is justified true belief; (2) S has a justified true belief p; (3) S does not know that p. These three propositions can not be true at the same time because they are inconsistent. To solve this paradox, you can partly or completely deny any one or two of those three propositions. Searching for the fourth condition is but one of the many proposed approaches to the Gettier problem.
Secondly, some epistemologists think that the Gettier problem results from the false premise of deduction, which directly leads to the nofalse premise solution (NFPS) to the problem. NFPS can either be strict or moderate. As a strong advocate of strict NFPS, the Chinese epistemologist Chen Jiaming interprets epistemic justification as ″verification,″ and then views Gettiers counterexample of the tencoin case not as an epistemic case but as the case of misusage of words. In fact, the strict NFPS is a disguised form of infallibilism which directly results in skepticism while the proponents of the moderate NFPS fail to make clear the meaning of ″essentially″ in the condition that ″Ss justification for p does not essentially depend on any false premise.″ The failure of NFPS arises from the misconception about the fundamental structure of the Gettier problem, which is a combination of fallibilist justification and veritic epistemic luck. Fallibilist justification may cause the justifiers of proposition to separate from the facts asserted by the proposition. That is to say, the justifiers of proposition are inconsistent with its truthmaker, which will usually induce a falsehood. But the good luck called ″veritic epistemic luck″ makes you magically avoid the falsehood and get the truth. In order to solve the Gettier problem, we should eliminate either the inconsistency between justifier and truthmaker or the effect of veritic epistemic luck. But you cannot eliminate the inconsistency completely; otherwise you will get into infallibilism. So it is a preponderant solution to the Gettier problem to eliminate the effect of veritic epistemic luck by means of the condition of ″safety.″ Finally, some epistemologists think that knowledge can be lucky, which means that the epistemic agents in Gettiers counterexamples actually know the target propositions instead of not knowing. The claim is based on the idea that knowledge varies in degree. Lucky knowledge is at the edge of ″knowing,″ but it is still ″knowing″ instead of ″not knowing.″ This argument hardly holds water because it begs the question. It is the case that knowledge is incompatible with veritic epistemic luck, but it is not the case that knowledge is incompatible with all kinds of luck. Furthermore, it is our universal intuition and daytoday practice that knowledge excludes veritic epistemic luck.
Key words: fallibilism; justification; luck; the Gettier Problem; sensitivity; safety; epistemology
葛梯尔1963年发表了经典论文《有合理证明的真信念就是知识吗》见E.Gettier, ″Is Justified True Belief Knowledge?″ Analysis, Vol.23, No.6(1963), pp.12123。“Justified”一词国内主要有三种译法,即“证实了的”(胡军译)、“有确证的”(陈嘉明译)和“得到辩护的”(陈真译);“justification”也就相应地译为“证实”、“确证”、“辩护”。“证实”和“确证”的译法看似很专业,实则欠妥。其原因有三个:(1)认知上的“justification”表明的是知识的评价性要素、价值要素,而非仅仅是描述性要素或事实要素,而“确证”或“证实”不容易被理解为价值评价,因为“命题p已经被证实了还是没有被证实(已经被确证了还是没有被确证)”,这更像是一个事实描述的问题,而非价值评价的问题。(2)认知上的“justification”有“可错论”(fallibilism)和“不可错论”(infallibilism)之分,而“确证”或“证实”容易让人想到“justification”一定是不可错的,“命题p是被证实了的(或确证了的),但p有可能是错的”,这话听起来是自相矛盾的,其原因就在于人们习惯于自觉地将“证实”或“确证”作不可错论的理解:p既然是被证实了的(或确证了的),那么它就不可能是错的。“命题p是得到合理证明的,但它仍然有可能是错的”,这话听起来并不自相矛盾,原因在于“合理证明”与可错论是可以相容的,当然也可以对“合理证明”作不可错论的理解。(3)认知上的“justification”是要表明信念的合理性,而非行为的合理性,人们习惯于将“辩护”理解成为行为提供合理性证明,而且“辩护”也不大容易被理解为知识/知道所具有的价值因素。我国台湾地区学者习惯于将“justification”译为“证成”,这同样患有(1)、(2)两项毛病。因此,笔者将“justification”一词译为“合理证明”;与此相对应,“justified”译为“有合理证明的”或“得到合理证明的”,“justify”译为“证明……是合理的”,“justifier”译为“理据”,即理由或证据。这种译法似乎有些笨拙,但有助于避免错误的理解,而且比较生活化,有助于知识论研究的成果被大众所接受。,提出了针对传统知识概念分析的两个反例,这立即引起了西方哲学界经久不衰的研究兴趣,相关的论著可谓汗牛充栋,却又争议不断对于葛梯尔反例所引发的知识论问题,国内学者虽有引介和探讨,但不够细致,更未形成百家争鸣的局面。。之所以如此,是因为葛梯尔揭示了人类认知可能面临的普遍困境。
一、 何谓“葛梯尔问题”
日常意义上的“知识”(knowledge)一词至少有两层意思:一是指作为过程的内在状态,即“知道”(knowing)的状态;二是指认识活动的真理性成果。“知道”的状态必定蕴含着真理性的认知成果;获得真理性认知成果必然蕴含着“知道”的状态。因此,哲学家们对“知识”概念的分析通常就是对“知道”这一意向状态的分析。
知识概念的传统定义是“有合理证明的真信念”(justified true belief)“Belief”一词与“knowledge”一样,有相同类型的过程/结果歧义,既可指一种作为过程的内在的意向状态,即“相信”的状态,也可指作为“相信”之结果的状态,即“信念”。这两层含义也是相互蕴含的,因而经常相互替换。,即S知道p,当且仅当:
(1)p是真的;
(2)S相信p;
(3)S的信念p是有合理证明的。参见E.Gettier, ″Is Justified True Belief Knowledge?″ Analysis, Vol.23, No.6(1963), p.121。S代表认知主体,p代表命题内容。
这种理解通常被简称为“JTB理论”或“三要素分析”,其来源可追溯到柏拉图的对话参见《美诺篇》(Meno)97e98a(《柏拉图全集》第1卷,王晓朝译,人民出版社2002年版,第532533页)和《泰阿泰德篇》(Theaetetus)201c202d(《柏拉图全集》第2卷,王晓朝译,人民出版社2002年版,第737738页)。。 “JTB理论”并不要求各要素之间具有相互的蕴含关系,也就是说,信念并非都是真的,真的东西也不一定已成信念;信念不一定都是有合理证明的,有合理证明的东西也不一定都已成信念;有合理证明的东西不一定都是真的,真的东西不一定都得到合理证明。因此,我们可以用相互交错的三个圆圈(图1)来表示它们之间的关系。
图1关于知识的JTB理论
只有三者交汇的区域7才可能代表“知识”,其背后的基本逻辑是很符合人们的常识或直觉的:
(1)人们可能错误地相信p,但不可能错误地知道p,“错误地知道”(knowing falsely)是一个自相矛盾的说法,“知道”只能是真的知道或正确地知道。因此,知道p蕴含着p是真的,此即“真理要素”。
(2)如果你确实知道p是真的,但不相信p,这显得很荒谬,甚至不可能,因此知道p蕴含着相信p,此即“信念要素”。
(3)前面两个要素意味着知识至少是“真信念”,是否所有的“真信念”都是知识呢?直觉告诉我们,有些“真信念”并非知识,比如凭空瞎猜而碰巧获得的真信念、仅因痴心妄想而凑巧获得的真信念,即区域2所代表的东西。因此,我们还需要加上另一个要素,无论这个要素的具体内容是什么,其作用都是给“真信念”赋予“知识”的资格,或者说将“真信念”转化为“知识”,该要素通常的名称是“合理证明”(justification)。
因此,知识等于有合理证明的真信念。这曾是哲学家们难得的共识,但葛梯尔却令人惊奇地表明这个等式极有可能是错误的。
葛梯尔原初的两个反例在葛梯尔提出其著名的反例之前,罗素曾于1912年提出了“首相姓氏首字母”的例子(见[英]罗素《哲学问题》,何兆武译,商务印书馆2000年版,第109110页),他又于1948年提出了“停止的时钟”的例子(见[英]罗素《人类的知识:其范围与限度》,张金言译,商务印书馆1983年版,第120、191页)。这两个例子与葛梯尔提出的反例结构相同,但罗素给出这两个例子的目的与葛梯尔是不一样的,罗素的直接目的是要说明知识并不等于“真信念”,葛梯尔是要挑战“JTB理论”。在葛梯尔反例引起重大反响之前,哲学家们并没有太在意罗素的这两个十分富有独创性的例子。虽在细节上有所不同,但其基本预设和内在结构一样,两者之间的区别并没有多大的实质意义,故在此仅举出“十枚硬币”的反例。史密斯和琼斯在申请同一份工作,史密斯有很好的理由和证据相信命题q,即“琼斯是将要得到那份工作的人,并且他口袋里有十枚硬币”;史密斯相信q的理由是r,即“公司总裁向史密斯保证琼斯最终会被选中,并且史密斯十分钟前才数了琼斯口袋里的硬币”;史密斯从q正确地推导出并相信命题p,即“将得到那份工作的人口袋里有十枚硬币”。然而,实际上是“史密斯将得到那份工作,并且他口袋里也碰巧有十枚硬币”,对此史密斯本人并不知道;在这种情况下,“命题p仍然为真,史密斯也相信p,而且史密斯的信念p是有合理证明的”E.Gettier, ″Is Justified True Belief Knowledge?″ Analysis, Vol.23, No.6(1963), p.122。如果觉得葛梯尔提供的理由或证据不足以支撑命题q,还可以增加证据。。这样一来,就出现了一个悖论:
(1)知识=“有合理证明的真信念”(即知识=JTB);
(2)S拥有一个“有合理证明的真信念p”(即S拥有一个JTB);
(3)S却不知道p(S却没有知识)。
这三者似乎每一个都是正确的,但放在一起却构成了一个矛盾,因而不可能同时为真,至少有一个是错误的。究竟该如何解答这个悖论,这就是所谓的葛梯尔问题有学者将葛梯尔问题与罗素悖论做对比,并认为,“在很多关键的地方,葛梯尔关于知识的论文应该被视为与罗素悖论相似。葛梯尔对知识论所做的事情正是罗素悖论对数学之基础所做的事情”。这种看法是比较有道理的。参见J.Bigelow, ″Gettiers Theorem,″ in S.Hetherington(ed.), Aspects of Knowing: Epistemological Essays, Oxford: Elsevier Ltd., 2006, p.207。。解决这个悖论的通常策略可以概括为两类:一是捍卫传统分析,否认葛梯尔的(或葛梯尔式的)例子葛梯尔的原初例子虽只有两个,但类似的例子可以举出无限多。肖普(Robert Shope)在其《分析知道》一书中曾举出过学者们提出的98个葛梯尔式的反例。参见R.Shope, The Analysis of Knowing, New Jersey: Princeton University Press, 1983; J.S.Crumley Ⅱ, An Introduction to Epistemology, Antario: Broadview Press, 2009, p.63。构成了传统知识概念分析的反例,即否定前提(2)或结论(3);二是修正传统分析,承认葛梯尔的(或葛梯尔式的)例子构成了传统分析的反例,传统的知识概念分析需要修正,因为传统分析是不充分的,需要增加“第四个条件”,即增加一个条件以缩小图1中区域7的范围,或者传统的分析既不充分又不必要,应以外在主义的合理证明代替内在主义的合理证明,即部分地否定前提(1)。
有一种流行的观点认为:“对葛梯尔反例的普遍反应是承认‘S知道P’的分析必须加入第四个条件。寻求这第四个条件就成为众所周知的‘葛梯尔问题’。”[1]17即“通过增加知识的第四个条件等方法, 来达到‘修补篱笆’ 的目的”陈嘉明《专名、摹状词与葛梯尔问题》,载《世界哲学》2008年第6期,第13页。相似的看法还可参见陈真《盖梯尔问题的来龙去脉》,载《哲学研究》2005年第11期,第42页;胡军《关于知识定义的分析》,载《华中科技大学学报(社会科学版)》2008年第4期,第22页。。这种广为流传的观点显然是不正确的,已尝试过的大部分解答方案都不是寻求“第四条件”的问题。只有当我们承认“信念”、“真理”和“合理证明”这三个要件是知识的必要而非充分条件时,才存在寻求“第四个条件”的问题;以捍卫传统分析的方式解答葛梯尔问题,或者以外在主义的合理证明概念代替内在主义的合理证明概念的方式来解答葛梯尔问题,都与“第四个条件”无关。 [2] N.Rescher, Epistemology: An Introduction to the Theory of Knowledge, Albany: State University of New York Press, 2003.
[3] [美]路易斯·P. 波伊曼: 《知识论导论:我们能知道什么》,洪汉鼎译,北京:中国人民大学出版社,2008年。[L.P.Pojman, What Can We Know?An Introduction to the Theory of Knowledge, trans. by Hong Handing, Beijing: China Renmin University Press, 2008.]
[4] M.Clark, ″Knowledge and Grounds: A Comment on Mr. Gettiers Paper,″ Analysis, Vol.24, No.2(1963), pp.4648.
[5] J.S.Crumley Ⅱ, An Introduction to Epistemology, Toronto: Broadview Press, 2009.
[6] R.Meyers, The Likelihood of Knowledge, Dordrecht: Kluwer Academic Publishers, 1988.
[7] R.Feldman, Epistemology, New Jersey: Prentice Hall, 2003.
[8] 陈嘉明: 《葛梯尔问题与知识的条件(下)》,《哲学动态》2001年第1期,第4245页。[Chen Jiaming, ″Gettier Problem and the Conditions of Knowledge(Ⅱ),″ Philosophical Trends, No.1(2001), pp.4245.]
[9] 陈嘉明: 《知识论的“葛梯尔问题”及其解决方式》,《东南学术》2004年增刊,第161163页。[Chen Jiaming, ″Gettier Problem and Its Solution in Epistemology,″ Southeast Academic Research, S1(2004), pp.161163.]
[10] L.P.Pojman, What Can We Know: An Introduction to the Theory of Knowledge, California: Wadsworth Publishing Company, 1995.
[11] R.Feldman, ″An Alleged Defect in Gettier CounterExamples,″ Australasian Journal of Philosophy, Vol.52, No.1(1974), pp.6869.
[12] A.Goldman, ″Discrimination and Perceptual Knowledge,″ The Journal of Philosophy, Vol.73, No.20(1976), pp.771791.
[13] N.Lemos, An Introduction to the Theory of Knowledge, Cambridge: Cambridge University Press, 2007.
[14] E.Gettier, ″Is Justified True Belief Knowledge?″ Analysis, Vol.23, No.6(1963), pp.121123.
[15] L.Zagzebski, On Epistemology, Belmont, CA: Wadsworth, 2009.
[16] O.Johnson, ″The Standard Definition,″ Midwest Studies in Philosophy, Vol.5, No.1(1980), pp.113126.
[17] A.Heathcote, ″Truthmaking and the Gettier Problem,″ in S.Hetherington(ed.), Aspects of Knowing: Epistemological Essays, Oxford: Elsevier Ltd., 2006,pp.151168.
[18] D.Statman, ″Moral and Epistemic Luck,″ Ratio, Vol.4, No.2(1991), pp.146156.
[19] L.Zagzebski, Virtues of the Mind, Cambridge: Cambridge University Press, 1996.
[20] 陈嘉明: 《专名、摹状词与葛梯尔问题》,《世界哲学》2008年第6期,第1218页。[Chen Jiaming, ″Proper Name, Descriptions, and Gettier Problem,″ World Philosophy, No.6(2008), pp.1218.]
[21] D.Armstrong, Belief, Truth, and Knowledge, Cambridge: Cambridge University Press, 1973.
[22] K.Lehrer & T.Paxson, ″Knowledge: Undefeated Justified True Belief,″ The Journal of Philosophy, Vol.66, No.8(1969), pp.225237. [23] A.Goldman, ″A Causal Theory of Knowing,″ The Journal of Philosophy, Vol.64, No.12(1967) , pp.357372.
[24] F.Dretske, ″Conclusive Reasons,″ Australasian Journal of Philosophy, Vol.49, No.1(1971), pp.122.
[25] R.Shope, The Analysis of Knowing, New Jersey: Princeton University Press, 1983.
[26] S.Hetherington, ″The Gettier Problem,″ in S.Bernecker & D.Pritchard(eds.), The Routledge Companion to Epistemology, New York: Rutledge, 2011, pp.119130.
[27] S.Hetherington, ″Knowing Failably,″ The Journal of Philosophy, Vol.96, No.11(1999), pp.565587.
[28] D.Pritchard, Epistemic Luck, Oxford: Oxford University Press, 2005.
[29]D.Pritchard, ″Knowledge Cannot Be Lucky,″ in M.Steup & E.Sosa(eds.), Contemporary Debates in Epistemology, Oxford: WileyBlackwell, 2014, pp.154161.
[30]R.Nozick, Philosophical Explanations, Cambridge, MA: Harvard University Press, 1981.
[31] E.Sosa, ″How to Defeat Opposition to Moore,″ Nos, Vol.33, No.S13(1999), pp.141153.