【美】奎因 著,翟玉章 译
转自翟玉章老师博客
http://blog.sina.com.cn/s/blog_4c6a7dda0100jrf7.html
鸣谢
26. One answer which readily suggests itself is this: any meaningful statement should at least have some statements of the basic language as consequences, or else be a consequence of some statements of the basic language.
一个很容易想到的现成的答案是这样的:任何有意义的陈述都至少蕴含一些基本语言中的陈述,或者为一些基本语言中的陈述所蕴含。
But on closer scrutiny even this requirement begins to appear too strong. I won’t here go into the technical reasons why; but merely say that the closer scrutiny in question is considerably aided by, if not dependent upon, methods of modern logic.
但是更仔细的考察表明,即使这样一个要求也还是太强的要求了。我不打算在这儿细说技术上的理由;我只是指出,现代逻辑方法的运用即使不是这种更仔细的考察的全部,也对它有相当的助益。
27. The most that can in general be demanded by way of connection is perhaps this: there must be some degree of confirmation—some degree of probability, in a sense—relating any meaningful statements to some statements of the basic language.
总的说来,对这种联系的要求最多可能只是这样:有意义的陈述相对于基本语言中的某些陈述能获得一定程度的确认,或(在某种意义上的)一定程度的可然性。
28. Then comes the problem of formulating what is meant by degree of confirmation. This is a vast problem, but, according to the considerations I have just sketched, it would seem to be an essential problem for empiricism; for, a degree of confirmation seems to be the one surviving connection between sciences, or common sense, and the direct experience which is the empiricist’s bedrock.
接下来是如何界定确认度的含义的问题。这是一个非常大的问题,而且根据我刚刚概略地谈到的那些考虑,这似乎也是经验主义的核心问题;因为,一定程度的确认似乎正是残存下来的充当科学或常识与作为经验主义基础的直接经验之间的联系的东西。
29. Carnap, Hempel, Helmer, Oppenheim, and others have been hard at word trying to formulate degree of confirmation, availing themselves liberally of the resources of modern logic. I for one am far from content with the results to date, but this much is clear: if a satisfactory formulation of confirmation is in the making, modern logic is playing the indispensible part in its creation; and if we are to decide rather, in time, that such an attempt is doomed to failure, any conclusive arguments to this negative effect will probably have to exploit the resources of modern logic in subtle ways that none of us have yet envisaged.
卡尔纳普、亨普尔、赫默尔、奥本海默,还有其他人正在致力于制定一个关于确认度的定义,他们都充分利用着现代逻辑的资源。我虽然也是他们中的一分子,但对迄今为止的结果并不感到满意。不过有一点是清楚的:如果一个令人满意的定义正在形成,那么现代逻辑一定在这当中发挥着不可或缺的作用;反过来,如果到头来我们不得不承认这样的努力终不免于失败,那么引向这个否定结果的论证也同样离不开对现代逻辑资源的利用,当然利用方式之微妙是我们迄今没有想到过的。【奎因在这篇文章中表现出来的对基本语言的怀疑,对令人满意的确认度概念的怀疑,在这以后进一步发酵了,并最终发展出了他自己独特的整体主义的科学观,把经验主义推向了一个更加首尾一贯的高度。】
30. Logic figures no less importantly, and perhaps yet more obviously, in another central problem of philosophy which is closely connected with that of empiricist reduction or confirmation; namely, the problem of the nature of a priori, or analytic, statements.
逻辑在另一个哲学中心问题中的重要性也并不稍逊,甚至还更加明显。这个问题就是先天陈述或分析陈述的本性问题,它是与前面关于经验主义还原或确认问题紧密联系在一起的。
31. When we make explicit how to reduce statements to basic ones about experience, or how to confirm them, we thereby implicitly determine the meanings of those statements and their constituent words—insofar as they may from an empiricistic point of view be said to have meanings at all. Empiricists and pragmatists from Peirce through Wittgenstein to Professors Lewis and Bridgman have agreed, in varying idioms, that the meaning of a statement consists in the methods of its empirical confirmation. Now as by-products of these methods there will be some statements, notably Stevenson’s old stand-by ‘No spinster is married’, which are automatically true by virtue of the meanings alone—by virtue, in other words, of the mere methods of empirical confirmation themselves, no matter what the content of experience may happen to be. Such statements are, in Kant’s terminology, analytic: true by virtue of the meanings alone.
当我们弄清陈述如何可以还原为关于经验的基本陈述,或陈述如何被关于经验的基本陈述所确认时,我们也就同时确定了这些陈述及其组成词汇的意义—就它们从经验主义的观点看所能具有的意义。从皮尔士,经过维特根斯坦,再到刘易斯和布里奇曼的经验主义者和实用主义者以各种不同的措辞认同:陈述的意义在于其被经验确认的方式。作为这些方法的一个后果,便会有一些仅仅根据意义—即仅仅根据经验确认的方法本身,而不管经验的内容碰巧会是什么—就自动成真的陈述,斯蒂文森爱挂在嘴边的“没有一个老处女是结婚的”就是一个显而易见的例子。这些陈述,用康德的术语讲,是分析的:仅仅根据意义而为真。
32. All those statements which are theorems of pure logic, or instances of theorems of pure logic, are generally viewed as falling into the category of analytic statements; so it is not surprising that a training in logic should contribute to an understanding of analytic statements, if anything will.
纯逻辑里的所有定理,或所有定理的实例,通常被看成是分析陈述的一部分;所以如果有什么途径来增进我们对分析陈述的理解,那么逻辑学的训练肯定要算一个,这一点是毫不足怪的。
33. In fact the development of modern logic has had an important bearing on Kant’s doctrine of a priori synthetic judgments: the doctrine that not all statements which are true independently of experience are analytic. Kant felt that arithmetic was largely a priori synthetic in this sense, as contrasted with logic, which is analytic. But Frege, in 1879—1888, showed that arithmetic was derivable from logic alone.
事实上,现代逻辑的发展和康德的先天综合判断的学说大有关系。康德主张并不是所有独立于经验的陈述都是分析的;算术大体上在这个意义上是先天综合的,与之对比的是,逻辑却是分析的。但是,弗雷格却表明算术可以仅仅由逻辑而推导出来(1879—1888)。
34. Kant might still maintain that logic in Frege’s sense, being stronger than the traditional logic, goes beyond the bounds of the analytic. If so, however, it would be up to Kant to show just where Frege, in his logic, crossed the line. In any case the problem delineating the bounds of the analytic becomes one which can hardly be profitably discussed any longer without reference to technicalities of modern logic.
康德可能会指出,弗雷格意义上的逻辑要比传统逻辑强,已经越出了分析性的边界。不过,如果是这样,那么康德就有责任表明弗雷格的逻辑是在什么地方越的界。无论如何,如果不考虑现代逻辑的技术,围绕如何划定分析性边界的问题的讨论几乎不会取得任何成果。
35. The question of analytic or a priori truth is, in the terms of the Aristotelian tradition, very close to the question of properties, or essential attributes, as opposed to accidents. I mention this only to emphasize how central the topic is to philosophy. Actually I have no doubt that it is far better to think of these matters in terms of a priori or analytic truth of statements than in terms essentiality of attributes. For, attributes are attributes of the thing, not adjuncts of their name; but whether something attaches essentially or accidentally, better a priori or a posteriori, depends not the thing but on the particular name or description through which the thing is referred to. It is essential that the speaker at the Philosophy Club have the faculty of speech, but accidentally that anyone of name Quine have that faculty, as opposed to being a deaf mute; so it is strictly meaningless to speak of that faculty as essential or accidental to me, who am at once Quine and the speaker of the Philosophy Club. Here incidentally is a simple case of the kind of clarification which is the business of a proper theory of the a priori; and such a theory, if it is to be adequate, is bound to lean heavily on logic.
分析真理或先天真理的问题很接近与偶有属性相对的必然属性(或性质)的问题(这些是亚里士多德传统中的术语)。我提到这一点,只是想强调这个话题在哲学中居于多么中心的地位。实际上,我倒是认为,思考这类问题时,用“分析真理”或“先天真理”这样的术语要远远强于“必然属性”这样的术语。因为,属性总是事物的属性,而不是事物名称的附属品;但是某个事物是必然地还是偶然地具有某个属性,却不依赖于这个事物,而是依赖于我们提到这个事物时所使用的名字或摹状词。哲学俱乐部的演讲者必然地具有演讲的才具,但任何叫奎因的人具有这种才具,而不是聋哑人,却是偶然的;所以,说这样一种才具是我—同时既是奎因又是哲学俱乐部的演讲者—的本质属性或偶然属性,严格说来是无意义的。而谈论真理的先天性和后天性要好得多。一个关于先天的理论涉及到这类概念上的澄清是理所当然的,当然这里插进来的只是这类澄清的一个简单的例子。这类澄清如果是适当的,就必定要大大地倚重逻辑。
36. It’s hard for me to talk for an hour without bringing up G?del’s theorem; and at this point I want to bring it up as an illustration of how logic, at its very subtlest, may reveal philosophical difficulties which nobody would otherwise have dreamed of.
如果我讲了一个小时而没有提到哥德尔定理,那可是说不过去的。现在我就来讲讲它,把它作为一个例证,说明逻辑,如果发挥到极致,是如何揭示出离开了逻辑做梦也无法想到的哲学困难的。
37. G?del’s theorem has to do with the elementary arithmetic, and algebra, of whole numbers. What he establishes is roughly this. Consider any system of rules and principles for proving laws or arithmetic. G?del shows that any such system, no matter how good or bad, is bound to be in one of these two predicaments: either (1) it enables us to prove a falsehood, so it is no good at all; or else (2) it is inadequate for proving certain truths of elementary arithmetic, even though those truths be capable of being written wholly in elementary arithmetical notation.
哥德尔定理与关于整数的算术和代数有关。他所得出的结论大致是这样的。考虑任何用于证明算术规律的规则和公理系统。哥德尔表明,任何一个这样的系统,不论好坏,都摆脱不了下面的两难处境:或者(1)它可以将某些谬误证明为真理,在这种情况下它当然是个不好的系统;或者(2)某些完全由基本的算术记号写就的真理无法被它证明。
38. In a word, every technique for recognizing or generating elementary arithmetical laws is bound to be either inconsistent or incomplete.
总而言之,每一个用来识别或生成算术规律的技术都注定了要么是不一致的要么是不完全的。
39. From an empiricist point of view, this is a baffling situation. The laws of arithmetic are not empirical hypotheses; to say they are true at all, then, must mean that they can be proved, rather than be verified by observation. Otherwise wherein can their truth consist?
从经验主义的观点看,这是一个令人困惑的局面。算术规律不是经验假说;说它们是真的,并不意味着它们可以被观察所证实,而是意味着它们可以得到证明。否则,它们的真理性到底是怎么回事呢?
Yet G?del’s argument shows that, no matter how good our technique, there must be some arithmetical truths that cannot be proved. His argument is mathematically rigorous and there is no escaping it. The empiricist could indeed throw up his hands and declare all arithmetic meaningless; but he is likelier to reconsider his empiricism.
但是哥德尔的论证表明,不管我们的技术有多好,总会有一些无法得到证明的算术真理。他的论证是在数学上是严格的,并无疏漏。经验主义者当然可以两手一摊,声称全部算术是无意义的;但他更有可能会转而反思毛病是不是出在他的经验主义上。
40. It is to be expected, surely, that any statement which is analytic should be capable of proof. So maybe G?del’s result may be interpreted that Kant was right in viewing arithmetic as a priori but synthetic. Undeservedly right, for Kant could have had no justifiable inkling of the incompletability of arithmetic.
想必我们会期待任何分析真理都可以得到证明。因此,哥德尔的结果也许可以被解释成康德关于算术的先天综合性的论断毕竟是正确的。但即使是这样,他也只是碰巧正确了而已,因为对于算术的不完全性,他并没有任何有理由的想法。
41. But then, as I remarked earlier, we have to show where logic stops being analytic and becomes synthetic; for the fact remains that arithmetic is translatable into modern logic without remainder. Logic itself, like arithmetic, is subject to the G?del’s predicament of incompletablity.
但是,就像我前面所说过的那样,我们得搞清楚逻辑是从什么地方由分析进入综合的。如果我们注意到算术可以无遗漏地翻译为逻辑这个事实,那么逻辑本身也无法逃出哥德尔的不完全难题。
42. Actually there is a line which can be drawn across modern logic, such that the logic to one side of the line does admit of complete systematization, and is not adequate to all of elementary arithmetic; while the logic on the other side of the line is incompletable, and contains implicitly elementary arithmetic and the rest of mathematics.
事实上逻辑中确实存在着这条分界线,在分界线的一侧允许有完全的系统,但不包含所有的初等算术;在分界线的另一侧的逻辑则是不完全的,它隐含了初等算术和其他数学部门。
43. The line is just this: it is the line between the part of logic which is neutral on the question of universals, and the part of logic which implies a Platonistic answer to the question of universals.
逻辑的这两个部分的区别在于,其中一个部分对共相问题是中立的,而另一个部分则蕴含了对共相问题的柏拉图主义的答案。
44. It seems natural enough to classify the Platonistic part of logic as a priori synthetic, rather than analytic, if true at all; for this removes the ontological problems of universals itself from the domain of problems whose answers are analytic.
将逻辑的柏拉图主义部分的真理归入先天综合真理而不归入分析真理,这看来是很自然的;因为这样做就把关于共相的本体论问题从分析真理的领域中移开了。
45. If it seems unnatural to classify part of logic as not analytic, this may merely reflect an unwise extension of the term ‘logic’ itself. As a matter of fact it would be equally in keeping with the old tradition to limit ‘logic’ to the hither side of the ontological line I have described, and call the farther side mathematics.
如果把逻辑的一部分不归入分析真理对一些人显得不那么自然,这可能只是反映了当前对“逻辑”一词的适用范围的不明智的扩展。事实上,把“逻辑”局限在本体论中立的一侧,而把另一侧称为“数学”,这种做法更符合人们对逻辑和数学的传统理解。
46. These developments suggest an interesting managing of two traditionally quite distinct philosophical thesis: the epistemological thesis of empiricism, which repudiates the a priori synthetic, and the ontological thesis of nominalism, which repudiates universal entities. A third repudiative thesis, less conspicuous traditionally, also appears to merge with them; viz., the repudiation of mathematics.
有趣的是,这些进展使两个原本截然不同的哲学论题会合到了一起。一方面作为认识论的经验主义拒绝先天综合真理,另一方面作为本体论的唯名论拒绝抽象实体(即共相)。作为这两个论题相联系的纽带是下面这第三个传统上不那么显眼的论题:对数学的拒绝。【数学既有对柏拉图主义又有对先天综合真理的承认。】
47. I have argued that the completable region of logic may conveniently be separated from an incompletable remainder of logico-mathematical science by drawing a boundary between the non-Platonistic and the Platonistic. My method thus far, in arguing that point, has been the single method of assertion. I must leave it at that; to elaborate it would involve me in another somewhat recondite result of G?del’s, the completeness of quantification theory, and various other matters too technical for the evening.
我已经表明,根据对柏拉图主义的态度划线,逻辑的完全性部分可以很方便地从逻辑的不完全性的部分中分离出来。到目前为止,我的论证方法只是断言而已。我必须止于断言,因为要展开我的论证,我必须提到哥德尔的另一个有些深奥的结果,即量词理论的完全性定理,还有其他许多事情,这些技术性很强的话题对于今天晚上的演讲是不相宜的。
48. One point involved is, of course, a criterion for deciding what theories are Platonistic and what ones are not. That is, a criterion for deciding whether a theory involves a commitment to abstract entities, or merely uses abstract words without presupposing corresponding entities.
上面的讨论还涉及到一个问题:用什么标准确定哪些理论是柏拉图主义的,哪些理论不是柏拉图主义的?即:我们据以说某个理论承诺了抽象实体,或仅仅只是使用了抽象词项而没有假定抽象实体的标准是什么?
49. The criterion which I use is this: a theory is committed to abstract entities just in case its variables of quantification—say the ‘x’ of ‘(x)(…x…)’ are sometimes used to refer to abstract entities.
我所使用的标准是:一个理论承诺了抽象实体,当且仅当它的量化变项—比方说‘(x)(…x…)’中的‘x’—有时用来指称抽象实体。
50. I have sometimes been misconstrued as meaning that the question whether there are abstract entities is a question of language. I almost wish I could agree to this, if only to sustain a bon mot due to Dr. James Grier Miller, who undertook to sum up the first of my papers on this topic with the formula ontology recapitulates philology.
我有时被误认为主张抽象对象存在与否的问题是语言问题。我倒真想同意这种误解,哪怕只是为了支持詹姆斯·格瑞尔· 米勒博士的名言:“本体论是语言学的浓缩”,这是他对我关于这个话题的第一篇论文所作的总结。【“本体论是语言学的浓缩”这句话后被奎因用作《语言和对象》一书的篇头语,它是对生物学家海克尔的名言“个体发展史是族群发展史的浓缩”(Ontogeny Recapitulates Phylogeny)的模仿。】
51. Actually I mean no such thing. My criterion is for deciding not whether there are universals, but whether a given theory, a given form of discourse, commits itself to the doctrine that there are universals, and here philology has a place.
实际上我的意思并不是这样的。我的标准【“存在就是成为约束变项的值”】不是用来确定是不是存在抽象对象的,而是用来确定某个理论,或某个述说,是不是对共相的存在作了承诺,在这里语言学确实有一席之地。
52. I haven’t tried to allow time to defend my criterion here, because I talked on that topic before this club last spring. Apropos of the prescribed theme of the present talk, I may merely add that this criterion of ontological commitment, if it is as adequate as I think it to be, is one more example of the use of modern logic in philosophy.
我没有打算留出时间在这里为我的标准辩护,因为这个话题我已经于今年春天在这个俱乐部谈过了。就今天晚上谈的这个话题来看,如果我的这个本体论承诺的标准像我所认为的那样是适当的,那么我想把它当作现代逻辑在哲学中的用处的又一个例子。
53. Before I stop I want to correct a wrong impression which I must have given when I suggested that empiricism, nominalism, and the repudiation of mathematics all converge. I must have given the impression that I thought empirical and nominalism were involved in absurdity.
在我结束之前,我想纠正一个我可能给出的错误印象。我曾指出,经验主义、唯名论和对数学的拒绝这三者是交会的,这可能会带来这样一个印象:我认为经验主义和唯名论会导致荒谬。【如果我们认为数学不是荒谬的,那么导致拒绝数学的经验主义(直接拒绝的是先天综合真理)和唯名论(直接拒绝的是它的对立面,即柏拉图所倡导的那种承认抽象实体的实在论)就是荒谬的。】
54. On the contrary, I am partial to empiricism and nominalism, and my taste for each is enhanced by the support which it proves to gain for the other.
相反,我非常喜欢经验主义,也非常喜欢唯名论,而且这两种喜欢是相互支持和放大的。
55. True, I also have a certain taste for mathematics. But repudiation of classical mathematics as a body of meaningful statements need not involve abandonment of mathematics altogether.
我实际上也对数学有一定的兴趣。拒绝承认经典数学是有意义陈述的集合体,并不意味着要抛弃数学本身。
56. For one thing, we may still pursue classical mathematics, and examine its structure, without us philosophers according genuine meaning to its formulas. And we may, as philosophers, still inquire into the reasons why such a system of meaningless formulas should be useful to science.
首先,我们可以继续经典数学的研究,仔细检查它的结构,而不用我们哲学家给它的公式委派真正的意义。作为哲学家,我们仍可以探究:为什么这样一个由无意义的公式组成的系统居然在科学中能派上用场?【对于这个问题,此时的奎因是没有现成的答案的,但他对这个问题的持续不断的研究终于导致了突破。】
57. Secondly, we may explore the possibilities of constructing a fragmentary neo-mathematics devoid of Platonistic commitments. However useful the classical mathematics may be, it would be important philosophically—from an empiricistic or nominalistic point of view—to show that a fragmentary non-Platonistic mathematics can be theoretically adequate for all applications to natural science.
第二,我们可以探讨建立一种免于柏拉图主义的承诺的新数学的可能性,虽然这种新数学从经典数学的角度看可能是支离破碎的。不管经典数学是多么的有用,如果能表明一种支离破碎的非柏拉图主义的数学在理论上能满足自然科学的所有要求,这从经验主义或唯名论的观点看,仍然有着重要的哲学意义。【奎因企图建立不承诺抽象对象的新数学的努力并不成功。1977年,在接受BBC节目专访时明确表示,他承认抽象对象的存在。】
58. In such a program, whose motivation is purely philosophical, techniques of modern logic would have to be exploited to the utmost.
离开了对现代逻辑技术的充分挖掘和利用,要实现这个出自纯粹哲学动机的方案是不可能的。
联系客服