Gottlob Frege · 戈特洛布·弗雷格
The Foundations of Arithmetic · 算术基础
Undated 未注日期
算术基础——散记
Reflections · Pages
我现在想明白为什么弗雷格在「算数基础」中不给出对数的定义,而只定义了“拥有某个数的概念”。那是因为首先,有不同的概念符合他给出的满足“拥有这个数”这个标准,但这些满足标准的概念可以被说是这个数的概念的对象吗?“七武士”可以说是“七”的对象之一吗?依照直觉,这似乎是可行的,那我们就得承认,一方面,作为概念的“七”拥有多个对象,但另一方面,“七”唯一的对象就是这个数字本身。如何调解这个矛盾?我们可以说,一些拥有数的概念是偶然成为这个概念的对象的。但我们必然会遇到如“方的圆”这样必然属于“零”的对象。我们也可以说,数字0是“零”的本质对象。但在何种意义上它是这样的对象呢?数字3并不满足“拥有三个对象的概念”这条标准,因为它根本不是概念。我们能说,数字3只是概念“三”的自我对象化吗?这也是不可能的,因为概念“三”拥有的对象要远远超过三个,因此概念“三”是不满足成为自己对象的标准的。因此似乎我们只能承认,类似“七武士”“三个火枪手”等等概念不是其对应的数的概念的对象。那么我们会陷入另一个困境:如果这些不是数的概念的对象,也就是说“三”的对象只有数字3,这个数本身是如何与其它概念产生联系的呢?为什么通过一一对应的操作得到的诸集合可以被统一地被称为“三”呢?弗雷格在这里很聪明,他达成了一个非直接的定义,也就是他其实说的是这些集合是“拥有三的概念”这个概念的对象,而避免了对数的直接处理。但也因此,至少在「算术基础」里,他很难澄清数作为一个对象是如何被诸概念分有并获得统一性的。
弗雷格论作为对象的数和作为概念的数
数的概念被弗雷格定义如下:“可以与被包含于某概念其下的对象一一对应的概念的外延”。因此,数的概念其实是一个二阶概念,即概念与概念之间的关系概念。只有当一一对应的关系存在于两个概念之间时,数才有意义。但是这仅仅只是作为概念的数;数同时还应该作为对象。数的对象性显然来自于其自足的分离性,数的概念需要数的对象作为内容。因为数本质上意指一种关系,这关系是不被任何一阶概念包含的,因此独立于概念而存在。比如说,2作为一个对象,其对象性来源于2是被施加于概念之上的性质,而本身并不被概念包含;在“2只猫”这个短语中,猫的概念无法包括2,就像猫的概念无法包括猫的毛色一样。然而,需要注意的是,“2”与“橘色”不同,“橘色”可以直接依附于现实对象上,但“2”只能作为概念与概念的比照关系——因此必须在知性中被构成——被理解,其对概念的内容的超出来自于其它概念,而不是感性对知性的超出。因此,其对象性毋宁说是一种超-概念性,或者说是概念间性。当然,这种概念间性本身需要是普遍的,否则就只会得到联系了两个特殊概念的特殊关系;由此,数需要成为一种普遍对象(既是概念,也是对象),或者是实存概念,其客观性不需要现实对象的担保。
我刚刚也想错了,应该是“拥有对应关系的概念的外延”
我感觉数是对象是对的,但我不知道这个对象是不是更像是理念对象,或者作为对象的概念?
就是理念对象
他的类型论里概念跟对象是严格区分开的,概念没法作为对象,只能是概念的外延作为对象
弗雷格对康德“存在不是谓词”的改造
康德的著名论断“存在不是谓词”奠基在一个前提之上:对他来说,谓词必然本质性地决定了主词的性质。因此,假如存在是谓词,仅仅意指可能性的总体概念对康德来说就会本质性地无法变得“拥有存在”,而拥有存在的特殊概念也将在本质上区别于总体概念。可是,本质上的区别就意味着特殊概念无法匹配总体概念。由此,判断将是不可能的。
弗雷格的改造就在于这个前提。对弗雷格来说,存在可以仅仅作为偶然的性质,而不是本质性的规定对概念起效。总体概念因此不会在本质上不存在,而仅仅是有条件地不存在;对特殊概念亦是如此。对于“猫”的总体概念来说,不存在并非是它的本质属性,而是相关于其下属特殊概念存在与否;相应的,对于“这只橘猫”的特殊概念来说,其存在也不是它的本质属性,而是关联于时间和空间的变量。那弗雷格为什么要将存在设定为一个关联于概念的非本质属性而不是康德式的关联于主词的状态呢?为什么他要坚持存在是一种属性?
表面上,康德和弗雷格的意见似乎并不冲突。说存在规定了主词的状态,即主词可以存在或不存在,与说存在非本质地干涉了概念,其背后的意义似乎是等价的。在两种说法中,存在都没有干涉概念的内容,而两者都指出了概念的可能性与并非-实在性。康德处的主语的存在与否取决于其内容是否得到了特殊概念及其直观对象的填充,而弗雷格处的概念存在与否则取决于其与下属概念和对象的对应。两者唯一的区别似乎仅仅在于构造的方式:总的来说,因为康德将“存在作为谓词”视为等价于“存在作为本质属性/内容”,缺少存在的概念内容便无法对应于拥有存在的概念内容,而这内容无法被改变,因此缺少存在不能转变为拥有存在;但在弗雷格处,这种转变是直接可以作用于概念内容上的。因此,似乎与其说弗雷格对康德的论述进行了改造,不如说他继承了康德。
然而,这里两者真正的分歧来自于弗雷格的深刻洞见:他很清楚地意识到,存在不仅仅是哲学问题,同时还是一个数学问题。存在需要成为概念的内容,是因为存在就意味着“大于零”。在康德处,存在似乎仅仅是主词的一个状态,而与其数学内容是断开的。“树存在”与“有三颗树”对于康德来说是两个不同的判断。但是对于弗雷格来说,“存在”意味着“大于零”,而“不存在”意味着“零”;零与其它数之间没有断裂而都作为概念的数学内容起作用。主词的存在与否无法决定其数学内容,因为我们并不会说“三个主词”,主词作为概念本身必须是不可分的,而只有对主词的个体化内容进行的数学描述才有意义。因此,弗雷格对康德的修正的动机并非是纯粹哲学的,而是来自于他的数学背景。
AI translation from the Chinese original. The original is the authoritative text.
I now understand why Frege does not define number in The Foundations of Arithmetic, but defines only “a concept possessing a certain number.” First, different concepts satisfy his criterion of “possessing this number,” but can these concepts be called objects of the concept of this number? Can “the seven samurai” be called one of the objects of “seven”? Intuitively, this seems possible. We must then acknowledge that, on the one hand, “seven” as a concept has many objects, while on the other its sole object is this number itself. How can this contradiction be reconciled? We might say that some number-bearing concepts become objects of this concept contingently. But we will inevitably encounter an object such as “a square circle,” which necessarily belongs to “zero.” We might also say that the number 0 is “zero’s” essential object. But in what sense is it such an object? The number 3 does not satisfy the criterion “a concept with three objects,” because it is not a concept at all. Can we say that the number 3 is merely the self-objectification of the concept “three”? This is also impossible, because the concept “three” has far more than three objects; it therefore fails to satisfy the criterion for becoming its own object. We seem forced to acknowledge that concepts such as “the seven samurai” and “the three musketeers” are not objects of their corresponding numerical concepts. But then we face another difficulty: if these are not objects of numerical concepts, if the only object of “three” is the number 3, how does this number itself relate to other concepts? Why can collections obtained through one-to-one correspondence be uniformly called “three”? Frege is very clever here. He achieves an indirect definition, effectively saying that these collections are objects of the concept “a concept possessing three,” thereby avoiding a direct treatment of number. Yet for this very reason, at least in the Foundations, he has difficulty clarifying how number as an object is shared by different concepts and acquires unity.
Frege on Number as Object and Number as Concept
Frege defines the concept of number as follows: “the extension of concepts that can be put into one-to-one correspondence with the objects falling under a certain concept.” The concept of number is therefore actually a second-order concept, a concept of relations between concepts. Number has meaning only when a one-to-one correspondence exists between two concepts. But this is merely number as concept; number must also be an object. Number’s objectivity clearly comes from its self-sufficient separateness. The concept of number requires the numerical object as its content. Because number essentially signifies a relation not contained in any first-order concept, it exists independently of concepts. For example, 2’s objectivity as an object comes from its being a property applied to a concept, without itself being contained in the concept. In the phrase “two cats,” the concept of cat cannot include 2, just as it cannot include the cat’s coat color. Yet “2” differs from “orange.” “Orange” can attach directly to an actual object, while “2” can be understood only as a comparative relation between concepts, and must therefore be constituted in the understanding. Its exceeding the concept’s content comes from other concepts, not from sensibility’s exceeding understanding. Its objectivity is thus better called supraconceptuality, or interconceptuality. This interconceptuality must itself be universal, of course, or we obtain only a particular relation connecting two particular concepts. Number must therefore become a universal object, both concept and object, or an existent concept whose objectivity needs no guarantee from actual objects.
I was mistaken just now too. It should be “the extension of concepts possessing a correspondence.”
I feel it is right that number is an object, but I do not know whether this object is more like an ideal object, or a concept as object?
It is an ideal object.
His theory of types strictly separates concepts and objects. A concept cannot be an object; only a concept’s extension can be an object.
Frege’s Revision of Kant’s “Existence Is Not a Predicate”
Kant’s famous claim that “existence is not a predicate” rests on a premise: for him, a predicate necessarily determines the subject’s properties essentially. If existence were a predicate, a general concept signifying only possibility would therefore be essentially incapable of coming to “possess existence,” while a particular concept possessing existence would also differ essentially from the general concept. But an essential difference means that the particular concept cannot correspond to the general concept. Judgment would therefore be impossible.
Frege revises this premise. For Frege, existence can affect a concept merely as a contingent property, rather than an essential determination. The general concept therefore does not essentially fail to exist, but merely fails to exist conditionally; the same holds for particular concepts. For the general concept “cat,” nonexistence is not an essential property, but concerns whether its subordinate particular concepts exist. Correspondingly, for the particular concept “this orange cat,” existence is not an essential property, but a variable related to time and space. Why, then, does Frege posit existence as a nonessential property pertaining to a concept, rather than a Kantian state pertaining to a subject? Why does he insist that existence is a property?
On the surface, Kant’s and Frege’s views seem compatible. To say that existence determines a subject’s state, that the subject can exist or not exist, seems equivalent in underlying meaning to saying that existence affects a concept nonessentially. In both formulations, existence leaves the concept’s content untouched, and both indicate the concept’s possibility and nonactuality. For Kant, whether the subject exists depends on whether its content is fulfilled by particular concepts and their intuitive objects; for Frege, whether the concept exists depends on its correspondence with subordinate concepts and objects. Their sole difference seems to lie in the mode of construction. Broadly, because Kant equates “existence as predicate” with “existence as essential property or content,” the content of a concept lacking existence cannot correspond to the content of a concept possessing existence. Since this content cannot change, lacking existence cannot turn into possessing existence. In Frege, however, this transformation can act directly upon conceptual content. It therefore seems more accurate to say that Frege inherits Kant’s account than revises it.
Their real divergence, however, comes from Frege’s profound insight: he clearly recognizes that existence is not only a philosophical problem, but also a mathematical one. Existence must become a concept’s content because existence means “greater than zero.” In Kant, existence seems merely a state of the subject, disconnected from its mathematical content. “Trees exist” and “There are three trees” are two different judgments for Kant. For Frege, however, “existence” means “greater than zero,” while “nonexistence” means “zero.” There is no rupture between zero and the other numbers; all function as the concept’s mathematical content. Whether the subject exists cannot determine its mathematical content, because we do not say “three subjects.” The subject as a concept must itself be indivisible; only mathematical descriptions of its individuated content are meaningful. Frege’s motive for revising Kant is therefore not purely philosophical, but arises from his mathematical background.