Edmund Husserl · 埃德蒙德·胡塞尔
Philosophy of Arithmetic and mathematical knowledge · 算术哲学与数学知识
Undated 未注日期
算术哲学与数学知识——散记
Reflections · Pages
胡叟后期数学论面临的难题
胡叟后期的数学论试图达成两个目标,一是对心理主义的克服,二是对逻辑与数学区域一致的本体论建设。最大的问题是,因为数学以多为对象,而逻辑以一为对象,要将这两者划为一致,首先要建立起多与一的等同或转换关系。然而,要将“多”视作一种“一”,就已经预设了这一种转换或变更是可能的。这种变更的合法性来自哪里呢?如果胡叟没有在本体论层面上给出恰当的奠基,那他将无法摆脱「算数哲学」时期心理主义的影子。不仅如此,逻辑与数学的区域一致性也会遭受挑战,因为哪怕其处理的对象都是总体抽象的一般对象,其处理对象的方式不同(将对象视作“一”与将对象视作“多”)也将本质性地决定两者的差异。更不用说,现代数学还有一个符号数学的疑难,也就是如何将在某种程度上与逻辑规则相抵触的符号数学还原为思与逻辑呢?根据胡叟自己的说法,恰恰是现代数学的不可理解性促使他完成了对现象学的构想,也就是对意义之可能性本身的探究的构想。在多大程度上,现代数学可以被视作是形式本体论呢?至少,符号运算本身是与形式本体论之思相独立的。真正的问题在于符号运算的法则是否与形式本体论的运思相符合。
胡赛尔的数是对心理行动的自反性归纳而得到的概念对象“任何事物”的组合;亚里士多德的数是对存在者的数量内容以“尺度”为单位进行抽象并且在根源处不能与决定了“尺度”是什么的存在本身割裂;柏拉图的数是数本身根据理念内容划分出的智性结构。
哪怕在亚那里理论数学所处理的“尺度”是一个尚未被确定种类的具有普遍性的“尺度”,那也是一个与存在者有着直接对应关系的对象,存在者在其中依旧被保持为数之潜能。
然而在胡赛尔处,根据Hopkins的论法,「任何事物」与「尺度」已经距离很远了。「任何事物」本身作为一个概念已经完全替代了对象。「尺度」的规定性来源于被测量的存在,有「尺度」就意味着有等待被确定的内容;而「任何事物」是没有等待被确定这样的内在要求的。
象征系统的断裂
象征系统对自然数本真直观的挑战表现在如下方面。1. 任何有限自然数列本身都可以被象征数列系统性地推导出,而作为数列重复的起始点的本真数列本身也被回溯性地构造为全然缺失本真性的象征符号。2. 数列排序的大与小并不是来自于对本真的多的概念直观比照,而完全是来自于符号演进的次序规则。3. 象征数的对象,如583这个字符所指涉的对象,并不是对583个单位的直观,而是对一个“如果没有象征数列的规则规定就不被任何范畴规定的多”的直观,在这个意义上,583这个字符所指涉的恰恰就是数的非本真概念,而不是数的本真概念;也就是说,583所指涉的依然是一个符号,因此自身是作为符号的符号被构造的。对这种与非本真概念的严格平行对应关系进行更加深入的分析,就可以发现,这种平行本身也是一种纳入:583所指涉的对象恰恰是在与582和584的并置关系中被明确的对象,因此这一种对象并不是外在于系统,而恰恰是内在于象征系统的。被象征的象征也是象征。在这里,本真概念对现代算术的影响几乎已经不可见了。这就导致了一个问题,也就是如果说现代算术是从古典(希腊)数学中发展而来的,以直观充实为己任的希腊数学科学是如何被转换成现代形式化的算术的呢?假如我们接受直观明见性对数学科学的奠基要求,我们又如何理解现代算术的特殊性与独立性呢?Hopkins认为胡赛尔在「算数哲学」中的失败并不是他对直观奠基的失败,而是他没有完成对这一转折的演示。也就是说,胡赛尔在将数学的本真直观扩充到非本真概念之后,这非本真概念却突然转化成了另一种造物,这一种转折成为了事件性的、与整体断裂的,甚至是完全不可理解的:为什么非本真概念符号会突兀地转折成象征系统?Hopkins认为,这是因为胡赛尔在理解非本真概念的时候,就已经有意识地去引入一个系统性了;而这一种引入未必是完全正当的。也就是说,对于胡赛尔,象征系统及其公理化是一个既定事实,也是数学的根本形态。因此,他忽略了这一种数学的形态是现代的,与希腊式的、也就是他在「算数哲学」开篇提出的设想是背道而驰、不可化约的。所以,对胡赛尔而言,这两种相反的思路同时存在的情况是不可理解的一种断裂,或者说是一种根本上的偶然性。在这个意义上,「算数哲学」的命运是非常辩证法式的,简单的对本真直观的量的扩充迫使了非本真直观的发生,而非本真直观的发生又转而需要迫使象征系统的发生,再回溯性地将自身设置成了一种必然结构。只是这样一种“精神现象学”式的结构,根据Hopkins的理解,似乎并不足以令他满意。而这当中的主要原因似乎在于,他认为胡赛尔在后一点上功亏一篑:胡赛尔成功构造出了从本真直观到非本真直观转换的必然性(通过设置“人类思维有限性”这个事实),却没有构造出非本真直观转换到象征系统的必然性。为什么非本真直观的象征最后必然被系统化?是什么允许了这个系统的发生?为什么不能只接受有暧昧的多?为什么我们需要并可以得到583,而不是比5个单位多的一群?或者,为什么要以这样的形式(即对自然数列的重复)去系统化?为什么不能以不断对多添加单位的方式(这个问题胡赛尔似乎做出了答复,也就是人类思维有限性要求了对基本直观元素的组合;可这依然预设了对本真直观的依赖,而这恰恰是被系统论否定的)?胡赛尔的解答归根结底可以说是:因为我们有一个这样的(比如说,十进制)系统。这也是Hopkins对胡赛尔的最根本的批评,也就是胡赛尔在这个问题上将偶然事件视为了绝对根据,而这与他自己的哲学立场是不统一的。
说直观经验是作为潜能潜在于现代数学的形式中也是不对的,因为潜能不能为知识辩护;而不能直接展现的直观就不是足以产生知识的明见性;那最终的后果就是要求必须每个数都被穷举
多与数
数是被决定的多。这意味着对多的直观可以是本真也可以是非本真的,而非本真的多也可以是不明确的、以数列中最开始的几个自然数为基本符号向外延拓的暧昧的多。而数则只可能是以数列中最开始的几个自然数(被本真给予的、自然的多)为基本符号,依照“大与小”的本真规则被重复的确定的非本真意向。然而数的符号化很快超越了这一“非本真概念的代名”的领域,作为机械算法替代了与其平行的概念意向。
多的无限性和概念扩张来自于其以本真直观作为符号直观整体的图像,而将整体视作了一个种类整全性被通过基本计数过程实现的可能性,因而获得了其无限性;而数的无限性与扩张被奠基于多的无限性与扩张,而要求一个以基本元素的重复为基础的简单符号系统,这种符号系统与数的概念的对应需要进一步阐明
数的概念无法脱离符号被思考,因为概念本身就是依托于意向性的符号,并在这个意义上并行于符号操作被思考。然而符号化本身就意味着这种符号可以被独立于概念意向被操作。其独立性的证据在于数字符号可以被回溯性使用在自然数列开头的几个可以被本真直观的数上。
弗雷格批评胡赛尔在「算术起源」中将心理主义引入了数学逻辑中,这个批评是不公允的。因为他没有意识到胡赛尔最根本的点是在对“起源”的讨论上。胡赛尔因此并不是心理主义,也没有否认数学逻辑的客观性与独立性,而是试图在心理现象中为差异于心理现象的逻辑现象奠基。所以如果一定要批评胡赛尔,那也不应该是认为胡赛尔混淆了逻辑与心理范畴最终导致写作的失败,而应该去关注当中对逻辑与心理关系更深刻且精微的刻画。最基本的一个事实是,胡赛尔的运思在「算术起源」中经历了一个巨大的转折:从试图将算术法则奠基于从属于心理经验与现象领域的自然数(多性)转向将自然数的多性反而奠基于数学符号的运算法则。这里的关系要如此把握:对胡赛尔来说,运算法则优先于数本身,而运算法则几乎总是符号化的,因此数学符号(在这种意义上)总是要优先于自然数。这并没有抹消数的概念与数的概念理想化这两个维度,而只是胡赛尔意识到了数学符号并非仅仅是一种对数的概念的替代,而反而是使普遍的运算成为可能的前提,并给了这一种普遍性它应有的地位而已。
AI translation from the Chinese original. The original is the authoritative text.
The difficulty facing Husserl’s later theory of mathematics
Husserl’s later theory of mathematics attempts to achieve two goals: overcoming psychologism and constructing an ontology in which the logical and mathematical regions coincide. The chief problem is that mathematics takes the many as its object, whereas logic takes the one. To bring the two into coincidence, one must first establish an equivalence or convertibility between many and one. Yet regarding ‘many’ as a kind of ‘one’ already presupposes that this conversion or modification is possible. Where does the legitimacy of such a modification come from? If Husserl fails to provide a proper ontological grounding, he cannot escape the shadow of psychologism from the period of Philosophy of Arithmetic. Furthermore, the regional coincidence of logic and mathematics will itself be challenged. Even if both deal with generic objects abstracted as wholes, their different ways of treating objects—as ‘one’ or as ‘many’—will essentially determine their difference. Modern mathematics also faces the difficulty of symbolic mathematics: how can symbolic mathematics, which to some extent conflicts with logical rules, be reduced to thought and logic? According to Husserl himself, it was precisely the unintelligibility of modern mathematics that prompted his conception of phenomenology, the investigation of the very possibility of meaning. To what extent can modern mathematics be regarded as formal ontology? Symbolic calculation itself, at least, is independent of the thinking of formal ontology. The real question is whether the laws of symbolic calculation conform to the thinking of formal ontology.
For Husserl, number is a combination of the conceptual object ‘anything whatsoever,’ obtained through a reflexive induction of psychic acts. For Aristotle, number is an abstraction of the quantitative content of beings in units of a ‘measure’ and cannot, at its root, be severed from Being itself, which determines what that ‘measure’ is. For Plato, number is an intelligible structure that number itself articulates according to its ideal content.
Even if the ‘measure’ with which theoretical mathematics deals in Aristotle is a universal ‘measure’ whose kind has not yet been determined, it remains an object directly corresponding to beings, which are still preserved within it as the potentiality of number.
In Husserl, however, according to Hopkins’ account, ‘anything whatsoever’ is already far removed from ‘measure.’ ‘Anything whatsoever,’ as a concept, has entirely replaced the object. The determinacy of ‘measure’ comes from the being measured: a ‘measure’ implies content awaiting determination. ‘Anything whatsoever,’ however, has no such intrinsic requirement to await determination.
The rupture of the symbolic system
The symbolic system challenges authentic intuition of natural numbers in the following ways. 1. Any finite series of natural numbers can itself be systematically derived through a symbolic numerical series. The authentic series that serves as the starting point for the repetition of the series is itself retrospectively constituted as symbolic signs entirely lacking authenticity. 2. Greater and lesser in the ordering of a numerical series do not come from intuitively comparing authentic concepts of multiplicity, but entirely from the sequential rules governing the progression of signs. 3. The object of a symbolic number—for example, the object referred to by the character sequence 583—is not an intuition of 583 units, but an intuition of ‘a multiplicity determined by no category except through the rules of the symbolic numerical series.’ In this sense, what 583 refers to is precisely the inauthentic concept of number, not its authentic concept. That is, 583 still refers to a sign and is itself constituted as a sign of a sign. A closer analysis of this strict parallel correspondence with the inauthentic concept reveals that the parallelism itself is also an inclusion: the object referred to by 583 is precisely an object clarified by its juxtaposition with 582 and 584. Such an object is therefore not external to the system but internal to the symbolic system. The symbolized symbol is also a symbol. Here the influence of authentic concepts on modern arithmetic has become almost invisible. This poses a problem: if modern arithmetic developed from classical Greek mathematics, how was Greek mathematical science, committed to intuitive fulfillment, transformed into modern formalized arithmetic? If we accept the demand that mathematical science be grounded in intuitive evidence, how can we understand the distinctiveness and independence of modern arithmetic? Hopkins holds that Husserl’s failure in Philosophy of Arithmetic is not a failure to ground intuition, but a failure to demonstrate this transition. After Husserl expands authentic mathematical intuition to encompass inauthentic concepts, these concepts suddenly turn into something else. This turn becomes an event, a rupture with the whole, even something entirely unintelligible: why do signs of inauthentic concepts abruptly turn into a symbolic system? Hopkins thinks this is because Husserl already consciously introduces systematicity when understanding inauthentic concepts, and this introduction may not be entirely legitimate. For Husserl, that is, the symbolic system and its axiomatization are an established fact and the fundamental form of mathematics. He thus overlooks that this form of mathematics is modern and irreducibly opposed to the Greek conception he himself proposes at the beginning of Philosophy of Arithmetic. For Husserl, the coexistence of these opposed lines of thought is therefore an unintelligible rupture, or a fundamental contingency. In this sense, the fate of Philosophy of Arithmetic is thoroughly dialectical. A simple quantitative expansion of authentic intuition forces inauthentic intuition to arise; inauthentic intuition in turn must force a symbolic system to arise, retrospectively establishing itself as a necessary structure. Yet according to Hopkins’ understanding, such a ‘Phenomenology of Spirit’-like structure apparently does not satisfy him. The chief reason seems to be that he considers Husserl to have fallen short on the latter point: Husserl successfully constructs the necessity of the transition from authentic to inauthentic intuition, by positing the fact of ‘the finitude of human thought,’ but does not construct the necessity of the transition from inauthentic intuition to the symbolic system. Why must the symbols of inauthentic intuition ultimately be systematized? What permits this system to arise? Why can we not simply accept an ambiguous multiplicity? Why do we need, and how can we obtain, 583 rather than a group of more than five units? Or why systematize in this form, through the repetition of the natural-number series? Why not continually add units to a multiplicity? Husserl seems to have answered this last question: the finitude of human thought requires combinations of basic intuitive elements. Yet this still presupposes a dependence on authentic intuition, precisely what the systemic account denies. Husserl’s answer ultimately amounts to: because we have a system of this kind—for example, a decimal system. This is also Hopkins’ most fundamental criticism of Husserl: on this question, Husserl treats a contingent event as an absolute ground, which conflicts with his own philosophical position.
It is also wrong to say that intuitive experience is latent as potentiality in the forms of modern mathematics, because potentiality cannot justify knowledge. An intuition that cannot be directly exhibited is not evidence sufficient to yield knowledge. The ultimate consequence would be to require every number to be exhaustively enumerated.
Multiplicity and number
Number is determined multiplicity. This means that an intuition of multiplicity can be authentic or inauthentic. Inauthentic multiplicity can also be an unclear, ambiguous multiplicity extended outward using the first few natural numbers of the series as basic signs. Number, however, can only be a determinate inauthentic intention repeated according to the authentic rules of ‘greater and lesser,’ using the first few natural numbers—authentically given, natural multiplicities—as basic signs. Yet the symbolization of number quickly exceeds this domain of ‘names standing for inauthentic concepts,’ replacing the parallel conceptual intentions with mechanical algorithms.
The infinity and conceptual expansion of multiplicity come from using authentic intuition as an image of the whole of symbolic intuition, treating the whole as a possibility whose generic completeness can be realized through the basic process of counting, and thereby obtaining infinity. The infinity and expansion of number are grounded in those of multiplicity and require a simple symbolic system based on the repetition of basic elements. The correspondence between this symbolic system and the concept of number requires further clarification.
The concept of number cannot be thought apart from signs, because the concept is itself a sign resting on intentionality, and in this sense is thought in parallel with symbolic operations. Yet symbolization itself means that these signs can be manipulated independently of conceptual intentions. Evidence of this independence lies in the retrospective application of numerical signs to the first few numbers in the natural-number series, which can be authentically intuited.
Frege’s criticism that Husserl introduces psychologism into mathematical logic in The Origin of Arithmetic is unfair. He fails to recognize that Husserl’s most fundamental point concerns the discussion of ‘origin.’ Husserl is therefore not a psychologist in this sense, nor does he deny the objectivity and independence of mathematical logic. Rather, he attempts to ground logical phenomena, distinct from psychic phenomena, in psychic phenomena. If one insists on criticizing Husserl, the criticism should therefore not be that he confuses logical and psychological categories and thereby causes the work to fail. One should instead attend to its deeper and subtler characterization of the relation between logic and psychology. A basic fact is that Husserl’s thinking undergoes a great turn within The Origin of Arithmetic: from attempting to ground arithmetical laws in natural numbers, or multiplicities, belonging to psychic experience and phenomena, he turns instead to grounding the multiplicity of natural numbers in the operational laws of mathematical signs. The relation should be understood as follows: for Husserl, operational laws precede numbers themselves, and operational laws are almost always symbolic. Mathematical signs, in this sense, therefore always precede natural numbers. This does not efface either the concept of number or the idealization of that concept. It is simply that Husserl recognizes mathematical signs as not merely substitutes for concepts of number, but the very precondition making universal operations possible, and accords this universality its proper place.