JOVANA
Explore Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
Back to the library
数学 1879

概念文字(Begriffsschrift):一种模仿算术而构造、用于纯粹思维的公式语言

戈特洛布·弗雷格

给逻辑配上量词的那套记号——「所有」与「有的」从此可以计算。

Choose your version
In depth · the introduction

1879 年之前,逻辑没法在同一句话里把「所有」和「有的」都说清楚。一本小书解决了这件事——也悄悄奠定了现代逻辑。

给逻辑一套说「所有」与「有的」的语法

两千年来,形式逻辑就是亚里士多德的三段论——对付「所有人都会死」绰绰有余,可面对「每个数都有更大的数」这种把一个「每个」套进一个「有的」里的句子,就无能为力了。弗雷格把逻辑从根上重建。他不再把句子读成「主词加谓词」,而是读成「函数填入自变量」,再为「普遍性」加上一个精确的符号。有了它,「所有」「有的」以及事物之间的「关系」,终于能被精确地写下来,并按规则推理。

耶拿的一位沉默教授

戈特洛布·弗雷格在耶拿大学教数学。1879 年,他出版了一本仅 88 页、标题古怪的小书——《概念文字》(Begriffsschrift,意为「概念的书写」)——通篇用他自创的、铺张的二维记号印成。他追的是莱布尼茨的一个古老梦想:一种精确到「推理几乎可以机械进行」的书面语言。几乎没人看得懂。最有分量的那篇书评,把它贬为布尔早已做过之事的劣化版本。此后数十年,弗雷格的天才大体无人识得,直到伯特兰·罗素等人看出了他究竟建起了什么。

它为何重要

弗雷格的概念文字,是一切现代逻辑的根基,并经由逻辑成为计算机科学的根基。「你可以把命题写得毫不含糊,再按固定规则逐步核验论证」——正是这一点,让计算机能够核验一个证明、执行一次数据库查询、跟住一段程序的逻辑。他还把逻辑里的「如果……那么……」钉死了:除非「如果」成立而「那么」不成立,否则它为真——这正是一条电路或一行代码所遵循的规则。

从句子到蓝图

想想「用流畅的散文描述一台机器」和「把它画成蓝图」的差别。散文丰富却含混;蓝图把每个零件、每处连接都定死,谁都能照着造,不必猜。弗雷格把逻辑那松散的散文,变成了蓝图。他的量词,是蓝图里「这对每个零件都成立」的说法;他的条件式,则是一根精确的导线:除非输入为开而输出为关,否则就有信号送出。

一个可交互面板,五个对象各有标着 F 与 G 的两个开关。随着切换,工具显示「所有 F 都是 G」与「有的 F 是 G」是否为真,并在「所有」那句失败时用红圈标出那唯一的例外。

它在知识谱系中的位置

弗雷格立于布尔与二十世纪诸般基础之间。布尔在 1854 年把逻辑变成了一种代数;弗雷格的谓词逻辑,则成了希尔伯特提出他 1900 年问题所用的语言,成了哥德尔证明(1931)「任何这样的系统都无法证明关于算术的全部真理」所用的语言,也成了图灵(1936)定义「何为计算」所用的语言。那些里程碑——本馆也都收录——无一不是用弗雷格发明的语法写成的。

The original document
Original source text
Gottlob Frege · Begriffsschrift, eine der arithmetischen nachgebildete Formelsprache des reinen Denkens · Halle a/S.: Louis Nebert · 1879
A note on this presentation: the booklet's two-dimensional notation cannot be reproduced in running text, and its German prose is summarised below rather than quoted. The full original is at the source link.
Preface
Frege explains that, while testing how far arithmetic could be carried by inference alone, he found ordinary language too pliable to keep a chain of reasoning free of unnoticed gaps; so he devised a written notation for 'pure thought' in which every assumption is made explicit. He invokes Leibniz's old dream of a universal characteristic — a calculus of reasoning — while disclaiming that he had achieved anything so vast. He likens his concept-script to a microscope and everyday language to the eye: the eye is versatile but limited in resolving power, the microscope useless for daily life yet unmatched for the single scientific purpose it is built for.
[ … ]
Part I — Definition of the symbols
Frege sets out his primitive signs: the judgment stroke, which asserts a content; the conditional, joining two contents and denied only when the first holds while the second does not; negation; the identity of content; and, decisively, the sign for generality — a concavity carrying a variable letter — which binds a variable and lets a statement speak of every object at once. Replacing the old subject–predicate split with a function–argument analysis, he can now express relations and nested generality that earlier logic could not.
[ … ]
Part II — Representation and derivation of some judgments of pure thought
From a small set of basic laws (axioms) and essentially a single rule of inference — detaching the consequent of an asserted conditional whose antecedent is also asserted — Frege derives a sequence of logical theorems, each step formally justified. It is the first worked demonstration of proof carried out inside a fully specified formal system.
[ … ]
Part III — Some elements of a general theory of sequences
Frege defines, in purely logical terms, what it means for one object to follow another in a series: he frames the notion of a property inherited along a relation, and from it the 'ancestral' of that relation. With no appeal to intuition or counting, this captures 'following in a sequence' — the logical seed of mathematical induction and of his later attempt to ground arithmetic in logic alone.
Jena · 1879