启程:阿基米德界的逻辑基石与集合论全景
在数学宇宙的深邃星图中,我们的探险者“证明者”从逻辑神殿的沉思中苏醒。他的第一个任务是理解这个宇宙的构成法则——那些统御一切数学对象存在与关系的根本原理。这引导我们进入集合论的领域,其中集合被视为最基本的存在单元,而 Zermelo-Fraenkel 公理系统(含选择公理,简称 ZFC)则扮演着宇宙宪法的角色。然而,要驾驭这套宪法,我们必须先掌握其立法语言:一阶逻辑(first-order logic)。本章将作为探险的起点,系统阐述从朴素集合论的危机到公理化体系建立的历程,揭示集合如何通过递归定义构建出整个数学世界,并探讨基数、序数以及连续统假设等核心概念在宇宙地图中的位置。最终,我们将通过“神殿的无限阶梯”这一隐喻,理解超限归纳法如何成为探索无限的可靠阶梯。
---
一、一阶逻辑:宇宙的立法语言
任何公理体系都必须建立在精确的形式语言之上。一阶逻辑(first-order logic)是集合论乃至整个数学的标准表述工具,它允许我们量化个体对象(集合),但不能量化集合的集合(即不能直接谈论“所有集合的集合”)。这种限制恰恰是避免悖论的关键屏障。
1.1 形式语言与语法
一阶语言 $\mathcal{L}$ 由以下符号组成:
- **逻辑符号**:$\neg$(非)、$\land$(且)、$\lor$(或)、$\rightarrow$(蕴含)、$\leftrightarrow$(等价)、$\forall$(全称量词)、$\exists$(存在量词)、$=$(等号)、变量 $x, y, z, \dots$
- **非逻辑符号**:取决于讨论的领域。在集合论中,通常只有一个非逻辑符号:$\in$(属于关系)。
定义 1.1 (项与公式 / Term and Formula) 在纯集合论语言 $\mathcal{L} = \{\in\}$ 中,项(term)递归定义为:所有变量都是项。没有常量或函数符号,因此项就是变量本身。公式(formula)递归定义为: 1. 若 $t, s$ 是项,则 $t \in s$ 和 $t = s$ 是原子公式。 2. 若 $\varphi, \psi$ 是公式,则 $\neg\varphi$、$\varphi \land \psi$、$\varphi \lor \psi$、$\varphi \rightarrow \psi$、$\varphi \leftrightarrow \psi$ 是公式。 3. 若 $\varphi$ 是公式,$x$ 是变量,则 $\forall x \varphi$ 和 $\exists x \varphi$ 是公式。 只有有限次应用以上规则得到的字符串才是合式公式(well-formed formula, wff)。
Definition 1.1 (Term and Formula) In pure set theory language $\mathcal{L} = \{\in\}$, terms are defined recursively: all variables are terms. There are no constant or function symbols, so terms are just variables. Formulas are defined recursively: 1. If $t, s$ are terms, then $t \in s$ and $t = s$ are atomic formulas. 2. If $\varphi, \psi$ are formulas, then $\neg\varphi$, $\varphi \land \psi$, $\varphi \lor \psi$, $\varphi \rightarrow \psi$, $\varphi \leftrightarrow \psi$ are formulas. 3. If $\varphi$ is a formula and $x$ is a variable, then $\forall x \varphi$ and $\exists x \varphi$ are formulas. Only strings obtained by finitely many applications of these rules are well-formed formulas.
语法中需注意变量的自由出现(free occurrence)与约束出现(bound occurrence)。在 $\forall x \varphi$ 中,$x$ 的所有自由出现被量词约束。一个公式若无自由变量,称为句子(sentence)。
1.2 语义:结构与模型
一阶逻辑的语义通过结构(structure)定义。在集合论中,一个结构 $\mathcal{M} = (M, \in^{\mathcal{M}})$,其中 $M$ 是非空集合(论域),$\in^{\mathcal{M}} \subseteq M \times M$ 是一个二元关系,解释为“属于”。注意:$\in^{\mathcal{M}}$ 不一定是真正的集合属于关系,它只是 $M$ 上的一个关系。
定义 1.2 (满足关系 / Satisfaction) 给定结构 $\mathcal{M} = (M, \in^{\mathcal{M}})$ 和赋值函数 $s: \text{Var} \to M$(将变量映射到 $M$ 中元素),递归定义 $\mathcal{M} \models \varphi[s]$($\varphi$ 在 $\mathcal{M}$ 下由 $s$ 满足):
- $\mathcal{M} \models t \in s [s]$ 当且仅当 $s(t) \in^{\mathcal{M}} s(s)$。
- $\mathcal{M} \models t = s [s]$ 当且仅当 $s(t) = s(s)$。
- $\mathcal{M} \models \neg\varphi [s]$ 当且仅当 $\mathcal{M} \not\models \varphi [s]$。
- $\mathcal{M} \models \varphi \land \psi [s]$ 当且仅当 $\mathcal{M} \models \varphi [s]$ 且 $\mathcal{M} \models \psi [s]$。
- $\mathcal{M} \models \forall x \varphi [s]$ 当且仅当 对所有 $a \in M$,$\mathcal{M} \models \varphi [s(x \mapsto a)]$,其中 $s(x \mapsto a)$ 是与 $s$ 几乎相同但在 $x$ 上取值 $a$ 的赋值。
若 $\varphi$ 是句子(无自由变量),则 $\mathcal{M} \models \varphi$ 表示对任意赋值 $s$ 都有 $\mathcal{M} \models \varphi [s]$,此时称 $\mathcal{M}$ 是 $\varphi$ 的模型。
Definition 1.2 (Satisfaction) Given a structure $\mathcal{M} = (M, \in^{\mathcal{M}})$ and an assignment $s: \text{Var} \to M$, we define recursively $\mathcal{M} \models \varphi[s]$:
- $\mathcal{M} \models t \in s [s]$ iff $s(t) \in^{\mathcal{M}} s(s)$.
- $\mathcal{M} \models t = s [s]$ iff $s(t) = s(s)$.
- $\mathcal{M} \models \neg\varphi [s]$ iff $\mathcal{M} \not\models \varphi [s]$.
- $\mathcal{M} \models \varphi \land \psi [s]$ iff $\mathcal{M} \models \varphi [s]$ and $\mathcal{M} \models \psi [s]$.
- $\mathcal{M} \models \forall x \varphi [s]$ iff for all $a \in M$, $\mathcal{M} \models \varphi [s(x \mapsto a)]$.
If $\varphi$ is a sentence, $\mathcal{M} \models \varphi$ means for every assignment $s$, $\mathcal{M} \models \varphi [s]$. Then $\mathcal{M}$ is a model of $\varphi$.
语义定义确保了我们可以精确讨论“一个句子在所有模型中成立”即逻辑有效(logically valid),以及“从一组句子推出另一个句子”即逻辑蕴含(logical consequence)。哥德尔完备性定理建立了语法推导与语义有效之间的等价性:$\Gamma \models \varphi$ 当且仅当 $\Gamma \vdash \varphi$。
1.3 形式系统与公理化方法
集合论公理化采用 Hilbert 式的形式系统:我们指定一组公理(axioms)和推理规则(主要是肯定前件和全称量词推广),从而定义可证明(provable)的概念。一个句子 $\varphi$ 在 ZFC 中可证明,记 $\text{ZFC} \vdash \varphi$,如果存在有限步推理从公理得到 $\varphi$。
公理化方法(axiomatic method) 的核心在于:不预先定义“集合”是什么,而是列出集合必须满足的性质(公理),任何满足这些性质的系统都是集合论的模型。这种方法将焦点从“本质定义”转移到“关系结构”,正是结构主义数学哲学(structuralism)的体现。
---
二、ZFC 公理系统:宇宙的宪法
ZFC 由以下公理组成(使用 $\in$ 语言)。我们以双语形式呈现每条公理,并解释其直观意义与必要性。
2.1 外延公理 (Axiom of Extensionality)
> 外延公理 > $\forall x \forall y (\forall z (z \in x \leftrightarrow z \in y) \rightarrow x = y)$ > Axiom of Extensionality > $\forall x \forall y (\forall z (z \in x \leftrightarrow z \in y) \rightarrow x = y)$
解释:两个集合相等当且仅当它们有相同的元素。这是集合作为“元素的聚合”的根本性质,排除了集合有“隐藏结构”的可能性。外延性确立了集合由其元素唯一确定,是集合论辨识对象的基石。
2.2 空集公理 (Axiom of the Empty Set)
> 空集公理 > $\exists x \forall y (y \notin x)$ > Axiom of the Empty Set > $\exists x \forall y (y \notin x)$
解释:存在一个没有元素的集合,记作 $\emptyset$。虽然可以从其他公理推导(如无穷公理给出无限集,再用分离公理得到空集),但单独列出强调了空集作为起点的重要性。空集是所有集合的子集,是累积层次 $V_0$。
2.3 配对公理 (Axiom of Pairing)
> 配对公理 > $\forall x \forall y \exists z \forall w (w \in z \leftrightarrow (w = x \lor w = y))$ > Axiom of Pairing > $\forall x \forall y \exists z \forall w (w \in z \leftrightarrow (w = x \lor w = y))$
解释:给定任意两个集合 $x, y$,存在一个集合 $z$ 恰好以 $x$ 和 $y$ 为元素,记作 $\{x, y\}$。当 $x = y$ 时,得到单元集 $\{x\}$。配对公理允许我们构造有限集合,是构建更复杂集合的基础。
2.4 并集公理 (Axiom of Union)
> 并集公理 > $\forall x \exists y \forall z (z \in y \leftrightarrow \exists w (z \in w \land w \in x))$ > Axiom of Union > $\forall x \exists y \forall z (z \in y \leftrightarrow \exists w (z \in w \land w \in x))$
解释:给定集合 $x$(其元素是集合),存在并集 $\bigcup x = \{ z : \exists w \in x (z \in w) \}$,即所有属于 $x$ 中某个集合的元素组成的集合。例如,$\bigcup \{\{1,2\}, \{2,3\}\} = \{1,2,3\}$。并集公理使我们能够“展平”集合的集合。
2.5 幂集公理 (Axiom of Power Set)
> 幂集公理 > $\forall x \exists y \forall z (z \in y \leftrightarrow \forall w (w \in z \rightarrow w \in x))$ > Axiom of Power Set > $\forall x \exists y \forall z (z \in y \leftrightarrow \forall w (w \in z \rightarrow w \in x))$
解释:给定集合 $x$,存在幂集 $\mathcal{P}(x) = \{ z : z \subseteq x \}$,即 $x$ 的所有子集组成的集合。幂集公理极大地扩展了集合的宇宙,因为 $|\mathcal{P}(x)| = 2^{|x|}$。康托尔定理 $|x| < |\mathcal{P}(x)|$ 依赖于此公理。
2.6 无穷公理 (Axiom of Infinity)
> 无穷公理 > $\exists x (\emptyset \in x \land \forall y (y \in x \rightarrow y \cup \{y\} \in x))$ > Axiom of Infinity > $\exists x (\emptyset \in x \land \forall y (y \in x \rightarrow y \cup \{y\} \in x))$
解释:存在一个无限集合 $x$,它包含空集,且若 $y \in x$ 则 $y \cup \{y\}$(即 $y$ 的后继)也在 $x$ 中。这样的集合包含所有有限序数 $\omega = \{0,1,2,\dots\}$。无穷公理是数学中无限对象的起点,没有它,集合论只能谈论有限集合。
2.7 替换公理模式 (Axiom Schema of Replacement)
> 替换公理模式 > 对任意一阶公式 $\varphi(x, y, \vec{p})$(其中 $\vec{p}$ 是参数),若 $\forall x \in a \exists! y \varphi(x, y, \vec{p})$,则 $\exists b \forall y (y \in b \leftrightarrow \exists x \in a \varphi(x, y, \vec{p}))$。 > Axiom Schema of Replacement > For any formula $\varphi(x, y, \vec{p})$, if $\forall x \in a \exists! y \varphi(x, y, \vec{p})$, then $\exists b \forall y (y \in b \leftrightarrow \exists x \in a \varphi(x, y, \vec{p}))$.
解释:替换公理是模式(schema),即对每个公式 $\varphi$ 都是一个公理。它断言:若 $a$ 是一个集合,且 $\varphi$ 定义了一个从 $a$ 中元素到“唯一” $y$ 的函数(可能跨集合),则这些 $y$ 的像构成一个集合 $b$。直观上,替换公理允许我们通过已知集合上的“函数”得到新集合。它是证明许多关键结果(如冯·诺依曼层次 $V_\alpha$ 对于极限序数 $\alpha$ 是集合)所必需的。没有它,我们无法证明所有序数构成一个真类而非集合。
2.8 正则公理 (Axiom of Regularity / Foundation)
> 正则公理 > $\forall x (x \neq \emptyset \rightarrow \exists y \in x (y \cap x = \emptyset))$ > Axiom of Regularity (Foundation) > $\forall x (x \neq \emptyset \rightarrow \exists y \in x (y \cap x = \emptyset))$
解释:每个非空集合 $x$ 包含一个元素 $y$ 使得 $x$ 与 $y$ 不相交。这排除了集合属于自身的循环($x \in x$)以及无穷递降属于链($\dots \in x_2 \in x_1 \in x_0$)。正则公理确保了集合的“良基性”(well-foundedness),使得超限归纳法成为可能。它也隐含了 $\in$ 关系的反对称性(若 $x \in y$ 则 $y \notin x$)。值得注意的是,大多数集合论