Level C1 · Advanced

A Mathematician's Paradise

"No one shall be able to expel us from the paradise which Cantor has created for us." David Hilbert

Level C1 closes Part I where the whole language has been heading: the full axioms of $\mathsf{ZFC}$, written out and read line by line; the Axiom of Choice examined from three angles that turn out to be exactly the same claim; and the von Neumann universe, which finally gives every set — finite or transfinite — a precise notion of depth.

What you'll learn

—The complete $\mathsf{ZFC}$ axiom list — congruence, extensionality, power set, union, regularity, replacement, infinity, choice — with what each formula actually says spelled out in plain language, not just its name.
—The Axiom of Choice in two equivalent forms, the intuitive "bag of bags" reading, and why it was controversial from the day Zermelo stated it in 1904.
—Three theorems provably equivalent to $\mathsf{AC}$ in $\mathsf{ZF}$: Zermelo's well-ordering theorem, Zorn's Lemma, and Hausdorff's maximal principle.
—A wide spectrum of $\mathsf{AC}$'s consequences — from "every vector space has a basis" and Tychonoff's theorem to the counter-intuitive Banach–Tarski paradox — and why its countable form $\mathsf{AC}_\omega$ has plenty of useful consequences of its own (a countable union of countable sets is countable, König's infinite lemma) but none of the paradoxical ones.
—The rank of a set as tree depth, why finite scales stop working once infinite sets appear, and how ordinals fix that.
—The von Neumann cumulative hierarchy $V_\alpha$ — the entire set-theoretic universe, built one level at a time from $\emptyset$.

1. Reading the axioms of paradise

Hilbert's line about Cantor's paradise wasn't idle praise — set theory really had, by the early 20th century, become the language every other mathematical theory could be translated into. But a paradise built on unrestricted comprehension has dark cellars (Level B2, Russell's paradox), so $\mathsf{ZFC}$ replaces "any property defines a set" with eight explicit construction rules: congruence and extensionality pin down what a set is; power set, union and replacement say how to build bigger sets from ones you already have; regularity rules out sets that contain themselves; infinity finally admits a genuinely infinite set; and choice — the most debated of the eight — says that a chosen element can always be extracted from every element of any collection of nonempty sets.

Reading axioms as sentences, not as opaque symbols, is the entire method of the book distilled: $\forall x\in A\,\exists!y\in x\,\langle x,y\rangle\in f$ is just "$f$ picks exactly one element from every element of $A$" — the bag-of-bags intuition made precise.

2. One axiom, three faces

Zermelo first stated the Axiom of Choice in 1904 to prove that every set can be well-ordered — and it turned out that this and two other, superficially unrelated statements are all provably equivalent to it in $\mathsf{ZF}$:

Axiom of Choice

Every family of nonempty sets has a choice function — one that picks exactly one element from each element of the family.

Zermelo's Theorem

Every set can be well-ordered: equipped with an order in which every nonempty subset has a least element.

Zorn's Lemma

If every chain in a partially ordered set has an upper bound, the set has a maximal element.

Gödel showed in 1938 that adding $\mathsf{AC}$ to the other $\mathsf{ZF}$ axioms can't introduce a contradiction if they weren't already contradictory; Cohen showed in 1963, with forcing, that $\mathsf{AC}$ can't be derived from $\mathsf{ZF}$ either. It sits there independent — true in some models of set theory, false in others — which is exactly why the book treats it as a genuine choice of mathematical worldview, not a settled fact.

The chapter's second big idea answers a quieter question: how "deep" is a set? For hereditarily finite sets, depth is just the longest chain of memberships down to $\emptyset$ — an ordinary natural number. But once infinite sets exist, that scale breaks: $\{\omega\}$ sits one membership-step above $\omega$ itself, yet its paths through the membership tree are no longer than $\omega$'s. The fix is to measure depth not in natural numbers but in ordinals, and to build the universe explicitly, level by level:

Von Neumann hierarchy

\[V_0=\emptyset,\qquad V_{\alpha+1}=\mathcal P(V_\alpha),\qquad V_\lambda=\bigcup_{\alpha<\lambda}V_\alpha\ \text{(}\lambda\text{ a limit)}\]

Every set has a rank — the least $\alpha$ such that it appears in $V_{\alpha+1}$ — and by the Axiom of Regularity, every set has one.

3. Try it yourself

The first five exercises from the chapter's problem set, from the English edition.

Translate the following statements into formulas of the language of $\mathsf{ZF}$ using the membership predicate $\in$ and logical connectives:
1. Set $A$ is a subset of set $B$.
2. Sets $A$ and $B$ have no common elements (are disjoint).
3. Set $A$ contains exactly two elements.
4. The set $P$ is the power set of $A$ (i.e. $P = \mathcal P(A)$).
Formulate the following concepts and theorems as single formulas in the language of Set Theory (using auxiliary predicates if necessary):
- Cantor's theorem: there is no surjection from $A$ to $\mathcal P(A)$.
- The Axiom of Countable Choice ($\mathsf{AC}_\omega$).
- Definition of a filter: a family $\mathcal F \subseteq \mathcal P(S)$ is a filter on $S$ if $S \in \mathcal F$ and $\emptyset \notin \mathcal F$; $\mathcal F$ is closed under intersection; and $\mathcal F$ is closed under supersets.
- Hausdorff's maximal principle: in any partially ordered set, every chain is contained in a maximal chain.
Let $S$ be an infinite set. Consider the collection $\mathcal F$ of all subsets $X \subseteq S$ such that $S \setminus X$ is finite. Prove that $\mathcal F$ is a filter.
Let $A = \{a, b, c\}$. Determine which of the following relations on $A$ are reflexive, symmetric, antisymmetric, or transitive:
1. $R_1 = \{\langle a, a\rangle, \langle b, b\rangle, \langle c, c\rangle\}$.
2. $R_2 = \{\langle a, b\rangle, \langle b, a\rangle, \langle a, a\rangle\}$.
3. $R_3 = \{\langle a, b\rangle, \langle b, c\rangle, \langle a, c\rangle\}$.
Let $J: \omega \times \omega \to \omega$ be defined by $J(a, b) = 2^a(2b + 1) - 1$. Prove that $J$ is a bijection.

Level C1 has 25 exercises in total, each with a full worked solution in the book's appendix.

The rest of Level C1 — and Part II beyond it

This digest skips the full derivation of the equivalence between the Axiom of Choice, Zermelo's theorem and Zorn's Lemma, the definable-mapping trick behind the Axiom of Replacement, and the chapter's closing section on how mathematical structures get constructed within $\mathsf{ZFC}$. From here, Part II — Proof- Theoretical Basis — revisits Language, Logic, Arithmetic and Set Theory at full formal rigor.

Pre-order on Springer Back to the book overview

← Level B2 — The Foundations of Mathematics Back to the book overview →