Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedSession-authored (Fable 5 assisted)precheck passjudge pass (z-ai/glm-5.2)verified 2026-07-29 (claude-fable-5) rests on unproved material (inherited)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Rests on 3 statements not proved in this library, by way of the results it cites. This item cites no such statement directly; it depends on results that do. The unproved premises it inherits are Cohen's first model: an infinite Dedekind-finite set of reals, Sierpiński 1947: the generalised continuum hypothesis implies the Axiom of Choice and The continuum hypothesis and its generalisation are independent of ZFC. Each is recorded with a citation to the literature and is not established here, because the track that would prove it has not yet been developed in this library. Everything else in this proof is proved here.

Hessenberg: κκ=κ\kappa \otimes \kappa = \kappa for every infinite cardinal κ\kappa, proved in ZF from the canonical well-order of κ×κ\kappa \times \kappa

Statement

Let κ\kappa be an infinite cardinal, that is a cardinal (Cardinal (initial ordinal) and cardinality) with ωκ\omega \le \kappa. Then

κκ=κ,equivalentlyκ×κ=κ\kappa \otimes \kappa = \kappa, \qquad \text{equivalently} \qquad \lvert \kappa \times \kappa \rvert = \kappa

(Cardinal sum κλ\kappa \oplus \lambda, product κλ\kappa \otimes \lambda and exponentiation κλ\kappa^{\lambda}, and why they are written apart from the ordinal operations).

This is a theorem of ZF and uses no choice principle. The well-order that carries the proof is written down from the ordinal order on κ\kappa; nothing is selected anywhere. That matters for this page: Hessenberg's theorem is exactly the part of "an infinite set is the same size as its square" that survives without choice.

Facts & Assumptions

Given: An infinite cardinal κ\kappa, in ZF. No choice principle is assumed. For ordinals ξ,η\xi, \eta write max(ξ,η)\max(\xi,\eta) for the \subseteq-larger of the two, which exists by comparability.

[L1]

For a well-orderable XX: XXX \approx \lvert X \rvert, the value is a cardinal, equinumerous sets receive the same one, αα\lvert \alpha \rvert \le \alpha, and α=α\lvert \alpha \rvert = \alpha exactly when α\alpha is a cardinal (A set equinumerous with some ordinal has a least such ordinal, that ordinal is a cardinal, and equinumerous sets get the same one; no choice principle is used).

[L3]

For cardinals, κλ\kappa \le \lambda iff κλ\kappa \preceq \lambda; and ABA \preceq B with both well-orderable gives AB\lvert A \rvert \le \lvert B \rvert (claim (a) of Commutativity, associativity, distributivity and monotonicity of \oplus and \otimes, the unit laws, the two exponent laws, and κλ\kappa \le \lambda if and only if κ\kappa injects into λ\lambda).

[L5]

N×NN\mathbb{N} \times \mathbb{N} \approx \mathbb{N}, that is ω×ωω\omega \times \omega \approx \omega (N×NN\mathbb{N} \times \mathbb{N} \approx \mathbb{N}, Finite, countably infinite, countable, uncountable).

[L6]

Every well-order is order isomorphic to exactly one ordinal, its order type; an order isomorphism is a bijection and carries the initial segment below a point onto the initial segment below its image (Every well-order has a unique order type, Order embedding and order isomorphism, Initial segment of a well-order).

[L7]

If SWS \subseteq W for a well-order (W,<)(W,<) satisfies "W<aSW_{<a} \subseteq S implies aSa \in S" for every aWa \in W, then S=WS = W (Transfinite induction).

[L8]

Ordinals: elements of ordinals are ordinals, αα\alpha \notin \alpha, αβ\alpha \subseteq \beta iff αβ\alpha \in \beta or α=β\alpha = \beta, trichotomy holds, every nonempty set of ordinals has an \in-least element, and every set of ordinals is well ordered by \in (Basic closure properties of ordinals, Trichotomy and well-ordering of the ordinals, Ordinal (von Neumann), Well-order and well-ordered set).

[L9]

ω\omega is the least limit ordinal and is an ordinal (ω\omega is the least limit ordinal); a bijection witnesses \approx and a subset inclusion is an injection (Equinumerous sets, ABA \approx B and ABA \preceq B, Injection, surjection, bijection).

Proof

technique · direct
1.1

For an ordinal α\alpha define (ξ,η)(ξ,η)(\xi,\eta) \lhd (\xi',\eta') on α×α\alpha \times \alpha to hold when max(ξ,η)max(ξ,η)\max(\xi,\eta) \in \max(\xi',\eta'), or the two maxima are equal and ξξ\xi \in \xi', or the two maxima are equal, ξ=ξ\xi = \xi' and ηη\eta \in \eta'; this is the lexicographic order on the triple (max(ξ,η),ξ,η)(\max(\xi,\eta), \xi, \eta) of ordinals, hence irreflexive, transitive and trichotomous by [L8], and a nonempty Sα×αS \subseteq \alpha \times \alpha has a \lhd-least element obtained by taking in turn the \in-least maximum occurring in SS, then the \in-least admissible ξ\xi, then the \in-least admissible η\eta, each of which is the least element of a nonempty set of ordinals and so is determined rather than chosen; therefore \lhd well-orders α×α\alpha \times \alpha.

L8
1.2

For (ξ,η)α×α(\xi,\eta) \in \alpha \times \alpha put γ=max(ξ,η){max(ξ,η)}\gamma = \max(\xi,\eta) \cup \{\max(\xi,\eta)\}, the successor of the maximum; then every (ξ,η)(ξ,η)(\xi',\eta') \lhd (\xi,\eta) has max(ξ,η)max(ξ,η)γ\max(\xi',\eta') \subseteq \max(\xi,\eta) \in \gamma, so ξ,ηγ\xi', \eta' \in \gamma, and the \lhd-initial segment of α×α\alpha \times \alpha below (ξ,η)(\xi,\eta) is contained in γ×γ\gamma \times \gamma.

L8
1.3

Base value: ω×ωω\omega \times \omega \approx \omega by [L5], so ω×ω=ω=ω\lvert \omega \times \omega \rvert = \lvert \omega \rvert = \omega by [L1] and [L4].

L1L4L5
1.4

Lower bound: for any ordinal μ\mu with 0μ0 \in \mu the map ξ(ξ,0)\xi \mapsto (\xi, 0) is an injection μμ×μ\mu \to \mu \times \mu, so μμ×μ\lvert \mu \rvert \le \lvert \mu \times \mu \rvert by [L3], and μ=μ\lvert \mu \rvert = \mu when μ\mu is a cardinal.

L1L3L9
2.1

Now let μ\mu be an infinite cardinal with ωμ\omega \in \mu, and assume the induction hypothesis that ν×ν=ν\lvert \nu \times \nu \rvert = \nu for every infinite cardinal νμ\nu \in \mu; for (ξ,η)μ×μ(\xi,\eta) \in \mu \times \mu and γ\gamma as in step 1.2 we have γμ\gamma \in \mu, because μ\mu is a limit ordinal by [L4] and max(ξ,η)μ\max(\xi,\eta) \in \mu, and moreover γ×γμ\lvert \gamma \times \gamma \rvert \in \mu: if γω\gamma \in \omega then γ×γ=γγωμ\lvert \gamma \times \gamma \rvert = \gamma \otimes \gamma \in \omega \subseteq \mu by [L4], while if ωγ\omega \subseteq \gamma then ν=γ\nu = \lvert \gamma \rvert satisfies ωνγμ\omega \le \nu \le \gamma \in \mu by [L1] and [L3], so ν\nu is an infinite cardinal in μ\mu and γ×γν×ν\gamma \times \gamma \approx \nu \times \nu by [L1] and [L2], whence γ×γ=ν×ν=νμ\lvert \gamma \times \gamma \rvert = \lvert \nu \times \nu \rvert = \nu \in \mu.

step 1.2L1L2L3L4
3.1

Under the same hypothesis, the \lhd-initial segment II of μ×μ\mu \times \mu below any (ξ,η)(\xi,\eta) has order type in μ\mu: Iγ×γI \subseteq \gamma \times \gamma by step 1.2 and II is well ordered by the restriction of \lhd by step 1.1, so Iγ×γμ\lvert I \rvert \le \lvert \gamma \times \gamma \rvert \in \mu by [L3] and step 2.1; and if the order type θ\theta of II satisfied μθ\mu \subseteq \theta then μθI\mu \preceq \theta \approx I by [L6] and [L9], giving μI\mu \le \lvert I \rvert by [L3], which contradicts Iμ\lvert I \rvert \in \mu.

step 1.1step 1.2step 2.1L3L6L9
4.1

Under the same hypothesis, μ×μ=μ\lvert \mu \times \mu \rvert = \mu: let δ\delta be the order type of (μ×μ,)(\mu \times \mu, \lhd) and g:δμ×μg : \delta \to \mu \times \mu the inverse of the collapsing isomorphism ([L6]); if μδ\mu \in \delta then gg carries the initial segment of δ\delta below μ\mu, which is μ\mu itself, onto the \lhd-initial segment below g(μ)g(\mu), whose order type would then be μ\mu, contradicting step 3.1; so δμ\delta \subseteq \mu and μ×μ=δδμ\lvert \mu \times \mu \rvert = \lvert \delta \rvert \le \delta \le \mu by [L1], while step 1.4 gives μμ×μ\mu \le \lvert \mu \times \mu \rvert.

step 3.1step 1.4L1L6L8
5.1

Apply [L7] to the well-order (κ{κ},)(\kappa \cup \{\kappa\}, \in) of [L8] and to S={μκ{κ}:μS = \{\mu \in \kappa \cup \{\kappa\} : \mu is not an infinite cardinal, or μ×μ=μ}\lvert \mu \times \mu \rvert = \mu\}: a μ\mu below which everything lies in SS is in SS, trivially if μ\mu is not an infinite cardinal, by step 1.3 if μ=ω\mu = \omega, and by step 4.1 otherwise; hence S=κ{κ}S = \kappa \cup \{\kappa\}, so κS\kappa \in S and κκ=κ×κ=κ\kappa \otimes \kappa = \lvert \kappa \times \kappa \rvert = \kappa.

step 1.3step 4.1L7L8

Remarks

Why the maximum comes first. Under the plain lexicographic order of Disjoint union, cartesian product, function space and power set respect equinumerosity, and for ordinals α,β\alpha, \beta the sets αβ\alpha \sqcup \beta and α×β\alpha \times \beta carry explicit well-orders, so their cardinalities exist in ZF, the initial segment below (1,0)(1,0) in ω×ω\omega \times \omega is the whole of {0}×ω\{0\} \times \omega, which is already infinite; the order type of ω×ω\omega \times \omega is then ωω\omega \cdot \omega, far above ω\omega. Ordering by the maximum first bounds every initial segment inside a square γ×γ\gamma \times \gamma with γ<κ\gamma < \kappa, and the induction hypothesis then says that square is small. The whole proof is that one change of order.

No choice, and it is worth saying why. Every place that invites a selection avoids it: the \lhd-least element of a nonempty set is found by three successive minimisations, the order type of a well-order is unique, and the bijection γγ\gamma \approx \lvert \gamma \rvert is used only through Disjoint union, cartesian product, function space and power set respect equinumerosity, and for ordinals α,β\alpha, \beta the sets αβ\alpha \sqcup \beta and α×β\alpha \times \beta carry explicit well-orders, so their cardinalities exist in ZF, which quantifies over existing bijections rather than picking one for each γ\gamma at once.

What the theorem does not say. It is about cardinals, that is about well-orderable sets. "Every infinite set AA satisfies A×AAA \times A \approx A" is a strictly stronger statement, and Tarski: the Axiom of Choice is equivalent to the statement that A×AAA \times A \approx A for every infinite set AA, so extending Hessenberg's theorem from the alephs to arbitrary sets is exactly as strong as choice shows it is equivalent to the Axiom of Choice. So Hessenberg's theorem is not a weaker version of Tarski's with a cheaper proof; it is the exact fragment that ZF proves.

Depends on

Used by

Dependency tree · next 3 levels

Direct dependencies and their dependencies through the next three levels: 108 results over 34 levels. An arrow runs from a result to what uses it, and this result sits at the bottom with a heavier outline. Click the chart to enlarge it.

Sources