Transfinite Numbers

Cantor’s extension of counting and ordering beyond the finite: the transfinite Ordinals (order types of well-orderings: ω, ω+1, …) and transfinite cardinals (sizes: ℵ₀, ℵ₁, …). Introduced to make precise the idea that infinite collections can be compared and arithmetic done on them. Given canonical set-theoretic form by the von Neumann ordinals.

In this vault

Backlinks

Linked Pages

Set Theory

The branch of mathematical logic that treats sets — arbitrary collections — as the basic objects from which all other mathematical structures can be built, and the de facto foundation of mathematics. Founded by Cantor, who showed (via Diagonalization) that infinite sets come in distinct, unboundedly increasing sizes (Cardinality, Cantor’s Theorem). Cantor’s “naive” theory permitted unrestricted comprehension and so fell to Russell’s Paradox; Zermelo’s axiomatization replaced comprehension with Separation, yielding the system that — strengthened with Replacement, Foundation, and Choice — became ZFC, the standard foundation. Ordinals and cardinals are realised concretely as sets (the von Neumann ordinals), indexing the Cumulative Hierarchy. The discipline’s deepest results are limitative: the Continuum Hypothesis and the Axiom of Choice are independent of ZFC (Gödel’s constructible universe + Cohen’s Forcing), the set-theoretic counterpart to Gödel incompleteness.

In this vault

Über eine elementare Frage der Mannigfaltigkeitslehre — Cantor 1891 (diagonal argument, Cantor’s theorem)
Untersuchungen über die Grundlagen der Mengenlehre I — Zermelo 1908 (axiomatization)
Zur Einführung der transfiniten Zahlen — von Neumann 1923 (ordinals as sets)
The Independence of the Continuum Hypothesis — Cohen 1963/64 (forcing, independence)
On Notation for Ordinal Numbers — Church & Kleene (constructive ordinals)
Über formal unentscheidbare Sätze der Principia Mathematica und verwandter Systeme I — Gödel 1931 (the limitative sibling)

Zur Einführung der transfiniten Zahlen

Reference: von Neumann, John (1923). “Zur Einführung der transfiniten Zahlen.” Acta Litterarum ac Scientiarum Regiae Universitatis Hungaricae Francisco-Josephinae, Sectio Scientiarum Mathematicarum (Szeged) 1, 199–208. English (“On the Introduction of Transfinite Numbers”) in van Heijenoort, From Frege to Gödel, Harvard 1967, 346–354.

Summary

The paper that gives ordinals a concrete set-theoretic identity. Cantor had introduced ordinals as order types — abstract equivalence classes of well-ordered sets sharing the same order structure — which left them mathematically real but representationally vague. Writing at twenty, von Neumann proposes instead to pick a canonical representative for each: an ordinal simply is the set of all ordinals below it. So 0 = ∅, 1 = {∅}, 2 = {∅, {∅}}, 3 = {∅, {∅}, {∅, {∅}}}, …, and the first infinite ordinal ω = {0, 1, 2, …}. Each ordinal is a transitive set well-ordered by the membership relation ∈; “less than” is “is a member of.” These are the von Neumann ordinals, still the standard definition.

The construction does real work beyond tidiness. Because each ordinal is built explicitly from its predecessors as sets, ordinals can be defined within axiomatic set theory rather than presupposed, and von Neumann justifies definition by transfinite recursion — defining a function on all ordinals by specifying its value at α in terms of its values below α — which Cantor had used but not rigorously grounded. Von Neumann’s later 1925/1928 axiomatization carries the idea further, distinguishing classes from sets (the precursor of NBG/von Neumann–Bernays–Gödel set theory) and proposing a limitation-of-size principle and the axiom of foundation, under which all sets sit in the cumulative hierarchy V indexed by exactly these ordinals.

For this vault the paper closes a loop. It is the classical, set-theoretic account of the ordinals whose constructive, computable counterpart is the Church–Kleene system of ordinal notations — the same objects approached from the standpoint of what a machine can name and recurse over. And it is of a piece with von Neumann’s lifelong habit, visible across his work here, of finding the minimal concrete structure that lets an informal mathematical notion be manipulated rigorously.

Key Ideas

Von Neumann ordinals: each ordinal is the set of all smaller ordinals; ∈ is the well-ordering, so “<” = “∈”.
Canonical representatives: replace Cantor’s abstract order types with concrete sets — 0 = ∅, n+1 = n ∪ {n}, ω = {0,1,2,…}.
Ordinals internal to set theory: definable from the axioms rather than assumed; transitive sets well-ordered by membership.
Transfinite recursion rigorously justified: define f(α) from {f(β) : β < α}.
Classes vs. sets (in the 1925 sequel): the NBG lineage, limitation of size, axiom of foundation.
Cumulative hierarchy V is indexed by the ordinals so defined.

Connections

Conceptual Contribution

Claim: Every ordinal can be canonically realised as a set — the set of its predecessors — so ordinals become objects definable inside axiomatic set theory rather than abstract order types taken on faith.
Mechanism: Define an ordinal as a transitive set well-ordered by ∈; build them by α+1 = α ∪ {α} and unions at limits; ground definition-by-transfinite-recursion on this representation.
Concepts introduced/used: Ordinals, Transfinite Numbers, Cumulative Hierarchy, Set Theory
Stance: foundational construction paper (set theory)
Relates to: Gives the classical set-theoretic ordinals whose computable analogue is Church–Kleene; supplies the ordinal scaffolding for the cumulative hierarchy built on Zermelo’s axioms; shares its author with Theory of Self-Reproducing Automata and Probabilistic Logics and the Synthesis of Reliable Organisms from Unreliable Components.

Über eine elementare Frage der Mannigfaltigkeitslehre

Reference: Cantor, Georg (1891). “Über eine elementare Frage der Mannigfaltigkeitslehre.” Jahresbericht der Deutschen Mathematiker-Vereinigung 1, 75–78. English (“On an Elementary Question of the Theory of Manifolds”) in Ewald, From Kant to Hilbert, vol. 2, Oxford 1996.

Summary

The four-page note in which Cantor introduces the diagonal argument and, with it, the modern proof that there are different sizes of infinity. Having already shown in 1874 (by a nested-intervals argument) that the real numbers cannot be enumerated, Cantor here gives a far simpler and more general method. He considers the collection of all infinite sequences over two symbols (he uses m and w) — equivalently, all functions from the natural numbers into a two-element set. Assume such sequences could be listed as E₁, E₂, E₃, …. Define a new sequence E₀ whose n-th entry is chosen to differ from the n-th entry of Eₙ (flip m↔w down the diagonal). Then E₀ differs from every Eₙ in at least one place, so it is absent from the list — contradiction. The collection of such sequences is therefore not enumerable.

Cantor immediately generalises. For any set M, the totality of functions from M into a two-element set — i.e. the power set of M — has a strictly greater cardinality than M itself. This is Cantor’s theorem: |M| < |𝒫(M)|. Its consequence is that there is no largest cardinality and no “set of all sets” of maximal size; instead the transfinite cardinalities form an unending ascending hierarchy. The argument is purely combinatorial — no analysis, no topology of the line — which is exactly why it became the template for every later self-referential limitative result.

The diagonal method is the paper’s lasting bequest. Gödel’s incompleteness construction, Turing’s and Church’s undecidability proofs, Russell’s paradox, and Tarski’s undefinability of truth are all recognisably the same move: build an object that disagrees, along the diagonal, with every member of a purported complete enumeration. The 1891 note is where that move is first made explicit and abstracted away from the particular case of the reals.

Key Ideas

The diagonal argument: from any enumeration of sequences, construct one differing from the n-th at position n; it cannot be in the list.
Uncountability of the set of infinite binary sequences (hence of the reals, of 𝒫(ℕ)).
Cantor’s theorem: for every set M, |M| < |𝒫(M)| — the power set is strictly larger.
No largest cardinal: transfinite cardinalities form an unbounded ascending hierarchy.
Combinatorial, not analytic: the proof needs nothing about the continuum beyond counting, making it general.
Template for self-reference: the structural seed of Gödel, Turing, Church, Russell, Tarski.

Connections

Conceptual Contribution

Claim: For any set M, there are strictly more subsets of M than elements of M; hence infinite sets come in distinct, unboundedly increasing sizes.
Mechanism: Suppose an enumeration of all sequences/subsets exists; diagonalise to manufacture one that differs from each listed item at its own index; the manufactured item is therefore unlisted — contradiction.
Concepts introduced/used: Diagonalization, Cardinality, Power Set, Cantor’s Theorem, Transfinite Numbers
Stance: foundational theorem paper (set theory / foundations of mathematics)
Relates to: Furnishes the diagonal method later turned into a limitative weapon by Gödel 1931 and Church 1936; the cardinality questions it opens (how large is 𝒫(ℕ)?) become the Continuum Hypothesis, settled as independent by Cohen 1963.

Ordinals

Numbers that generalise “position in a well-ordering” into the transfinite (0, 1, …, ω, ω+1, …). In modern set theory each ordinal is realised as a von Neumann ordinal — the set of all smaller ordinals, well-ordered by ∈ — so that “<” is literally “∈”. Ordinals index transfinite recursion and the Cumulative Hierarchy. Their constructive, computable fragment is the subject of the Church–Kleene ordinal-notation system.