Infinity comes in sizes.
Georg Cantor's 1891 diagonal argument is the cleanest proof in mathematics that some infinities are strictly larger than others. List the reals between 0 and 1 however you like; then read straight down the diagonal and build a number nobody listed. Cantor 1874, 1891 — the moment set theory left the nursery.
Infinity is not one thing. ℵ₀ counts the naturals 1, 2, 3, …; the real line carries a strictly larger infinity. Cantor turned the gut feeling «the line is denser than the integers» into a proof.
Suppose you list reals in [0,1] as r₁ = 0.1415…, r₂ = 0.2718…, r₃ = 0.5772…, and so on. Walk down the diagonal — pick a different digit at each step — and you have a real that disagrees with every row by construction.
Cantor's argument seeded modern set theory, exposed Russell's paradox a decade later, and gave Gödel and Turing the diagonal trick that drives incompleteness and the halting problem. Modern logic stands on it.
The naturals are the smallest infinite set. Their size is named ℵ₀ — aleph-null — and any set you can label with naturals (the integers, the rationals, all finite strings) has the same size. Countable means «small enough to enumerate».
Assume, for contradiction, the reals in [0,1] are countable: a sequence r₁, r₂, r₃, … containing every real exactly once. The order is yours to choose. Write each row as a decimal expansion 0.d_{n,1} d_{n,2} d_{n,3} … and lay them down as an infinite table.
Read straight down the diagonal: d_{1,1}, d_{2,2}, d_{3,3}, …. Build a new real s whose n-th digit s_n differs from d_{n,n} for every n. A safe recipe swaps each digit for one that isn't 0 or 9 (avoiding the 0.999… = 1.000… ambiguity). By construction s ≠ r_n at the n-th decimal, for every n — yet s is a perfectly good real in [0,1].
| r1 | 0. | 4 | 1 | 4 | 9 | 4 | 7 | 0 | 1 | … |
| r2 | 0. | 6 | 7 | 0 | 1 | 4 | 9 | 8 | 7 | … |
| r3 | 0. | 8 | 9 | 8 | 7 | 8 | 1 | 2 | 3 | … |
| r4 | 0. | 6 | 9 | 6 | 1 | 8 | 5 | 4 | 3 | … |
| r5 | 0. | 4 | 5 | 6 | 9 | 6 | 5 | 6 | 7 | … |
| r6 | 0. | 6 | 3 | 0 | 3 | 8 | 7 | 0 | 5 | … |
| Constructed real | ||||||||||
| s | 0. | 5 | 8 | 1 | 2 | 7 | 8 | · | · | … |
The amber cells are the digits Cantor reads down the diagonal. The rose row beneath is the new real s — each digit chosen so s differs from row n at the n-th place. No matter which list you tried, s is not in it.
The list was supposed to contain every real, and we just built one it missed. The assumption collapses: no enumeration of the reals can be complete. The reals have cardinality c = 2^ℵ₀, strictly larger than ℵ₀. A new kind of infinity, hand-delivered.
Strictly larger than ℵ₀, the size of the naturals. The first uncountable infinity mathematics actually hands you.
Cantor (1878) asked whether any cardinality lives strictly between ℵ₀ and c. Gödel (1940) showed CH cannot be disproved from ZFC; Cohen (1963, Fields medal 1966) showed CH cannot be proved from ZFC. The question is independent of the axioms — formally undecidable, and yet the question remains.
| Year | Who | Result |
|---|---|---|
| 1878 | Cantor | Asks whether any size lives strictly between ℵ₀ and c. |
| 1891 | Cantor | Diagonal argument: |ℝ| > |ℕ|. |
| 1940 | Gödel | CH cannot be disproved from ZFC (constructible universe L). |
| 1963 | Cohen | CH cannot be proved from ZFC (forcing). Fields medal 1966. |
The continuum hypothesis is independent of ZFC: the axioms simply don't decide it.
Cantor's theorem says |𝒫(X)| > |X| for every set X — the power set is always strictly larger. Iterating gives an unending ladder of ever-larger infinities: ℵ₀, 2^ℵ₀, 2^(2^ℵ₀), … There is no biggest one.
Cantor's theorem: every set is strictly smaller than its power set. Iterating gives an unending ladder. Click a rung to read what it is.
Click any rung to expand.
Pick your own listing of reals, watch the diagonal pull a new number out of it, and toggle the digit-swap rule. The contradiction is alive on screen.
→ Open the ExplorerHow and where this technique lives in the world today.
Topics in the same vein.
