The first great shock in Greek mathematics was not a complicated theorem, but a simple square. If the side has length one, the diagonal has length \(\sqrt 2\), and no pair of whole numbers measures both side and diagonal exactly. A culture that trusted whole-number ratios suddenly had to separate arithmetic from geometry.
This page follows the crisis from commensurable magnitudes to Euclid's theory of proportion. The point is not only that \(\sqrt2\) is irrational; the deeper lesson is that mathematics advances when a familiar language becomes too small for a phenomenon it has already discovered.
Commensurability and magnitude
Definition. Two positive magnitudes \(a\) and \(b\) are commensurable if there exists a common unit \(u\) and integers \(m,n\) such that \(a=mu\) and \(b=nu\). In modern notation this is equivalent to \(a/b\in\mathbb Q\).
For counting objects, whole numbers are enough. For comparing lengths, areas, and diagonals, the same intuition fails: a geometric magnitude can be perfectly well defined even when no rational number names its ratio to a chosen unit.
The diagonal argument
Theorem. The diagonal of the unit square is incommensurable with the side; equivalently, \(\sqrt2\notin\mathbb Q\).
Proof. Suppose \(\sqrt2=m/n\) with \(m,n\) coprime positive integers. Then \(m^2=2n^2\), so \(m^2\) is even and hence \(m\) is even. Write \(m=2k\). Then \(4k^2=2n^2\), so \(n^2=2k^2\), and \(n\) is even. This contradicts coprimality.
Eudoxus and Euclid's repair
The Greek response was subtle. Instead of forcing every magnitude to become a number, Eudoxus developed a theory of proportion that compares ratios through inequalities among multiples. Euclid's Book V records this move: ratios of magnitudes are understood by how their multiples dominate one another.
Proposition. Eudoxus-style comparison anticipates the modern order-theoretic idea behind real numbers: a magnitude is determined by the rational comparisons it makes true. This is the ancestor of Dedekind's later cut construction.
From crisis to method
The proof of incommensurability also sharpened the role of contradiction. A geometric fact was established by assuming a numerical representation and showing that the representation destroys itself. This pattern becomes one of the durable habits of mathematics: when direct construction is unavailable, impossibility can still be proved with complete force.
The longer vein runs from Greek magnitude theory to modern completeness. The rational numbers are dense but incomplete; between any two rationals there are more rationals, yet Cauchy sequences and cuts reveal missing limits. The first crisis therefore anticipates the later need to distinguish density from completeness.
From Greek proportions to completeness
The traditional story associates the discovery with the Pythagorean school; Theaetetus studied classes of irrational magnitudes; Eudoxus supplied the proportion theory that protected geometry from collapse; Euclid organized the repair into an axiomatic monument. Much later, Dedekind turned the same pressure into a construction of \(\mathbb R\).
The diagonal that escaped arithmetic
The discovery attached by later tradition to the Pythagorean school was not simply that \(\sqrt2\) has an inconvenient decimal expansion. Greek arithmetic treated numbers as ratios of whole numbers, while geometry treated lengths as magnitudes. The diagonal of a unit square forced those languages apart: geometry produced the segment immediately, but no pair of integers could measure it exactly.
A proof in the language of parity. Suppose \(\sqrt2=m/n\) in lowest terms. Then \[ m^2=2n^2, \] so \(m^2\), hence \(m\), is even. Write \(m=2k\). Then \[ n^2=2k^2, \] so \(n\) is even as well, contradicting lowest terms. The contradiction is finite and exact; it does not depend on decimals or approximation.
Euclid's response in Book X of the Elements was not to force every magnitude into arithmetic. He developed a theory of commensurable and incommensurable magnitudes, while Eudoxus's theory of proportion compared ratios without assuming that they were numbers in the later real-number sense. This was an early example of mathematical progress by enlarging the language while preserving rigor.
From geometric magnitude to the real line
For nearly two millennia, irrationality remained geometrically secure but arithmetically unfinished. Decimal expansions, algebraic equations, and calculus made irrational numbers indispensable, yet a definition of the full continuum arrived only in the nineteenth century. Dedekind cuts and Cantor's Cauchy-sequence construction finally made every point of the line an arithmetic object.
The episode established a recurring pattern. A contradiction does not necessarily destroy a subject; it can reveal that familiar objects were being asked to carry more structure than their definitions allowed. The repair is then conceptual: distinguish number from magnitude, or later enlarge the number system with a completeness axiom.