Why √2 is Irrational — The Prime-Parity Proof You Can See

The classical proof that \sqrt{2} is irrational runs through fractions in lowest terms and chases contradictions. It works, but it hides the reason why. Here is a visual proof that cuts straight to the heart: count prime factors. If you count the factors of 2 in p^2 and the factors of 2 in 2q^2, one count is always even and the other is always odd. An even number can never equal an odd number. The equation p^2 = 2q^2 has no integer solution. Therefore \sqrt{2} cannot be written as p/q.

The beauty of the prime-parity proof is that it generalises instantly. The same argument shows that \sqrt{n} is irrational whenever n is not a perfect square. No case analysis, no "assume p is even, then q must be even too, contradiction." Just: count the factors, compare parity, done.

The setup

Suppose — for the sake of argument — that \sqrt{2} is rational. Then there exist positive integers p and q with

\sqrt{2} = \frac{p}{q}

Squaring both sides:

2 = \frac{p^2}{q^2} \implies p^2 = 2q^2

The entire proof now comes down to analysing this single equation. If p^2 = 2q^2 is impossible for any positive integers p, q, then the assumption that \sqrt{2} is rational was wrong, and \sqrt{2} is irrational.

The key move: count the factors of 2

Every positive integer n has a unique prime factorisation (the Fundamental Theorem of Arithmetic). Define v_2(n) — the 2-adic valuation of n — as the exponent of 2 in that factorisation. That is, if n = 2^k \cdot m with m odd, then v_2(n) = k.

Some examples:

v_2(12) = 2 because 12 = 2^2 \cdot 3.
v_2(40) = 3 because 40 = 2^3 \cdot 5.
v_2(17) = 0 because 17 is odd (no factor of 2).
v_2(2^{10}) = 10.

Two properties of v_2 that we will use:

Multiplicative. v_2(ab) = v_2(a) + v_2(b). Why: write a = 2^{v_2(a)} \cdot m_a and b = 2^{v_2(b)} \cdot m_b with m_a, m_b odd. Then ab = 2^{v_2(a) + v_2(b)} \cdot (m_a m_b), and m_a m_b is odd (odd times odd is odd). So the exponent of 2 in ab is v_2(a) + v_2(b).
Squaring doubles the count. v_2(n^2) = 2 \cdot v_2(n), which is always even.

That second property is the whole proof.

Counting both sides

Take the equation p^2 = 2q^2 and apply v_2 to both sides.

Left side: v_2(p^2) = 2 \cdot v_2(p). This is even — a multiple of 2.

Right side: v_2(2q^2) = v_2(2) + v_2(q^2) = 1 + 2 \cdot v_2(q). This is odd — one more than a multiple of 2.

But p^2 and 2q^2 are the same integer, and an integer has exactly one prime factorisation, so v_2 of that integer has exactly one value. It cannot be simultaneously even and odd.

Contradiction. Therefore no positive integers p, q satisfy p^2 = 2q^2, and \sqrt{2} is irrational. \blacksquare

The parity picture

p = 12: q = 7:

Pick any $p$ and $q$. The top (red) bar shows $v_2(p^2) = 2 v_2(p)$ as paired blocks \u2014 always even. The bottom (green) bar shows $v_2(2q^2) = 1 + 2 v_2(q)$ as pairs plus one orange unpaired block \u2014 always odd. Even $\\neq$ odd, so $p^2 = 2q^2$ has no solution.

Why "pairs" and "pair plus one" is the whole story

Squaring any integer doubles each prime's exponent. If p = 2^3 \cdot 5 \cdot 7, then p^2 = 2^6 \cdot 5^2 \cdot 7^2 — every exponent doubled. Doubling preserves evenness and guarantees it: starting from any non-negative integer v_2(p), the doubled value 2 v_2(p) is even.

Now look at the right side, 2q^2. The q^2 contributes an even count of 2s, and the leading factor of 2 throws one more into the pile. Even plus one is odd.

The picture is of pairs versus pair-plus-one. On one side, factors of 2 come in perfect pairs (from squaring). On the other, the same pairs appear plus one unpaired extra. You cannot match pairs to pair-plus-one by any rearrangement. This is the same parity argument used in the classic checkerboard-with-corners-removed puzzle, where you cannot tile a board with dominoes if an odd number of squares of one colour remain.

The proof generalises

Replace 2 with any prime p. The same argument gives:

\sqrt{p} \text{ is irrational for every prime } p

Why: if \sqrt{p} = a/b, then a^2 = p \cdot b^2. Take v_p of both sides. Left: v_p(a^2) = 2 v_p(a), always even. Right: v_p(p b^2) = 1 + 2 v_p(b), always odd. Even ≠ odd. Contradiction.

More generally, \sqrt{n} is irrational whenever n is a positive integer that is not a perfect square. Pick any prime p whose exponent in n's factorisation is odd. Then the same parity argument applies to v_p of both sides of a^2 = n b^2.

This is why only a sparse set of integers have rational square roots: the perfect squares. Every other positive integer has an irrational square root, guaranteed by prime-exponent parity.

Why the classical proof secretly used parity too

The "classical" proof of \sqrt{2}'s irrationality — "if \sqrt{2} = p/q in lowest terms, then p^2 = 2q^2, so p is even, so p^2 is divisible by 4, so q^2 is even, so q is even, so both are even, contradicting lowest terms" — is a special case of the prime-parity proof. Every step is the v_2 bookkeeping, done in words rather than in symbols. The prime-parity argument is the same proof, stated in the language that makes it generalise.

You can see why mathematicians prefer the valuation form. It turns a two-paragraph case-chase into two lines of algebra, and it works for any prime — not just 2.

A note on what "irrational" actually says

Irrationality is a statement about representability, not about existence. The number \sqrt{2} exists: it is the length of the diagonal of a unit square, which you can draw with a ruler and compass. What is impossible is writing \sqrt{2} as a ratio of integers. The decimal expansion never terminates and never repeats. The prime-parity proof shows that this non-representability is not a lucky accident of notation — it is forced by the multiplicative structure of the integers themselves.