Why Is the Square Root of 2 Irrational? The Classic Proof

Can you write $\sqrt2$ exactly as a fraction? Not a close one like $1.41421356$. An exact one, some whole number divided by another whole number, that lands on $\sqrt2$ dead on.

You can't. And this isn't "nobody's found it yet." It's proven impossible, and it has been for about 2,500 years.

The story goes that this exact fact broke a group of ancient Greek mathematicians called the Pythagoreans, who believed every length in the universe could be measured as a ratio of whole numbers. Then someone in their own school proved that the diagonal of a simple square couldn't be. Legend says they drowned him for it. True or not, the discomfort was real: this proof cracked open a hole in numbers that took centuries to properly understand.

So let's build the proof ourselves, piece by piece, until you can see exactly why $\sqrt2$ refuses to be a fraction.

Can't we just try really hard to find one?

Before reaching for a proof, it's worth asking the obvious question: what if we just search? Pick a denominator $b$, find the whole number $a$ that gets $a/b$ closest to $\sqrt2$, and check how close $(a/b)^2$ lands to 2.

Slide through different values of $b$, or hit scan and watch them all go by. Notice the bars never touch that dashed line. $a^2 - 2b^2$ jumps around, sometimes close to zero, sometimes not, but it never actually lands on it. No matter how far you search, this quantity refuses to hit exactly 0.

That's suggestive, but it's not proof. We checked 20 fractions. There are infinitely many. To actually settle this, we need an argument that covers every possible $a$ and $b$ at once, not just the ones we tried.

What does "lowest terms" actually buy us?

Here's the strategy: assume $\sqrt2$ IS a fraction, $\sqrt2 = a/b$, and follow that assumption until it breaks.

But there's a sneaky move we get to make first. Any fraction, no matter how bloated, can be reduced by dividing out whatever numerator and denominator have in common. $84/60$ reduces to $7/5$. $168/120$ also reduces to $7/5$. So if $\sqrt2$ really does equal some fraction, it equals one written in lowest terms too, where the top and bottom share no common factor at all.

Watch what happens as you cancel the shared blocks. No matter how big $k$ gets, there's always a bottom layer left standing where nothing more can be divided out. We're allowed to assume $a$ and $b$ share no common factor, because every fraction can be pushed down to a version that doesn't. We write that as $\gcd(a, b) = 1$.

That one assumption is about to do all the work.

Why does $a^2$ being even force $a$ to be even?

We're going to need one more tool before the trap can close: a fact about even and odd numbers.

Claim: if $a^2$ is even, then $a$ itself has to be even. Watch it hold across a whole range of numbers before we say why.

Hover across the strip. The color of $n$ on top always matches the color of $n^2$ below it, every single time. Here's why it has to. If $n$ is even, write $n = 2m$. Then $n^2 = 4m^2 = 2(2m^2)$, which is 2 times something, so it's even. If $n$ is odd, write $n = 2m+1$. Then $n^2 = 4m^2+4m+1 = 2(2m^2+2m)+1$, which is 2 times something plus 1, so it's odd. Even numbers square to even numbers, odd numbers square to odd numbers, and there's no third option. So if you know $a^2$ is even, $a$ can't be odd. It has to be even.

Now we have both pieces: the lowest-terms assumption, and this parity lemma. Time to put them together.

So what actually happens when we assume $\sqrt2 = a/b$?

Here's the whole proof, laid out one forced step at a time.

Step through it slowly. Every box follows from the one before it, no leaps, no guessing. We assumed $\sqrt2 = a/b$ in lowest terms. That forced $a^2 = 2b^2$. That forced $a$ to be even. Writing $a = 2k$ and substituting back forced $b$ to be even too. And now both $a$ and $b$ are even, which means they share the factor 2, which flatly contradicts $\gcd(a,b) = 1$ from step one.

We didn't choose that contradiction. It was forced, the same way N+1 is forced to beat any candidate for "largest number." The only assumption we made anywhere in this chain was that $\sqrt2$ could be written as a fraction in the first place. Since everything after that was airtight, that's the assumption that has to give.

There is no pair of whole numbers $a$ and $b$ with $\sqrt2 = a/b$. $\sqrt2$ is irrational.

Why does this matter beyond one square root?

This isn't just a fact about $\sqrt2$. It's the moment numbers stopped being simple.

Before this proof, "number" basically meant "whole number, or a ratio of whole numbers." That covers a lot of ground: every measurement you could physically construct with a ruler and a compass, everyone assumed, would land on some fraction eventually. $\sqrt2$ is not exotic. It's just the diagonal of a unit square, about as basic a shape as geometry has. And it sits completely outside the world of fractions.

The same proof structure works for $\sqrt3$, $\sqrt5$, and most square roots of whole numbers that aren't perfect squares. Once you see it here, you can rebuild it anywhere a "no fraction can do this" claim shows up: assume it can, force out a property that can't coexist with the setup, and the assumption collapses.

The short version

Assume $\sqrt2 = a/b$ in lowest terms, meaning $a$ and $b$ share no common factor. Squaring gives $a^2 = 2b^2$, so $a^2$ is even, which forces $a$ to be even. Write $a = 2k$ and substitute back in, and $b^2$ turns out even too, which forces $b$ to be even as well. But if $a$ and $b$ are both even, they share the factor 2, contradicting the fact that we started with no common factor at all.

The assumption breaks. $\sqrt2$ can't be written as a fraction. It's irrational, and it always was.

All visualizations are interactive React components running entirely in your browser, built with plain SVG. No libraries beyond React.