Proof by Contradiction: What It Is and How It Works

How do you prove something is true without ever showing it directly?

Sometimes you can't build the thing you're trying to prove, piece by piece, in a straight line. There's no picture to draw, no formula that walks you there. So mathematicians reach for a strange trick instead: they assume the claim is false, follow that assumption honestly, and let it destroy itself.

That's proof by contradiction. It sounds backwards, and it kind of is. But once you see it work, you'll notice it everywhere.

Why would assuming the wrong thing ever help?

Think like a detective for a second. A suspect claims they were at the harbor all night, never left. You don't have a video of them committing the crime. So instead, you take their story at face value and follow it.

The harbor's checkout log shows their badge scanning out at 8:52pm. The theater's ticket scanner shows the same badge entering at 8:55pm. Three minutes to cover forty minutes of road...

That's not evidence against the alibi. That's the alibi collapsing under its own weight. The suspect's own story now demands they be in two places at once, and that's impossible. So the alibi has to be false, and you never had to prove where they actually went.

That's the whole shape of proof by contradiction: assume the opposite of what you want, follow it honestly, and if it forces something impossible, the assumption was never true.

Step through the chain above. Notice that every link is forced by the one before it. Nobody chose to make the alibi fail. It failed because two true facts, the badge times and the distance, couldn't both hold alongside it.

We write this idea as: assume $\lnot P$ (read "not P"). If $\lnot P$ forces a contradiction, we write that as $\bot$. And once you've reached $\bot$, you're allowed to conclude $P$.

What does this look like with numbers instead of alibis?

Here's the cleanest version of the trick: there is no largest whole number.

You probably already believe that. But watch what happens if we refuse to believe it, and actually try to find one.

Pick any candidate. The number just past it is always waiting: your number, plus one. Try the leap button and jump to something enormous. It doesn't matter how big your candidate is, adding 1 to it always produces something bigger. So "N is the largest number" can't survive contact with N+1, for any N you choose.

We write it like this: assume some $N$ is the largest integer. Then $N+1$ is also an integer, and $N + 1 > N$. Contradiction. So no largest integer exists, and that holds for every candidate at once, not just the ones we tried.

Does the same trick work in the other direction, toward zero?

Now try a mirror version: there is no smallest positive fraction.

Say you pick some tiny positive number $q$ and claim it's the smallest positive rational there is. Well... what about $q/2$?

Keep halving. Every single time, $q/2$ is still positive, and still smaller than $q$. There's no floor. No matter how far you zoom toward zero, there's room for one more fraction squeezed in underneath your "smallest" one. The claim doesn't just fail once, it fails at every scale, forever, which is exactly why "smallest positive rational" was never a real thing to begin with.

Why does hitting a contradiction actually prove anything?

This is worth slowing down on, because it isn't a cute coincidence. It's a law.

A statement and its opposite can't both be true at the same time. Not "usually can't", can't, ever. That's called the law of non-contradiction, and it's one of the oldest rules in logic. If your reasoning ever forces a fact to be both true and false at once, you haven't stumbled onto some exotic new possibility. You've proven that one of your starting assumptions was wrong.

Watch what happens when you run the assumption. The point gets pulled toward a spot that would need to belong to both regions at once, a spot that, by definition, cannot exist. That's not a visual trick. It's the actual mechanism. Reaching an impossible state doesn't mean logic is broken. It means your assumption was. Reject the assumption, and whatever you were trying to prove in the first place is left standing.

So when should you actually reach for this?

Direct proofs are usually cleaner when you can get them: build the thing, step by step, arrive at the answer. But some claims resist that. "There's no largest prime." "$\sqrt2$ isn't a fraction." "No finite list of these things can ever be complete." Claims about nonexistence, about something never happening no matter how far you search, are exactly where contradiction shines. You can't walk through infinitely many cases one at a time. But you can assume they all fail to break the claim, and show that assumption self-destructs instead.

The short version

Proof by contradiction assumes the opposite of what you want, follows that assumption wherever it honestly leads, and treats an impossible outcome as proof the assumption was wrong all along. It works because a fact can't be both true and false at once, so if your reasoning ever forces that, the fault was never in reality, it was in your starting guess.

Assume the false thing. Let it talk. It'll trip over itself eventually.