Lowest-Terms Contradiction: The Finisher on Every Irrationality Proof

Read almost any textbook proof that some number is irrational and you will notice a strange déjà vu. The setup is always "assume \sqrt{2} = p/q in lowest terms." The ending is always "but then p and q share a common factor — contradiction." The middle is where the algebra varies, but the opening and the finisher are fixed. This satellite is about the finisher. Once you see why "both p and q are divisible by the same prime" is the only place these proofs can land, you can steer toward that finish line in any new irrationality problem.

The shape

A number x is rational if and only if it can be written as p/q with integers p, q, q \neq 0, and — crucially — you can insist \gcd(p, q) = 1. The "lowest terms" assumption is the one you drive the contradiction through.

So the opening is: assume x = p/q with \gcd(p, q) = 1.

And the finisher is: show p and q share a prime factor.

These two statements cannot both be true. So the opening was wrong. So x is not rational. So x is irrational.

Why "same factor" is the natural contradiction

The assumption you made — "lowest terms, \gcd(p, q) = 1" — is a statement about shared factors: there are none, beyond \pm 1. The only way to contradict a no-shared-factor statement is to produce a shared factor. Anything else — p and q being big, being negative, being unequal — is irrelevant, because the assumption didn't constrain those things.

So when you are steering an irrationality proof, you are steering toward a line that reads "\therefore k \mid p and k \mid q" for some prime k. That is the only shape the ending can take.

The canonical finish

Look at the tail of the \sqrt{2} proof with a magnifying glass.

You derived p^2 = 2q^2.
Therefore p^2 is even.
Therefore p is even (because odd squared = odd).
So p = 2k.
Substituting, 4k^2 = 2q^2, hence q^2 = 2k^2, hence q^2 is even, hence q is even.
So p and q are both divisible by 2. But \gcd(p, q) = 1. Contradiction.

The underlined ingredients:

A prime in the algebra (here, 2, which came from p^2 = 2q^2).
A divisibility rule for that prime ("prime divides square \Rightarrow prime divides base"). This is Euclid's lemma for primes.
Symmetry: the same argument that shows p is divisible by the prime also shows q is.

All three are always present in a successful proof. If you have produced an equation with a prime factor that appears once on one side and an even power on the other, the divisibility + symmetry argument will fire. That is why the template works for \sqrt{p} for any prime p: pick the prime on the right-hand side, apply Euclid's lemma, re-substitute, apply Euclid's lemma again.

Visual: the squeeze toward a shared factor

Both branches land on the same place: $2$ divides $p$, and $2$ divides $q$. That is the contradiction with $\gcd(p, q) = 1$. The proof's entire job is steering both arrows onto the same prime.

How to spot this finish coming

When you are partway through an irrationality proof and don't know how to end, the signal to look for is: is one side of my equation a multiple of a specific prime, and the other side a square (or cube)? If yes, Euclid's lemma fires. You get a divisibility statement on the base — p is divisible by the prime — and then substituting that back gives you the same statement on q.

If the proof stalls, the usual cause is that you haven't yet reached the "prime divides a square" configuration. The fix is almost always more algebra: square once more, cube once more, isolate a single surd on one side, square to remove it.

Adaptations for other bases

The finisher works for any prime, not just 2.

For \sqrt{3}: the finisher is "3 \mid p and 3 \mid q."
For \sqrt{5}: the finisher is "5 \mid p and 5 \mid q."
For \sqrt{p} where p is prime: same pattern, "p \mid both."

What about \sqrt{6}? 6 = 2 \cdot 3, not prime. But the same argument works: from p^2 = 6q^2, the right-hand side is divisible by 2, so p^2 is divisible by 2, so p is even. Write p = 2k: 4k^2 = 6q^2, so 2k^2 = 3q^2. The left is even, so the right is even — but 3 is odd, so q^2 must be even, so q is even. Both even. Contradiction with \gcd(p, q) = 1. You picked 2 as your contradiction prime; you could equally have picked 3 and done it differently.

Adaptations for higher roots

For cube roots, the finisher is the same shape but the intermediate algebra uses cubes. \sqrt[3]{2} = p/q gives p^3 = 2q^3, so 2 \mid p^3, so 2 \mid p (primes still have this property for cubes), so p = 2k, so 8k^3 = 2q^3, so q^3 = 4k^3, so 2 \mid q^3, so 2 \mid q. Contradiction.

The finisher — "both p and q divisible by the same prime" — doesn't change. The tool getting you there (Euclid's lemma for cubes) doesn't change either.

Why lowest terms, and not just any p/q?

One common student question: why not assume p/q in any form, not specifically lowest terms?

Because then the "both divisible by 2" wouldn't be a contradiction — it would just mean "oh, this fraction isn't in lowest terms, let me cancel." You would learn that p/q is reducible, but you wouldn't learn that it doesn't exist. The whole point of the lowest-terms clause is to preempt the cancellation. You say, up front: "I have cancelled everything I can. No more cancelling is possible." Then when you discover another cancellation is possible, you have a hard contradiction.

It is the mathematical equivalent of a witness saying under oath "I have told you everything I know" and then letting slip one more fact. The oath is what turns the slip into a problem.

The reflex, in one line

When you see "prove x is irrational," drive the algebra toward the shape "prime k divides both p and q". That is the only finish line. Everything else is how you get there.