You are partway through the \sqrt{2}-is-irrational proof. You have just shown p^2 = 2q^2, and you declare "so p^2 is even, which means p^2 = 2k for some integer k." A little voice at the back of your head says: wait — isn't every even square actually a multiple of four? Shouldn't p^2 = 4m, not 2k? And if p^2 = 2k is somehow wrong, isn't that already the contradiction? Can I stop the proof here?

No — you cannot stop, and the line "p^2 = 2k" is not a contradiction with anything. This misread is documented and surprisingly common, so let us unpick it carefully.

The sharp claim

p^2 = 2k is the definition of "p^2 is even." That is all it says. It does not claim p^2 is twice an odd number, or twice a prime, or anything else about k. The letter k is a placeholder for "some integer" — and "some integer" includes the case k = 2m, where the 4m structure hides.

Why this is the right framing: "even" means divisible by 2, full stop. If n is even, then n = 2k for some integer k. The k can itself be even, odd, zero, negative — anything. Writing p^2 = 2k commits you to nothing beyond "p^2 is divisible by two." The fact that p^2 also turns out to be divisible by four (because every even square is 4m) is a stronger statement that you simply have not asserted yet.

So the line "p^2 = 2k" is underpowered, not wrong. It is a correct, weak statement. You can always sharpen it later to "p^2 = 4m" once you know more about p, but the weak form is what the proof needs at this step.

Where the misconception comes from

Here is the thinking that usually generates the confusion:

  1. "I know every even square is a multiple of 4. So if p^2 is even, then p^2 = 4m."
  2. "But I just wrote p^2 = 2k. If the truth is p^2 = 4m and I wrote 2k, those disagree — so I have a contradiction!"

The flaw is in step 2. p^2 = 2k and p^2 = 4m do not disagree — they are compatible. p^2 = 4m is the special case k = 2m. Writing p^2 = 2k does not deny p^2 = 4m; it just does not commit to the stronger form. Both sentences are true simultaneously.

Analogy: writing "the sum of the angles of a triangle is more than 170°" is a true, weak statement about every triangle, even though the true value is exactly 180°. The weak statement is not a contradiction with the strong one — it is implied by it.

Where the contradiction actually lives

The contradiction in the \sqrt{2} proof does not appear at "p^2 = 2k" or even at "p is even." It appears two steps later, when the same move has been done twice:

Now you have "p is even and q is even" — they share a factor of 2. That is the line that clashes with the assumption \gcd(p, q) = 1. That pairing — "\gcd(p, q) = 1" (from the opening) and "\gcd(p, q) \ge 2" (from the two passes) — is the contradiction.

Before you reach both even, nothing has collided. Stopping early at "p^2 = 2k" or "p is even" would be claiming victory from a half-finished proof.

Where the real contradiction lives in the root-two proofFour numbered boxes arranged horizontally, connected by arrows. Box one labelled p squared equals two q squared is the starting point. Box two labelled p equals two k contains the line that students mistake for a contradiction. Box three labelled q equals two l follows after substitution. Box four, outlined in red, says gcd p q greater than or equal to two, contradicting the lowest terms assumption. The red box is the actual contradiction. p² = 2q² starting line p = 2k NOT a contradiction q = 2ℓ second pass gcd(p, q) ≥ 2 vs gcd = 1 ✗ contradiction students often stop here — don't "p = 2k" is just what "p² even" means it does not collide with any earlier line the real contradiction collides with the opening "p/q in lowest terms"
The chain of the $\sqrt{2}$ proof. The step $p = 2k$ is not a contradiction — it is the definition of $p^2$ being even unpacked. The real contradiction lives at the far right, where $\gcd(p, q) \ge 2$ (derived) clashes with $\gcd(p, q) = 1$ (assumed). That is the collision point. Stopping before it means stopping before the proof is done.

The two-line self-check

Whenever you suspect you have hit a contradiction mid-proof, run these two questions:

Question 1 — can you name the earlier line it collides with? Not a vague "something feels wrong." The specific sentence, quoted. For "p^2 = 2k," there is no earlier sentence it denies — the proof never asserted "p^2 \ne 2k."

Question 2 — is the collision across all integers, or only in the case at hand? "p^2 = 2k" is true for some integer k (namely k = q^2). That makes it a valid consequence, not an impossibility. A real contradiction is false for every possible witness.

If Question 1 has no answer, keep going — you are in the middle of a derivation, not at its end.

The correct stopping point

Claim. \sqrt{2} is irrational.

Proof sketch with the contradiction check inline.

  1. Assume \sqrt{2} = p/q, \gcd(p, q) = 1. (Opening; creates the constraint to be violated later.)
  2. p^2 = 2q^2. (Not a contradiction: just an equation about integers. Check 1 fails — nothing to collide with.)
  3. p^2 is even, so p = 2k. (Not a contradiction: p = 2k is compatible with everything assumed so far. Check 1 fails.)
  4. 4k^2 = 2q^2, so q^2 = 2k^2. (Not a contradiction: another equation. Check 1 fails.)
  5. q^2 is even, so q = 2\ell. (Not a contradiction yet: you have derived "q is even" but have not paired it with the opening.)
  6. p and q both even, so \gcd(p, q) \ge 2. Contradicts the opening \gcd(p, q) = 1. (Check 1 passes: the colliding sentence is step 1. Check 2 passes: \gcd cannot simultaneously be 1 and \ge 2.)

The stop-sign is at step 6, not step 3. Every line before step 6 is derivation, not collision. \square

The misconception is honest — you have heard "every even square is a multiple of 4," and you expected that fact to show up. It does, but it is not what creates the contradiction. What creates the contradiction is the second application of the even-implies-halves argument, which forces the shared factor of 2 into the denominator as well as the numerator. That is the move that kills the assumption.

Related: Proof by Contradiction · Lowest-Terms Contradiction: The Finisher on Every Irrationality Proof · When Is It Really a Contradiction (vs Just Something Weird)? · Why the √2 Proof Needs 'p and q Share No Common Factor' · Assume the Negation — But Still Check Your Moves Preserve Meaning