Proof by Contradiction
Detectives do this every day: assume the suspect is innocent, follow that assumption through, and if
it leads somewhere that plainly cannot be true — the suspect would have had to be in two cities at
once — then innocence is ruled out. Mathematicians turned that everyday trick into one of their most
powerful tools.
Here is one of the sneakiest, most satisfying strategies in all of mathematics. To prove that
something is true, you begin by boldly assuming it is false. You
take that false assumption completely seriously and follow it, step by careful step, wherever the
logic leads — until it marches you straight into something impossible: a
square that is round, a number that is both even and odd, a fraction that is somehow smaller than
itself.
When correct reasoning delivers an absurdity, only one thing can be to blame: the assumption you
started with. So the assumption must be false — which means the original statement
was true all along. You have proved it, not by building it up, but by watching its
opposite self-destruct. It feels like a magic trick, and it powers some of the most beautiful
results ever discovered.
Worked example 1 — the classic: \sqrt{2} is irrational
A number is irrational if it cannot be written as a fraction of whole numbers.
We'll prove \sqrt{2} is irrational. Trying to build that directly is
awkward — so instead, assume the opposite: suppose
\sqrt{2} can be written as a fraction, and moreover in
lowest terms (all common factors already cancelled):
\sqrt{2} = \frac{a}{b}, \qquad \text{with } \frac{a}{b} \text{ in lowest terms.}
Square both sides and rearrange:
2 = \frac{a^2}{b^2} \quad\Longrightarrow\quad 2b^2 = a^2.
So a^2 is even, which forces a to be even
(an odd number squared stays odd), say a = 2c. Substitute back:
2b^2 = (2c)^2 = 4c^2 \quad\Longrightarrow\quad b^2 = 2c^2.
Now b^2 is even too, so b is even. But if
a and b are both even, the fraction
\tfrac{a}{b} was not in lowest terms — flatly
contradicting how we chose it. The assumption has destroyed itself, so
\sqrt{2} is irrational. \blacksquare
Worked example 2 — there is no largest even number
Claim: there is no biggest even number. Directly, that's hard to picture — how do you argue about
"all even numbers at once"? Contradiction makes it easy.
Assume the opposite: suppose there is a largest even number; call it
N. Now look at
N + 2.
Since N is even, N+2 is also even — and it is
bigger than N. That contradicts
N being the largest even number. The assumption fails, so no
largest even number exists. \blacksquare
Notice the flavour: assuming a "biggest" thing exists, then constructing something bigger, is a
pattern you'll see again and again — including in Euclid's proof about primes below.
Worked example 3 — a warm-up in one line
Claim: if n^2 is even, then n is even.
Assume the opposite: suppose n is odd, so
n = 2k+1. Then
n^2 = (2k+1)^2 = 4k^2 + 4k + 1 = 2(2k^2 + 2k) + 1,
which is odd. But we were told n^2 is even — a
contradiction. So n cannot be odd; it must be even.
\blacksquare (This is exactly the little fact the
\sqrt{2} proof leaned on.)
The pattern
Every proof by contradiction has the same shape: flip the claim to its exact opposite, follow the
logic honestly, and wait for the impossible to appear.
To prove a statement P by contradiction:
-
assume the opposite — suppose P is
false (the exact logical negation of P);
-
reason with correct, valid steps until you reach a
contradiction — something genuinely impossible;
-
conclude that the assumption must have been wrong, so
P is true;
-
classic results proved this way: \sqrt{2} is
irrational, and there are infinitely many primes.
Two things sink more contradiction proofs than anything else — get them right and the method is
bulletproof:
-
Assume the EXACT negation. The opposite of "x > 5" is
"x \le 5", not "x < 5". The
opposite of "all swans are white" is "at least one swan is not white", not "no swans are
white". If you negate the statement carelessly, you have set off to disprove the wrong thing, and
the whole proof is worthless — even if every step after it is flawless.
-
Reach a REAL contradiction, not a surprise. You must arrive at something
impossible — 1 = 0, a number both even and odd, a fraction
not in lowest terms after you insisted it was. A result that is merely unexpected, ugly, or
"surely can't be right" is not a contradiction. If the absurdity isn't
undeniable — something no valid reasoning could ever produce — you haven't finished the proof.
Proof by contradiction has produced some of the most beautiful results in mathematics — and one of
its most dramatic legends. Around 300 BC, Euclid proved there are
infinitely many primes in a few devastating lines. Assume the opposite: that there
are only finitely many, say p_1, p_2, \dots, p_n. Multiply them all
together and add one:
N = p_1 p_2 \cdots p_n + 1.
This N leaves remainder 1 when divided by
every prime on the list, so none of them divides it — yet every number has a prime factor.
Contradiction! So the list can never be complete: the primes never run out.
Even older is the \sqrt{2} result. The Pythagoreans held a near-religious
belief that every number was a ratio of whole numbers — that the universe was built from tidy
fractions. The discovery that \sqrt{2} is irrational shattered that
creed, and legend has it the cult was so scandalised that they drowned the man,
Hippasus, who let the secret out. Small wonder the mathematician G. H. Hardy called proof by
contradiction "a mathematician's finest weapon" — sharper, he said, than any gambit in chess. You
can watch its opposite, the demolishing power of a single
counterexample,
elsewhere.
See it explained