Why Does Proving the Negation Is False Prove the Original Is True?

You have just finished your first proof by contradiction, and a doubt creeps in. "I showed the negation leads to a contradiction, so the negation is false. But wait — what if the original is also somehow false? Could they both be false? How do I know ruling out one automatically confirms the other?"

This is a completely natural worry, and resolving it takes you straight to the logical foundation that proof by contradiction is built on. The short answer: in standard mathematics, a proposition and its negation cannot both be false — exactly one of the pair is true. Ruling out one therefore forces the other. The slightly longer answer is worth understanding, because it sharpens what a "proposition" really is.

The law of the excluded middle

The principle at work is called the law of the excluded middle (LEM). In plain language:

For every proposition P, either P is true or \lnot P is true. No third option.

Equivalently: P \lor \lnot P is always true, whatever P is. This is not something you prove from other axioms — it is a foundational rule of classical logic, built into the definition of what "true" and "false" mean.

Why LEM makes the "both could be false" worry dissolve: by definition, a proposition is a sentence that is either true or false — there is no middle ground, no "somewhat true," no "undefined." If you rule out the negation by showing it leads to absurdity, the only remaining option in the pair is that the original is true. The worry "what if both are false" describes a third status — "both assertive but both not true" — which LEM explicitly forbids.

You can think of it like a light switch. If the switch is not off, it is on — there is no third setting. You do not need to separately verify "on-ness" after ruling out "off-ness"; the two-valued structure of the switch guarantees one from the other.

What "proposition" actually means

The subtle part is that LEM is a statement about propositions, and propositions are a specific kind of sentence — not every English sentence qualifies. A proposition must be definite: it makes a clear claim that is either true or false, with no vagueness and no self-reference.

"\sqrt{2} is irrational" — a proposition. Either it is true (as the proof shows) or it is false. There is no third status.
"This sentence is false" — not a proposition. It attempts self-reference and fails to have a stable truth value. LEM does not apply because the object does not qualify.
"The current king of France is bald" — not a straightforward proposition either (the referent does not exist). Philosophers argue about this one.

In everyday mathematical practice, every claim you care about — \sqrt{2} is irrational, there are infinitely many primes, every even integer greater than 2 is the sum of two primes, and so on — is a proposition. LEM applies, and the "both could be false" scenario is impossible.

The law of the excluded middle allows two configurations: $P$ true or $\lnot P$ true. It forbids the configuration where both are false (the third row). This is why ruling out $\lnot P$ via contradiction directly confirms $P$ — there is no "both false" escape hatch.

The formal argument

Let us spell out the reasoning step by step, to see why "both could be false" is not a loophole.

LEM says: P is true, or \lnot P is true.
You do a proof by contradiction: assume \lnot P, derive an absurdity. This shows \lnot P is not true (because true statements do not produce absurdities — they produce consistent consequences).
So \lnot P is false. By step 1, the only remaining option is that P is true.
Therefore P is true. \square

The crucial move is step 3 to step 4. It uses the two-valuedness of propositions: "\lnot P is false" and "P is true" are the same statement by definition. A proposition and its negation have opposite truth values — if one is false, the other is true. No gap, no third option.

This is not a sleight of hand; it is the definition of negation. If P and \lnot P are both false, then \lnot P is a statement whose falseness means P is not false, which means P is true — contradicting the assumption that P was false. The only consistent assignment of truth values makes exactly one of them true.

What constructivists say (and why it does not change your practice)

There is a school of mathematics — constructive mathematics — that refuses to assume LEM as a universal axiom. Constructivists argue that proving "P is true" should require constructing a witness for P, not merely ruling out the alternative. For them, ruling out \lnot P gives a weaker result: you have shown \lnot \lnot P, which is not automatically P in their framework.

In practice, this matters only for certain advanced settings (computer-verified proofs, some foundations research, certain branches of topology). In Indian school mathematics, JEE-level mathematics, and nearly all of research mathematics, LEM is accepted and proof by contradiction is fully valid. Your intuition "ruling out the negation proves the original" is correct under the standard rules.

Why the mainstream accepts LEM: mathematics is enormously more powerful with LEM than without it. Many classical theorems — including the irrationality of \sqrt{2}, the infinitude of primes (one proof), and the fundamental theorem of algebra — have their cleanest proofs via contradiction. Constructive mathematicians develop alternative proofs when they can, but the classical framework with LEM is the default for good reason: it is expressive, internally consistent, and matches the way mathematicians have reasoned for two thousand years.

A sanity check: what if the proof is flawed?

The worry "what if both could be false" sometimes masks a different concern: "what if my proof has a bug?" That is a real and healthy worry — but it is a worry about your proof, not about the logic.

If your derivation of the contradiction has an algebra error, you have not actually shown \lnot P is impossible. You just think you have. The remedy is to re-check every step.
If your negation of P is wrong (you assumed the opposite of the opposite, or you negated an "all" as "none" instead of "some not"), you have proved something about a different statement. The remedy is to redo the negation carefully.
If the derivation is clean and the negation is correct, the conclusion follows. LEM is not a point of weakness in your proof; it is part of the bedrock.

The doubt resolved on a concrete claim

Claim. There is no smallest positive real number.

Proof by contradiction. Assume there is a smallest positive real, call it \varepsilon > 0. Then \varepsilon / 2 is also a positive real (half a positive number is positive), and \varepsilon / 2 < \varepsilon. But \varepsilon was assumed to be smallest — nothing positive is strictly less than \varepsilon. Contradiction. Therefore no smallest positive real exists. \square

Where does the doubt live here? "Could both 'there is a smallest' and 'there is no smallest' be false?" No — these are exact opposites (negation of an existence claim is a non-existence claim). LEM says exactly one is true. The proof ruled out the first; the second follows automatically.

If you still feel uneasy, notice that the conclusion matches your intuition: the positive reals really do not have a smallest element (you can always halve). The proof is not producing a counterintuitive result; it is just formalising what halving shows. The logic checks out, the conclusion matches reality, and the "both false" worry does not apply.

The short story: proof by contradiction works because propositions are two-valued by definition. Ruling out one value forces the other. The "what if both are somehow false" possibility is not a gap in the method — it is a scenario that the very rules of logic exclude. Trust the structure, verify the algebra, and write the conclusion.