Direct vs Contradiction — Two Paths to the Same Theorem, Side by Side

When a theorem can be proved in more than one way, comparing the proofs side by side is one of the best ways to understand what contradiction actually buys you. For some statements, the direct proof is shorter and cleaner. For others, the contradiction path is the only way in. And for a few, both work — and the one to pick is the one that makes the hidden structure most visible.

This page sets three theorems on split screens. For each, the left panel is a direct proof; the right panel is a contradiction proof. You watch both unfold; you compare lengths, clarity, and where the work is concentrated.

Theorem 1: "if n^2 is even, then n is even"

Direct path

This is actually quite awkward directly. You would want to assume n^2 is even (n^2 = 2k) and conclude n = 2m for some integer m. But unpacking n^2 = 2k does not factor cleanly into information about n itself. You end up saying "well, if n were odd, n^2 would be odd, contradiction" — which is a proof by contradiction hiding inside a direct-proof wrapper.

A genuinely direct proof would need a lemma like "every integer is either even or odd" (which is a fact) and then argue case-by-case. Doable, but it takes space.

Contradiction path

Assume n^2 is even and, for contradiction, n is odd. Then n = 2k+1, so n^2 = 4k^2 + 4k + 1 = 2(2k^2 + 2k) + 1, which is odd. But n^2 was assumed even. Contradiction. So n is even. \blacksquare

Comparison: Four lines. The contradiction path is clearly the natural one. The direct path is possible but tangled — it replicates the contradiction argument inside a case split.

Why contradiction is the right tool here: the statement "n^2 even ⇒ n even" is the contrapositive of "n odd ⇒ n^2 odd." The latter has a natural direct proof. Proof by contradiction essentially lets you prove the contrapositive without formally invoking it — and that is why it works so smoothly here.

Theorem 2: "there are infinitely many primes"

Direct path

You could try to directly exhibit an infinite sequence of distinct primes — e.g., the Euclid-Mullin construction: start with any prime, then each next term is the smallest prime factor of (product of all previous + 1). This generates 2, 3, 7, 43, 13, 53, 5, 6221671, \ldots. Every term is a prime, every term is new. Done.

This is a legitimate direct proof — no contradiction needed. It is constructive: you produce an infinite sequence, which requires infinitely many primes to exist.

Contradiction path

Assume there are only finitely many primes, list them, multiply them all, add 1. The result N has a prime factor not on the list — contradiction. See Euclid's Infinite-Primes Tree for the full argument.

Comparison: Both work. The contradiction form is a little crisper because it focuses on one clean algebraic move ("p_i \mid 1 is impossible") instead of the dynamics of generating a sequence. But the direct form has the advantage of being constructive — it actually builds primes, rather than just showing they exist. Mathematicians often prefer constructive proofs when available, and philosophical schools like constructivism accept only the direct form here.

Theorem 3: "\sqrt{2} is irrational"

Direct path

You cannot prove "\sqrt{2} cannot be written as p/q" directly in any natural way — the conclusion is "no such p, q exists," a negative existential, and negative existentials are the hardest thing to prove directly. You would need to argue over every possible pair (p, q) and rule each one out, which is infinitely many cases.

There is no reasonable direct proof of this theorem.

Contradiction path

Assume \sqrt{2} = p/q in lowest terms; square; derive p, q both even; contradict lowest terms. Four clean steps. See Proof by Contradiction.

Comparison: Direct is essentially impossible in any useful sense. Contradiction is the only approach. Negative existentials generally force you into contradiction or contrapositive.

The interactive: toggle between paths on three theorems

Drag the first slider to pick a theorem; drag the second slider to pick a path. Watch length and clarity change.

Theorem:

Three theorems with their two proofs laid out side by side. For "$n^2$ even ⇒ $n$ even," direct is awkward and contradiction is natural. For "infinitely many primes," both are about the same length; the direct version happens to be constructive. For "$\sqrt{2}$ is irrational," there is no direct proof — the statement is a negative existential, and contradiction is the only tool that works. The length column shows that contradiction is roughly four lines in every case — contradiction tends to be short when it applies.

When contradiction is clearly better

Four situations force your hand toward contradiction:

The conclusion is a negative existential — "there is no x such that \ldots", "\alpha cannot be written as p/q." Direct proofs have nothing to start with because the conclusion is about the non-existence of an object. Contradiction supplies an object to manipulate.
The conclusion is a universal negative — "no integer of this form is prime." Same issue: direct reasoning has nothing positive to build, while contradiction hypothesises the forbidden object and finds the wall.
The conclusion asserts uniqueness — "there is exactly one x with property P." Assume two distinct candidates both have P, derive a contradiction with their distinctness.
The hypothesis gives you no handle — if you stare at the forward chain and see no algebraic move, reversing to "assume the conclusion fails" often hands you new equations that direct reasoning was withholding.

When direct is clearly better

Three situations favour the direct path:

The conclusion is constructive — "there exists x such that \ldots" When you need to produce an object, produce it directly. A contradiction proof of existence ("if no x existed, …") is philosophically contested (constructivists reject it) and rarely as informative.
The hypothesis is rich — algebraic identities, factoring, substitution theorems. When you can manipulate the hypothesis into the conclusion in a couple of steps, contradiction is overkill.
Pedagogical clarity — when teaching or communicating, a direct chain is usually easier to follow than a contradiction, because the reader knows at every step why they are doing this step. In contradiction, the reader is walking around a doomed assumption, which is cognitively more expensive.

The rule of thumb

Try direct first. Almost always, a claim of the form "if P then Q" should be attacked directly. If the attempt stalls after unpacking the definitions, then ask whether the conclusion is negative, non-existence, or uniqueness — and switch to contradiction if so.

Never reach for contradiction just because it is the "clever" move. If direct works, use direct. The direct version is almost always more informative — it tells you not just that something is true, but why the algebra marches in the direction it does.

The meta-comparison

	Direct	Contradiction
Works when	Conclusion is positive, algebraic chain is visible	Conclusion is negative, no direct chain exists
Style	Forward	Hypothesise negation, find wall
Constructive?	Yes	Sometimes not
Informativeness	Shows why	Shows that
Typical length	Varies	Usually short once the wall is found
Best signal	Clean hypothesis and forward chain	Negative existential or uniqueness

Both paths are legitimate mathematics. The art is in picking the one that makes the argument shortest and most transparent — and for that, seeing side-by-side comparisons like the one above is the single best way to build intuition.