P4 1.1 - Proof by contradiction

P4 1.1 - Proof by contradiction

Prove a statement by showing that its negation leads to an impossibility. Build precise assumptions, prove that 2\sqrt2 is irrational and that there are infinitely many primes, then adapt the same reasoning to unfamiliar statements.

Assuming the negation

Could any finite list contain every prime number? Checking larger and larger primes will never settle that question. A proof can instead assume that a complete finite list exists, then show why that assumption defeats itself. First learn how to make such an assumption precisely.

To prove “if PP, then QQ”, a contradiction argument keeps PP and assumes that QQ is false. The negation of “every allowed value works” is “at least one allowed value fails”, not “every value fails”. The original domain still applies.

Proof by contradiction

To prove a statement by contradiction, assume its negation, derive a consequence that conflicts with a known fact or an assumption, and conclude that the negation is impossible. Therefore the original statement is true.

An integer is either even, 2k2k, or odd, 2k+12k+1, for an integer kk. It cannot be both. These forms turn a parity claim into algebra.

Worked example 1: an even square

Prove by contradiction that if n2n^2 is even, then the integer nn is even.

Plan. Keep the even-square assumption and suppose that nn is odd.

Assume n2n^2 is even but n=2k+1n=2k+1, where kk is an integer. Then

n2=(2k+1)2=4k2+4k+1=2(2k2+2k)+1.n^2=(2k+1)^2=4k^2+4k+1=2(2k^2+2k)+1.

The bracket is an integer, so n2n^2 is odd, contradicting its assumed evenness.

Answer. The assumption that nn is odd is impossible; therefore nn is even.

Check. The contradiction concerns the same quantity, n2n^2: it was assumed even and proved odd. No special value of nn was used, so the argument covers every integer satisfying the premise.

Change only the target parity from even to odd. The negated conclusion must change too.

Worked example 2: an odd square

Prove by contradiction that if n2n^2 is odd, then the integer nn is odd.

Plan. Assume the square is odd but the integer is even, then calculate the square's parity.

Assume n2n^2 is odd and n=2kn=2k for an integer kk. Then

n2=(2k)2=4k2=2(2k2),n^2=(2k)^2=4k^2=2(2k^2),

which is even. This contradicts the assumption that n2n^2 is odd.

Answer. nn cannot be even; hence it is odd.

Check. Every integer that is not odd is even, so the assumed case was the complete negation of the conclusion. The algebra establishes a conflict with the retained premise.

The assumption should contain both the given information and the opposite of what you must prove.

Watch out

To prove “if PP, then QQ”, do not assume “not PP and QQ”. That describes a different situation. In Example 1 the contradictory assumption is “n2n^2 even and nn odd”.

Apply the same approach after a small change to the expression.

Irrationality of square root of two

No terminating decimal calculation can prove that 2\sqrt2 is irrational. Instead, assuming a rational representation forces both its numerator and denominator to be even, even when the fraction was chosen in lowest terms.

A rational number is expressible as a/ba/b, where a,ba,b are integers and b≠0b\ne0. A real number with no such representation is irrational. Any rational number can be written in lowest terms by cancelling common factors.

Coprime integers

Integers a,ba,b are coprime if their only common positive factor is 11. A fraction in lowest terms has coprime numerator and denominator. If both were even, 22 would be a common factor, contradicting that choice.

From the previous part, an even integer square must have an even base. That result is essential; writing “a2a^2 even, so aa even” now has an established reason.

Worked example 3: irrationality of square root of two

Prove by contradiction that 2\sqrt2 is irrational.

Plan. Assume a lowest-terms rational representation, square it, and use parity to force a common factor.

Assume 2=a/b\sqrt2=a/b, with positive coprime integers a,ba,b. Squaring gives

2=a2b2⟹a2=2b2.2=\frac{a^2}{b^2} \quad\Longrightarrow\quad a^2=2b^2.

Thus a2a^2 is even, so aa is even. Write a=2ka=2k, for an integer kk. Substituting gives

4k2=2b2⟹b2=2k2.4k^2=2b^2 \quad\Longrightarrow\quad b^2=2k^2.

So b2b^2, and hence bb, is even too. Then a,ba,b have the common factor 22, contradicting their being coprime.

Answer. A rational representation is impossible, so 2\sqrt2 is irrational.

Check. The contradiction requires both numbers to be even. One even numerator would not conflict with a fraction being in lowest terms. Every parity step follows from an exact integer equation, not a decimal approximation.

Change only the number inside the root from 22 to 2/32/3. The odd denominator changes the algebra but preserves the parity argument.

Worked example 4: adapting the parity argument

Prove by contradiction that 2/3\sqrt{2/3} is irrational.

Plan. Assume a coprime fraction, clear the denominator, and again force both integers to be even.

Assume 2/3=a/b\sqrt{2/3}=a/b, with positive coprime integers a,ba,b. Then

3a2=2b2.3a^2=2b^2.

The right side is even. Since 33 is odd, a2a^2 must be even: an odd a2a^2 would make 3a23a^2 odd. Hence a=2ka=2k, for an integer kk, and

3(2k)2=2b2⟹b2=6k2=2(3k2).3(2k)^2=2b^2 \quad\Longrightarrow\quad b^2=6k^2=2(3k^2).

Thus bb is even as well, contradicting coprimality.

Answer. 2/3\sqrt{2/3} is irrational.

Check. The critical odd-factor step can be checked algebraically: if a=2j+1a=2j+1, then 3a2=12j2+12j+3=2(6j2+6j+1)+13a^2=12j^2+12j+3=2(6j^2+6j+1)+1, which is odd. It therefore cannot equal the even integer 2b22b^2. The proof does not assume irrationality of another root.

The lowest-terms choice turns a common-factor conclusion into an actual contradiction.

Watch out

“The numerator and denominator are even” is possible for an unreduced fraction, such as 2/42/4. State coprimality at the start; otherwise you have not explained why evenness is impossible.

Change the odd denominator from 33 to 55.

Infinitely many primes

A finite list can always be challenged by a number built from its own entries. Multiplying all the listed primes and adding one produces a number that none of those primes can divide.

Continue this lesson.

You’re previewing NovaLearn without an account. Create a free account to read the rest.

Create free account