Indirect arguments are useful when a negated statement gives a concrete restriction that the original conclusion does not. Two common methods are proof by contrapositive and proof by contradiction. Their relationship was introduced in Converse, Inverse, and Contrapositive; here the focus is choosing assumptions carefully and finishing actual proofs.
We use classical logic. A contradiction must be a specific conflict with a hypothesis, definition, or established fact. Being unable to continue a calculation is not a contradiction. MIT’s introductory proof notes provide further examples of how assumptions support a proof [1][1] T. Leighton and R. Rubinfeld, “What Is a Proof?,” 2006. MIT 6.042/18.062J lecture notes, September 7, 2006. https://web.mit.edu/neboat/Public/6.042/proofs.pdf.
Choose the starting assumptions before calculating
| Method for proving | What to assume | What to establish |
|---|---|---|
| Direct | ||
| Contrapositive | ||
| Contradiction | and | A contradiction |
Domain restrictions remain active in every row. For a universal theorem, a contradiction proof assumes a counterexample exists and chooses one satisfying the hypothesis while violating the conclusion. For a nonexistence theorem, assume an object with the forbidden properties exists. Write the full negation, including its quantifiers, before manipulating symbols.
The two indirect methods can overlap: one may prove a contrapositive using a contradiction. Their names describe the organization of the argument, not mutually exclusive categories.
Contrapositive: use a workable representation
For every integer , if is even, then is odd.
Starting from the evenness of the polynomial gives little immediate information about . The negation of the conclusion is more useful: for integers, “not odd” means “even,” which supplies the representation .
We prove the contrapositive. Let with . Then
The bracketed quantity is an integer, so the polynomial is odd and therefore not even. This proves the contrapositive, hence the original statement.
This proof never assumes that the polynomial is even. A contradiction proof would assume both that it is even and that is not odd, then obtain a number that is simultaneously even and odd.
Contradiction: name the incompatible facts
The positive real number is not rational.
The assumption of rationality produces a fraction, while choosing lowest terms gives a property that divisibility can contradict. We use the already established parity fact that an integer with an even square is even.
Suppose , where are positive integers with no common divisor greater than one. Such a lowest-terms representation can be chosen for any positive rational number. Squaring gives
Thus is even, so is even. Write with . Substituting and simplifying yields
Consequently is even as well. Both and are divisible by , contradicting the lowest-terms condition. Therefore is irrational.
Merely finding a representation with an even numerator and denominator would not be a contradiction: is a perfectly valid fraction. It conflicts specifically with the choice of a coprime numerator and denominator. Naming that conflict is what completes the proof.
A finite-list contradiction
The prime numbers cannot be exhausted by a finite list.
We use the elementary fact that every integer has a prime divisor. One justification is to take its least divisor ; if were composite, a smaller divisor greater than one would divide , contradicting minimality.
Suppose all primes form the finite list . The list is nonempty because is prime. Set
Since , choose a prime divisor of . It must appear in the supposedly complete list, so it also divides the product . It would then divide the difference , which no prime does. This contradiction proves the claim.
The constructed need not be prime. The proof needs only a prime divisor outside the list. Replacing that statement by “the product plus one is always prime” would introduce a false claim.
Common logical slips
Negating “both and are odd” gives “at least one is not odd,” not “both are even.” Negating “ or ” gives both strict inequalities and . Integer assumptions may then sharpen each to a bound of ; the same sharpening is invalid for real numbers.
Also distinguish proving a claim from proving its converse. Showing an even integer has an even square does not establish that an even square has an even integer root. When unsure, write explicitly and compare the assumed statement with the target before continuing.
Exercises
Prove irrational. First explain why implies for integers .
Show solution
The possible remainders of modulo are , whose squares have remainders . Thus only remainder zero can give a square divisible by .
If in lowest positive terms, then , hence . Substitution gives , so is also divisible by . This contradicts coprimality. The residue argument supplies the divisibility fact instead of silently assuming it for every integer divisor.
Prove that there is no smallest positive rational number.
Show solution
Suppose is positive and no positive rational is smaller. Then is rational and satisfies , contradicting minimality. No lowest-terms representation is needed here. Equivalently, the construction proves directly that every positive rational has a smaller positive rational.
Prove by contradiction that if an integer has an even square, then is even. Identify the pair of incompatible facts.
Show solution
Assume is even and is not even. Integer parity makes odd, so write . Then is odd. The contradiction is that is both even and odd. If the proof instead starts only from odd and concludes odd , it is a direct proof of the contrapositive.
Prove that the sum of an irrational real number and a rational number is irrational.
Show solution
Let and . Suppose is rational. Then is rational, because rational numbers are closed under subtraction, contradicting the choice of . Explicitly, if and with nonzero integer denominators, then
This argument does not establish that the sum of two irrational numbers is irrational; refutes that different claim.
For integers , prove that odd implies both and are odd.
Show solution
The contrapositive assumes at least one factor is even. By symmetry, call it with . Then is even. The case where is even has the same argument with names exchanged, so the entire contrapositive is proved.
For integer , prove that if is even, then is even.
Show solution
Prove the contrapositive. If , then is odd, with . Therefore the original implication holds.
For integers , prove that implies or . Does the statement remain true for real ?
Show solution
Assume the negation of the conclusion: and . Because these are integers, each is at most , so . This proves the contrapositive. For real numbers the statement is false: gives sum but neither reaches .
From negation to induction
Use a contradiction when the negated target gives an object or extremal condition that cannot exist. Use a contrapositive when the negated conclusion has a convenient representation. For statements indexed by natural-number size, the more useful information may be the truth of smaller instances; that is the subject of Mathematical Induction.
References
- [1] T. Leighton and R. Rubinfeld, “What Is a Proof?,” 2006. MIT 6.042/18.062J lecture notes, September 7, 2006. https://web.mit.edu/neboat/Public/6.042/proofs.pdf ↩
Comments