by Carl Pierer
Among the many tools available to mathematicians attempting to prove a statement is something called “proof by contradiction” or reduction ad absurdum. The general method of the proof is a very smug one: Let the statement to be proved be Φ. The strategy, then, is to suppose that Φ is false and to consequently derive a contradiction. Now there are quite a few, infinitely many one might say, ways of going about this. This shan't be our concern. What is of interest here is the question as to what kind of contradiction forms the end of such a proof. Let us distinguish 2 cases:
- Internal contradiction. The proof takes the form of:
Suppose ¬Φ. Then γ.
Δ.
∴, ¬γ.
Contradiction! ∴ Φ.
Here, we deduce an immediate consequence (γ) from the assumption that ¬Φ and then proceed by a sequence of logical steps (Δ) to show that this leads to ¬γ – a blatant contradiction. For example, let us say our statement to prove is:
Φ: is not a rational number
Then, ¬Φ is ” is a rational number” and an immediate consequence of this would be:
γ: Then we can write √2 = m/n , where m,n are natural numbers, such that m,n are coprime. That is, m/n is an irreducible fraction.
One standard proof then goes as follows:
1. Then, 2 = m2/n2
2. So, 2n2 = m2
3. This means, m is even.
4. So m can be written as m = 2k for some natural number k.
5. Thus: 2n2 = (2k)2
6. Ergo, 2n2 = 4k2
7. And hence: n2 = 2k2
8. So n is even.
9. Then 2 divides both n and m.
10. Therefore, ¬γ.
Now, we have the desired contradiction, and therefore we conclude Φ.
- External contradiction. Here, the proofs take the following form:
Suppose ¬Φ. Then:
Δ.
∴ μ.
But ¬μ! Contradiction. Therefore, Φ.
Here, ¬μ is some generally accepted mathematical truth, such as 1 ≠ 2. In contrast to the internal contradiction, it need not be a statement that is a consequence of ¬Φ. ¬μ is some statement in the general corpus of mathematical truths (i.e. proven statements) and is not necessarily linked in its content to Φ.
Now, the astute reader might question the meaningfulness of the distinction between internal and external contradiction. In particular, since μ appeared in the proof of Φ, there must be a sequence of logical derivations that relate the two. So, one might ask, how much sense does it make to distinguish between a “direct” consequence of ¬Φ and one that is related by a longer sequence of logical steps? Certainly, formally speaking, such an objection is valid. But there seems to be an intuitive sense in which some consequences are more immediate than others, and this is all that is needed at this stage.

