You are on page 1of 1

Inferential Equivalence

Definition 1. Two statements are said to be inferentially equivalent if, and only if, each implies
the other.

Many theorems are of the form of an inferential equivalence. Let us consider Wilson’s Theorem
as an example.

Wilson’s Theorem. n is a prime if, and only if, n is an integer > 1 such that n | (n – 1)! + 1.

Let A be the statement ‘n is a prime’, and let B be the statement ‘n is an integer > 1 such
that n | (n – 1)! + 1’.

Then Wilson’s Theorem can be stated as ‘A if, and only if, B’.

So, an inferential equivalence is a theorem of the form ‘A if, and only if, B’.

Definition 2. The right-pointing arrow → means ‘only if’, and the left-pointing arrow  means
‘if’, and the double-pointing arrow  means their logical conjunction.

So, an inferential equivalence is a theorem of the form ‘A  B’. In other words, ‘A  B’ is an


abbreviation for ‘(A → B) and (A  B)’, which can be also written as ‘(A → B) and (B → A)’.

It sometimes happens that more than two statements of interest are all inferentially equivalent to
each other. For example, suppose that three statements A, B, and C are all inferentially
equivalent to each other. In such a case, a common proof technique is to prove them in a circle.
For example, we prove A → C and then we prove C → B and then we prove B → A. This
establishes that they are all inferentially equivalent to each other. In making this circle, we can
choose any order we wish. Going from one inferentially equivalent statement to another can vary
greatly in difficulty. That is why we wrote C after A in this example – to emphasize our option to
choose the order.

To appreciate the difference in difficulty, let us consider the ‘A’ and ‘B’ components of Wilson’s
Theorem again. Going from A to be is very difficult, but going from B to A is very easy. Another
such example is the characterization of even perfect numbers, known as the Euclid-Euler
theorem, with part of the characterization being given by Euclid, and the rest of the
characterization being given two millennia later by Euler.

The following statements are all inferentially equivalent to each other:

The Axiom of Choice


Zorn’s Lemma
The Well-Ordering Theorem
Tukey’s Lemma

Going from the Axiom of Choice to the Well-Ordering Theorem is difficult, but going from
Zorn’s Lemma to the Well-Ordering Theorem is relatively easy, and going from the
Well-Ordering Theorem to the Axiom of Choice is extremely easy.

(end of document)

You might also like