Zorn
Zorn
Zorn
DANIEL R. GRAYSON
1. Introduction
Zermelo gave a beautiful proof in [6] that every set can be well ordered, and
Kneser adapted it to give a direct proof of Zorn’s lemma in [3]. Sources such as [4],
[5], [2, p. 63], and most recently, [1], describe this proof, but it still doesn’t seem to
be generally known by mathematicians.
2. The proof
A partially ordered set is a set X equipped with a relation x ≤ y satisfying x ≤ x
and x ≤ y ≤ z ⇒ x ≤ z and x ≤ y ≤ x ⇔ x = y. (The last property is easily
obtained by considering the quotient set for the equivalence relation x ∼ y ⇔ x ≤
y ≤ x.) A totally ordered set is a partially ordered set where x ≤ y ∨ y ≤ x. A well
ordered set is a totally ordered set where every nonempty subset has a minimal
element. A closed subset Y of a partially ordered set X is a subset satisfying
x ≤ y ∈ Y ⇒ x ∈ Y ; we write Y ≤ X, and if Y 6= X, too, then we write Y < X. If
X is well ordered and Y < X, and we take x to be the smallest element of X − Y ,
then Y = {y ∈ X | y < x}.
Lemma 2.1. Suppose X is a partially ordered set, and F is a collection of subsets
S of X. Suppose also that for any C, D ∈ F ,
which are well ordered by the ordering
either C ≤ D or D ≤ C. Let E = C∈F C. Then E is well ordered, and for each
C ∈ F we have C ≤ E.
Theorem 2.2 (Zorn’s lemma). A partially ordered set X with upper bounds for its
well ordered subsets has a maximal element.
Proof. Suppose not. For each well ordered subset C ⊆ X pick an upper bound
g(C) ∈ / C. A well ordered subset C ⊆ X such that c = g({c0 ∈ C | c0 < c}) for
every c ∈ C will be called a g-set.
Intuitively, a g-set C, as far as it goes, is determined by g. For example, if C
starts out with {c0 < c1 < c2 < . . . }, then necessarily c0 = g({}), c1 = g({c0 }),
c2 = g({c0 , c1 }), and so on. A pseudoproof of the theorem might go like this. We
start with an empty collection of g-sets and add larger and larger g-sets to it. At
each stage let W be the union of the g-sets encountered previously. We see that
W 0 = W ∪ {g(W )} is a larger g-set, and we add it to our collection. Continue this
procedure forever and let W be the union of the g-sets encountered along the way;
it’s again a g-set, and we can enlarge it once again, thereby encountering a g-set
that isn’t in our final collection and providing a contradiction. The problem with
this pseudoproof is in interpreting the meaning of “forever”, so now we turn to the
real proof.
Date: January 22, 2007.
1
2 GRAYSON
References
[1] Akihiro Kanamori. The mathematical import of Zermelo’s well-ordering theorem. Bull. Sym-
bolic Logic, 3(3):281–311, 1997.
[2] Irving Kaplansky. Set theory and metric spaces. Chelsea Publishing Co., New York, second
edition, 1977.
[3] Hellmuth Kneser. Eine direkte Ableitung des Zornschen Lemmas aus dem Auswahlaxiom.
Math. Z., 53:110–113, 1950.
[4] T. Szele. On Zorn’s lemma. Publ. Math. Debrecen, 1:254–256, erratum 257, 1950.
[5] J. D. Weston. A short proof of Zorn’s lemma. Arch. Math., 8:279, 1957.
[6] Ernst Zermelo. Beweis, daß jede Menge wohlgeordnet werden kann. Math. Ann., 59:514–516,
1904.