This document presents a proof that every preorder can be embedded into a total preorder. It defines preorders and total preorders, and states the theorem to be proved. It then proves the theorem in two lemmas: Lemma 1 shows that the set of total preorders containing a given preorder has maximal elements, and Lemma 2 shows that the maximal elements must be the total preorder on the entire original set. Together these imply the theorem that every preorder has a totalization.
This document presents a proof that every preorder can be embedded into a total preorder. It defines preorders and total preorders, and states the theorem to be proved. It then proves the theorem in two lemmas: Lemma 1 shows that the set of total preorders containing a given preorder has maximal elements, and Lemma 2 shows that the maximal elements must be the total preorder on the entire original set. Together these imply the theorem that every preorder has a totalization.
This document presents a proof that every preorder can be embedded into a total preorder. It defines preorders and total preorders, and states the theorem to be proved. It then proves the theorem in two lemmas: Lemma 1 shows that the set of total preorders containing a given preorder has maximal elements, and Lemma 2 shows that the maximal elements must be the total preorder on the entire original set. Together these imply the theorem that every preorder has a totalization.
Abstract. In some contexts, like decision theory, it is useful to know that a
preorder can be embedded into a total preorder. I show that (subject to the axiom of choice) this can always be done.
A preorder is a pair (S, ≤S ) where S is a set and ≤S is a reflexive, transitive,
binary relation on S. The preorder is total if x ≤S y or y ≤S x, for all x and y in S. A totalization of (S, ≤S ) is a total preorder (S, ≤0S ) such that x ≤0S y whenever x ≤S y. Theorem. Every preorder has a totalization. Suppose (S, ≤S ) is a preorder. For any A ⊂ S, we obtain a preorder (A, ≤A S) defined so that, for any x, y ∈ A, x ≤A S y ⇐⇒ x ≤S y. Consider the set X of all preorders (A, ≤A ) such that A ⊂ S and (A, ≤A ) is a totalization of (A, ≤AS ). Then (X, ≤X ) is a preorder, where
[(A, ≤A ) ≤X (B, ≤B )] ⇐⇒ A ⊂ B and (A, ≤A ) = (A, ≤A
B) . In fact, (X, ≤X ) is a partial order : x ≤X y and y ≤X x together imply x = y. The following two lemmas imply the theorem. Lemma 1. X contains maximal elements. In other words, there exists x ∈ X such that x ≤X y ⇐⇒ x = y. Proof. Zorn’s Lemma yields the existence of maximal elements, subject to two hypotheses. First, X must not be empty. Well, suppose A is a one-element subset of S; then (A, ≤A S ) lies in X. Second, consider any increasing sequence (A1 , ≤1 ) ≤X (A2 , ≤2 ) ≤X · · · in X. Zorn’s Lemma requires that there exist (A, ≤A ) ∈ X with (An , ≤n ) ≤X (A, ≤A ) for all n. Set A := ∪∞ n=1 An and define ≤A in the following way: for any x, y ∈ A there exists some n with x, y ∈ An ; put x ≤A y ⇐⇒ x ≤n y. Lemma 2. If (A, ≤A ) ∈ X is maximal, then A = S. Proof. Suppose there exists s ∈ S − A. Set B := A ∪ {s}. Define ≤B by the rules x ≤B y ⇐⇒ x ≤A y s ≤B s x ≤B s ⇐⇒ x ≤S s s ≤B x ⇐⇒ [s ≤S x or not x ≤S s] for all x, y ∈ A. Then (A, ≤A ) ≤X (B, ≤B ). This contradicts the maximality of (A, ≤A ).