The Preconditions of Infinity

A common topic being discussed on YouTube is how some infinities are bigger than others. We formulate that precisely using cardinality. We first make some preliminary notations.

Definition 1. Let K_1, K_2 be sets. Write K_1 \sim K_2 whenever there exists a bijection f : K_1 \to K_2.

Lemma 1. The relation \sim satisfy the following properties:

  • For any set K, K \sim K.
  • For sets K_1, K_2, K_1 \sim K_2 implies K_2 \sim K_2.
  • For sets K_1,K_2,K_3, K_1 \sim K_2 and K_2 \sim K_3 together imply that K_1 \sim K_3.

By Russell’s paradox, we cannot call \sim an equivalence relation on the set of sets.

Denote [0]:= \{\} \equiv \emptyset and [n]:= \{1,\dots,n\} for n \in \mathbb N^+. For the rest of this writeup, we will adopt the convention \mathbb N = \mathbb{N}_0.

Theorem 1. Let K \subseteq \mathbb N. Exactly one of the following is true:

  • K \sim [n] for some n \in \mathbb N.
  • K \sim \mathbb N.

In the first case, we say that K is finite. In the second case, we say that K is infinite.

Proof. Firstly, if K = \emptyset, then K = \emptyset \sim \emptyset = [0]. Next, suppose K \neq \emptyset. Define K_0 := K. By the well-ordering principle, for each i \in \mathbb N, whenever K_i \neq \emptyset, iteratively define

y_i := \min K_i,\quad K_{i+1} := K_i \backslash \{y_i\}.

Define L := \{y_i : i \in \mathbb N\} \subseteq K.

If the process ends, then there exists n \in \mathbb N such that K_{n+1} = \emptyset. In this case, L = \{y_1,\dots,y_n\} and L \sim [n] via the bijection k \mapsto y_k.

If the process never ends, then L \sim \mathbb N via the bijection k \mapsto y_k. We need to prove that K = L. It suffices to show that K \subseteq L.

Fix x \in K. Define M := \{y_i \in L : y_0 \leq y_i < x\}. We claim that M is finite. Suppose otherwise. Then for any n \in \mathbb{N}^+,

y_n \geq y_{n-1} + 1 \geq \cdots \geq \underbrace{ 1 + \cdots + 1 }_n = n.

We can make this calculation airtight using a quick exercise in mathematical induction. In particular, y_{x+1} \geq x.

On the other hand, y_{n+1} \in M, so that y_{x+1} < x \leq y_{x+1}, a contradiction. Thus, M is finite and we can write M = \{y_0,y_1,\dots,y_{n-1}\} for simplicity, so that y_n \geq x.

Clearly, x \in K \backslash M. On the other hand, by construction,

y_n = \min (K \backslash M) \leq x.

Therefore, x = y_n \in L, as required.

Corollary 1. A set K is countable precisely when K \sim L for some L \subseteq \mathbb N. For any countable set K, exactly one of the following is true:

  • K \sim [n] for some n \in \mathbb N.
  • K \sim \mathbb N.

In the first case, we say that K is finite. In the second case, we say that K is infinite.

Proof. The result follows by the transitivity of \sim.

Now, we have almost the necessary groundwork to discuss cardinality. We need to ensure that the notion of size as “number of items” is a consistent (i.e. well-defined) notion, at least for finite sets.

—Joel Kindiak, 4 Nov 24, 0951H

,

Published by


Response

  1. The Size of Finity – KindiakMath

    […] Previously, we showed that bijections preserve sets in a rather friendly manner. We almost have the necessary groundwork to discuss cardinality. […]

    Like

Leave a comment