1

This is an equivalence proof by myself. I need someone to tell me if the proof is alright or to point out its flaws.

Definition 1: A set M is countable iff $|M| = |\Bbb N|$ or |n $\in \Bbb N $| = |m $\in M$| (here I mean that if M is finite then you can just pair the element m $\in$ M to a natural number n $\in \Bbb N$ and it will become a finite bijection between all the elements of M and the same amount of numbers of $\Bbb N$).

Definition 2: A set M is countable iff $|M| \le |\Bbb N|$, i.e. there is some injective $g: M \to \Bbb N$.

So now I want to prove equivalence of the two definitions:

1 $\to$ 2: This is easy because if $|M| = |\Bbb N|$ or |n $\in \Bbb N $| = |m $\in M$|then also always $|M| \le |\Bbb N|$.

2 $\to$ 1: I split in two cases.

Case 1: $|M| \le |\Bbb N|$, i.e. there is some injective $g: M \to \Bbb N$, and rng(g) is finite. Then this leads naturally to the second variant of definition 1 because we can just take the n elements of rng(g) and map it into M which must be bijective.

Case 2: $|M| \le |\Bbb N|$, i.e. there is some injective $g: M \to \Bbb N$, and rng(g) is infinite. So it must be |rng(g)| = |$\Bbb N$| = |M| which lets us arrive at variant 1 of definition 1.

But why is |rng(g)| = |$\Bbb N$|? Because rng(g) $\subseteq \Bbb N$ which means that rng(g) is well-ordered as well and so every non-empty subset of rng(g) has a minimum element. So you can construct the following chain:

a_minimum $\in$ rng(g)

b_minimum $ \in$ rng(g)\a_minimum

c_minimum $ \in$ rng(g)\a_minimum, b_minimum

...

Since rng(g) is infinite this goes on endless and you can cleary see the structure of the natural numbers rising, so that there's a bijection possible between rng(g) $\to \Bbb N$, so that |rng(g)|=|$\Bbb N$|.

But why |rng(g)| = |M|? Because of the definition of rng(g) we get that the inverse function rng(g) $\to$ M must be surjective but the function rng(g) $\to$ M is injective by assumption of definition 2: $|M| \le |\Bbb N|$. That's why |rng(g)| = |M|.

  • 1
    You need to add $M$ not being finite to your hypothesis. Otherwise you can take $g: \emptyset \to \mathbb N$ or $g:{ 15 } \to \mathbb N$, both are injective, yet the domains are finite. – Steve Jun 21 '21 at 21:27
  • I fixed it by including finite-countability. –  Jun 21 '21 at 21:44
  • 1
    Now your first definition is wrong. ${ 1 }$ is countable, but there is no bijection ${ 1 } \to \mathbb N$. – Steve Jun 21 '21 at 21:47
  • No, because look at the last part where countable-finite is defined. I think its ok now, don‘t u think? –  Jun 21 '21 at 21:49
  • Now it's all good! :) – Steve Jun 21 '21 at 21:55

1 Answers1

2

Let $g(M)=H_0$ be an infinite subset of $\mathbb N$. You may define $H_i$ for $i=1,2,\dots$ recursively:

  • $H_{i-1}$ is infinite non-empty by hypothesis, so by the well-ordering property of $\mathbb N$ it must have a minimal element, say $h_{i-1}$.
  • Let $H_i = H_{i-1} \setminus\{h_{i-1}\}$.
  • If $H_i$ was finite, then $H_{i} \in H_i \cup \{ h_{i-1} \}$ would be a union of two finite sets so a finite set, a contradiction. Thus $H_i$ is infinite.
  • Observe $h_{i-1} \notin H_i$ and $H_i\subset H_{i-1}$.

Observe $H_0 \supset H_1 \supset H_2 \supset \cdots$. We will show in a moment that $\{h_i\}$ are all distinct.

Let $h: \mathbb N\to g(M)$ be defined by $h(n) = h_n$. Observe it is an injective map, otherwise $\exists i>j$ such $h(i) = h(j)$ i.e. $h_i = h_j$. Then $$h_j=h_{i} \in H_i \subset H_{i-1} \subset \cdots \subset H_{j+1}$$ but $h_{j} \in H_{j+1}$ which contradicts our observations. Thus $h$ must be injective and we have $|\mathbb N|\leq |g(M)|$. Clearly if we restrict $g$ to be $g:M\to g(M)$ it is a bijection so $|M| = |g(M)| \leq |\mathbb N|$ Is is clear from here?

Edit: added working with the image of $M$.

Edit: some typos.

Edit: format.

Steve
  • 555
  • Way over my head. Can u check my proof and point out things that are false? That would help me probably more in a first step because that is how i would think to prove the equivalence. –  Jun 21 '21 at 22:36
  • You claim that if a set $M$ is infinite then $|M|\geq |\mathbb N|$. If said claim is somewhere in you book or you have a way to prove it, then your proof is all good. – Steve Jun 21 '21 at 22:39
  • That was intuitive thinking due to laziness because it would make everything easier. No idea, I hope someone can help, but of course why would my textbook present a rather complicated proof if it was so easy possible. That makes it probable that it is no theorem. –  Jun 21 '21 at 23:08
  • Well, this is the very thing I have shown above. I don't know if there is a significantly simpler way. Note it only holds if you assume axiom of choice. Which I suppose you do. – Steve Jun 21 '21 at 23:15
  • So do you think that the following is a theorem: M is an infinite set iff $|\Bbb N| \le |M|$? Because if it is my proof works and my proof is much simpler than yours. –  Jun 21 '21 at 23:26
  • I changed my proof further. Are u more/completely convinced now? –  Jun 21 '21 at 23:57
  • You edited your first definition again, and now it is incorrect, again. There is a bijection $f': \mathbb N \to \set{2,3,4,5,\cdots}$ but it is not countably finite. And no, I am far from convinced. You can't just willy-nilly say "yeah I think this works" and pass it as a proof. If you claim something holds "naturally", you should be able to write it out, you know, naturally. – Steve Jun 22 '21 at 00:21
  • The second part of the first definition means: some M is countable iff you have a bijection between some n elements of IN and the elements of M which can be called countable finite in opposition to the first case where countable infinite is defined. –  Jun 24 '21 at 21:45
  • I think the latest version of my proof should be pretty close. If you answer my last question with YES then the proof actually is legit. –  Jun 24 '21 at 21:47