In $ZF$ without any form of $AC$, is it true the following statement: "A well-ordered set $S$ is Dedekind-infinite iff it is finite" ?

I know that a finite set is always Dedekind-finite, and that if $AC$ (or countable choice) holds, then also the converse is true. What about if we have a well-order only on $S$?

I think that this is true, as I found here: (i) implies (iv) and in the answer here, but I am looking for a proof. I have the following one: can you confirm that it does not use (even indirectly) $AC$ in any point?

Let $S$ Dedekind-finite and well ordered. Then it is isomorphic through $f$ to a von Neumann ordinal $\alpha$. If $\alpha$ were not a finite ordinal, then it would contain the least infinite ordinal $\mathbb{N}$. But $\mathbb{N}$ is Dedekind-infinite and then, by the bijection $f$, $S$ would contain a Dedekind-infinite subset $T$. But this would imply that $S$ is Dedekind-infinite, since the non-surjective injection $\varphi$ of $T$ to all $S$ by posing $\varphi(x)=x$ for $x \in S \setminus T$.

  • $\begingroup$ If $<$ well-orders $S,$ and $S$ is Dedekind-finite then any non-empty $T\subset S$ has a $<$-max member. Otherwise if $f(x)=\min_<\{y\in T: x<y\}$ when $x\in T,$ and $f(x)=x$ when $x \in S$ \ $T$ then $ f:S\to S$ is injective with $\min_< T \not \in f[S].$ $\endgroup$ Commented Mar 27, 2020 at 8:30

1 Answer 1


Yes, your proof is correct. One theorem you should take note of that would simply your proof a bit towards the end is that a set is Dedekind-infinite if and only if it has a countably infinite subset (in fact this is sometimes used as the definition). Your last sentence proves one direction of this equivalence.

For the other direction (which you don't need for this proof), let $f$ be a non-surjective injection $S\to S$ and let $x\in S$ not in its image, and then show that $ \{f^n(x): n\in \omega\}$ is a countably infinite subset of $S.$

  • $\begingroup$ Yes, I know this, and in fact I proved directly one direction of the equivalence. I wanted to make (possibly) a self-contained proof. Thanks! $\endgroup$
    – Eupho
    Commented Mar 15, 2020 at 8:04

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .