Consider the following statement (in $\mathsf{ZF}+\text{AC}_\omega (\mathbb{R})$):

There exists $(\varphi_\alpha)_{\alpha\in\omega_1}$ with $\varphi_\alpha : \alpha \rightarrow \mathbb{N}^\mathbb{N}$ injective and $\text{ran}(\varphi_\alpha)$ is closed and $\text{rank}_{CB}(\text{ran}(\varphi_\alpha)) = 1$, i.e. $\text{ran}(\varphi_\alpha)$ is made of isolated points, for all $\alpha\in \omega_1$.

where $\mathbb{N}^\mathbb{N}$ is the Baire space (the space of infinite sequences of natural numbers with its usual topology) and $\text{rank}_{CB}$ is the Cantor-Bendixon rank. The statement trivially holds if we assume $\text{AC}_{\omega_1}$.
My questions are:

  1. How much choice do we need (at least) to prove the above statement? Or more generally what known "weak" axiom (on top of $\mathsf{ZF}+\text{AC}_\omega (\mathbb{R})$) can be assumed to prove the statement?
  2. Is it equivalent to the statement "There exists an $\omega_1$ subset of the reals"? In case it is not, is it consistent $\mathsf{ZF}+\text{AC}_\omega (\mathbb{R})+$ The above statement $+$ "It does not exists an $\omega_1$ subset of the reals"?


  • 2
    $\begingroup$ By the way, you might be interested to know that existence of an $\omega_1$-sequence of reals is equivalent to the weaker assertion that that there is a choice of injections $(\varphi_{\alpha})_{\alpha<\omega_1}: \alpha \rightarrow \mathbb{R}.$ $\endgroup$ Nov 14, 2021 at 1:10

1 Answer 1


Your statement is equivalent to the assertion that there is a function choosing an enumeration of every countable ordinal. From an enumeration of $\alpha,$ you can easily inject it into a countable set of isolated points in Baire space. For the other direction, if $\varphi_{\alpha}$ injects $\alpha$ into Baire space, with the range a set of isolated points, there is a surjective partial map from basic open sets onto $\alpha,$ by sending $U \mapsto \beta$ if $\beta$ is unique such that $\varphi_{\alpha}(\beta) \in U.$ From this we can easily enumerate $\alpha.$ Note that by the Cantor-Bendixson analysis, any closed countable set of reals can be canonically enumerated, by sending basic open sets to the unique point in $U$ of maximal rank if there is such a point.

This principle implies there is an $\omega_1$-sequence of reals, since an enumeration of an ordinal can be canonically coded by a real. It is strictly stronger than the existence of an $\omega_1$-sequence of reals. For example, it clearly implies $\omega_1$ is regular, while in Figura's model ($\mathcal{M}36$ in Consequences of the Axiom of Choice), there is an $\omega_1$-sequence of reals yet $\omega_1$ is singular.

Note that everything above is purely in ZF. The possibility remains that countable choice for reals plus an $\omega_1$-sequence of reals implies your principle, though that seems unlikely. I would guess adding $\omega_1$ Cohen reals to the Solovay model would result in a model with an $\omega_1$-sequence of reals and DC but no choice of enumerations for every countable ordinal. I have not verified this however.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.