Let $X$ be a metric space and assume that, for every $\varepsilon>0$ there is a countable open cover $(A_i)$ of $X$ with $diam(A_i)\le \varepsilon$ for each $i$.

Of course I can assume the cover is made only of basic open sets: if $\varepsilon$ is fixed then, for each $i$, we can pick $x_i\in A_i$ and consider the ball $B_i$ with center $x_i$ and diameter $2\varepsilon$ to obtain a $2\varepsilon$-cover of $X$.

The question is: can I assume the open cover is made of open balls even without $AC$ (or, in this case, without $AC(\omega)$)? Is there a way to build a ball cover from a generic open cover that is not dependent on the choice of a point for each set in the cover?

  • 1
    $\begingroup$ I don't get your question. What is the statement you want to prove without choice? $\endgroup$ – Asaf Karagila Oct 6 '18 at 17:12
  • $\begingroup$ Passing from a generic open cover to a ball cover. Let me rephrase it in a clearer way $\endgroup$ – Manlio Oct 7 '18 at 9:59

The answer is unfortunately negative, at least if you want to keep the cover countable.

In Cohen's first model there is an infinite Dedekind-finite set of reals which is dense in $\Bbb R$. This is a set which is infinite, but has no countably infinite subset. And as a set of reals, it is of course a metric space, and by being dense, it is also not bounded.

Now, cover $A$ with rational intervals (or their intersection with $A$) with diameter $\varepsilon$, this is a countable cover. But if you want these to be open balls of diameter $\varepsilon$, you need to choose centers for these balls, which would necessitate a countably infinite subset of $A$.


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.