Assume we have a metric space $X$, a subset $E\subseteq X$, and a limit point $p$ of $E$.

Proofwiki and Rudin both "construct" a sequence that converges to $p$ using the fact that every neighborhood of radius $\frac1n$ ($n = 1, 2, 3, ...$) is nonempty because $p$ is a limit point, and picking an arbitrary point in $N_{\frac1n}(p)$ as $s_n$.

What I'm not sure about is that since there's infinitely many points in each neighborhood, is whether they've really proven a sequence exists that converges to $p$ exists. I understand you can keep getting points closer to $p$, we have pleanty of points to make a sequence out of.

Is that enough, or am I right in feeling like the Axiom of (countable, right?) Choice has been quietly used to actually say any such sequence exists?


It is a construction proof.

You can use a similar condition that you let D$_1$ = 1, and then you know there is an element of E, x$_1$, distinct from p so d(x$_1$,p) < 1. Now you can choose D$_2$ = Min(d(x$_1$,p),1/2). Again, you know that since every nbd of p contains a point in E distinct from p, there is a point x$_2$ (distinct from x$_1$) in E, such that d(x$_2$,p)< D$_2$.

If you keep doing this, you will get a sequence of elements from E, (x$_n$) that converge to p since for every $\epsilon$ > 0, choose N to be a natural number such that N > 1/$\epsilon$ which guarantees that for all n>N, you will have d(x$_n$,p) < D$_n$ and we know by construction D$_n$ $\leq$ 1/n < 1/N < $\epsilon$.

  • $\begingroup$ You do get a sequence of distances, but not explicitly a sequence of points, right? I edited the original question a little to better explain what I'm wondering about. $\endgroup$ – Jake Oct 14 '13 at 1:36
  • $\begingroup$ The axiom of choice is really not a necessity here, as the defn of limit point guarantees that for every nbd of p, a limit point of E, there exists a point x in E, distinct from p, in the nbd. You are using more the properties of open sets than AC. To answer your other question, yes you do get a sequence of points that converges to p. The definition of "limit(x$_n$) = p" says that for every $\epsilon$ >0, there exists a natural number N such that for all n>N, d(x$_n$,p) < $\epsilon$. If you go back and look at our construction you will see that the sequence (x$_n$) satisfies this condition. $\endgroup$ – Matt Brenneman Oct 14 '13 at 2:49

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.