[Math] Closure and limit of a sequence

metric-spacesreal-analysis

Let $E$ be a subset of a metric space $(S,d)$. I'm trying to show that an element is in $\overline{E}$ if and only if it is the limit of some sequence of points in $E$.

Suppose there is a sequence $(p_n) \subseteq E$. Then because $E \subseteq \overline{E}$, I know that $(p_n) \subseteq \overline{E}$. Since $\overline{E}$ is closed, $\overline{E}$ contains the limit of every convergent sequence of points in $\overline{E}$. So $p \in \overline{E}.$

This is how I approached the converse: I assumed $p \in \overline{E}$ and that there does not exist a sequence $(p_n) \subseteq E$ such that $p_n \to p$. Then I concluded that there exists $\varepsilon > 0$ such that $d(p, x) \geq \varepsilon \ \forall x \in E$.

Then there exists an open ball centered at $p$ that does not contain a point of $E$ (I wasn't sure how to make this statement rigorous). So $p \notin \overline{E}$.

Could I get some feedback on my proof?

Best Answer

You haven’t actually justified the conclusion that there is an $\epsilon>0$ such that $d(p,x)\ge\epsilon$ for each $x\in E$. Justifying it properly requires essentially the same work as a direct proof that if $p\in\operatorname{cl}E$, then there is a sequence $\langle p_n:n\in\Bbb Z^+\rangle$ in $E$ converging to $p$, so you might as well give the direct proof.

There are two cases.

  1. If $p\in E$, there’s a trivial sequence in $E$ that converges to $p$; what is it?

  2. If $p\in(\operatorname{cl}E)\setminus E$, you have to work just a little bit. For each $n\in\Bbb Z^+$ there’s a point $p_n\in E$ such that $d(p,p_n)<\frac1n$; why? What does this get you?