Let $(x_n)$ be a sequence in a metric space converging to a point $x$. Prove that $S=\lbrace x_1, x_2, x_3, . . . \rbrace \cup \lbrace x \rbrace$ is sequentially compact.

I understand that there already exists a similar question (Show that $\lbrace x_n : n \in \mathbb{N} \rbrace \cup \lbrace x \rbrace$ is a compact subset of $(X,d)$). But In that question, the answers are all using the notion of open cover, which I have not learnt yet. So in my question, I need to prove that every sequence from $S$ has a subsequence which converges to an element of $S$. I am not sure how to do that. Can someone help me (with a few details), please?

  • $\begingroup$ Recall that every subsequence of a convergent sequence is also convergent. And note that any sequence in $S$ is a subsequence of $\{x_n\}.$ $\endgroup$ – Timon Apr 20 '17 at 8:36

Hint: For convenience, define $x_{\infty} := x$. Then you have a 'rank' map $r : S \to \mathbb{N} \cup \{\infty \} : x_k \mapsto k$. Now, let $y : \mathbb{N} \to S : k \mapsto y_k$ be a sequence. You can consider the compose map $r \circ y : \mathbb{N} \to \mathbb{N} \cup \{\infty\}$, which allows you somewhat to count how many times the sequence 'passes by' any point of $S$.

Here are two (non mutually exclusive) possibilities (you have to prove this): Either the map $r \circ y$ is constant on a subset $X \subset \mathbb{N}$ of infinite cardinality or it is strictly increasing on a subset $Y \subset \mathbb{N}$ (of infinite cardinality). To prove this, it might be useful to prove that you can always find a subset $Z \subset \mathbb{N}$ of infinite cardinality on which the function $y$ is a (not necessarily strictly) increasing function; This observation underlies the second part of timon's comment to your question$^1$. In the first case, the subsequence $ \left. y \right|_X$ is constant (and thus converges); In the second case, one has to show that the subsequence $\left. y \right|_Y$ converges to $x_{\infty}$ (which relies on the first part of timon's comment).

$^1$ To be a bit picky about timon's comment, I point out for instance that the sequence $(y_k = x_{\infty})_k$ is a sequence in $S$, but not a subsequence of the sequence $(x_n)_n$. The composition $x \circ r \circ y$ formalises in what sense $y$ could be (generally abusively) understood as a subsequence of $x$.


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.