I'm working through a proof in which $(X,\tau)$ is a $T_1$-space that is metrizable. The author says, "since $(X,\tau)$ is metrizable, there is a countably infinite, closed (in $X$) and discrete subspace, $(Y,\tau_{|Y})$, of $(X,\tau)$."

Is this something I should know about or should I provide more details of what's going on and what I'm trying to prove so that I can get some help figuring this out.

  • $\begingroup$ If $X$ is compact, there cannot be an infinite, closed, discrete subset; did you mean to have non-compact instead? $\endgroup$ – Brian M. Scott Nov 15 '12 at 23:23
  • $\begingroup$ Yes, sorry. Assume $X$ is not compact. $\endgroup$ – user41728 Nov 15 '12 at 23:28

If $X$ is a metrizable space that is not compact, then there it has a countably infinite closed discrete subset $Y$. This follows immediately from the fact that a metrizable space is compact iff it is sequentially compact. Since $X$ is not compact, it is not sequentially compact, and there is therefore a sequence $\sigma=\langle x_n:n\in\Bbb N\rangle$ in $X$ with no convergent subsequence; let $Y=\{x_n:n\in\Bbb N\}$.

Let $x\in X$ be arbitrary; $\sigma$ has no subsequence converging to $x$, so there are an open nbhd $V_x$ of $x$ and an $m_x\in\Bbb N$ such that $x_n\notin V_x$ whenever $n\ge m_x$. Clearly $V_x\cap Y$ is finite, so $F_x=(V_x\cap Y)\setminus\{x\}$ is closed, and $U_x\triangleq V_x\setminus F_x$ is an open nbhd of $x$. $U_x\cap Y=\varnothing$ if $x\notin Y$, and $U_x\cap Y=\{x\}$ if $x\in Y$, so $Y$ is closed and discrete. If $Y$ were finite, there would be a $y\in Y$ such that $A\triangleq \{n\in\Bbb N:x_n=y\}$ was infinite; but then $\langle x_n:n\in A\rangle$ would be a subsequence of $\sigma$ converging to $y$, contradicting the choice of $\sigma$.


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.