I need help to understand definition of c-closed. In here https://www.sciencedirect.com/science/article/pii/0166864180900279, author said that $X$ is c-closed iff every countably compact subset of $X$ is closed. Equivalently, every non-closed subset $A$ of $X$ contains a sequence which has no a cluster point in $A$.

I can not understand the equivalence. Is every non-closed containing a sequence which has no a cluster point in $X$? Then every space is c-closed?

I know it's wrong. For example, the ordinal space $\omega_1+1$ with the order topology is not c-space since $\omega_1$ is countably compact but isn't closed. How to understand this space isn't c-closed using second definition?


The cluster point of the non-closed set $A$ is considered to be in $A$, as you first wrote it.

The second definition is just the contrapositive of the first:
If every sequence of $A$ has a cluster point in $A$, then it is countably compact, hence it must be closed if $X$ is c-closed.

For the specific example, we would need to find a sequence in $\omega_1$ that tends to $\omega_1$ (the only missing point), which doesn't exist, as $\omega_1$ is regular.


A set $A$ is countably compact iff every countable subset (or sequence of distinct points, if you prefer) of $A$ has a cluster point in $A$.

The alternative formulation is

$(\ast$) For all $A \subseteq X$: if $A$ is not closed there exist $a_n \in A$ such that $A$ contains no cluster point for $(a_n)$.

The proposed equivalence is simply a matter of contrapositive reasoning:

Suppose $X$ is $C$-closed. Let $A$ be non-closed. This means that $A$ is not countably compact. So there is a sequence $x_n$ from $A$ without a cluster point in $A$. Hence $X$ obeys ($\ast$). For the reverse, suppose $X$ obeys the property $(\ast)$. Then $X$ is $C$-closed: let $A$ be countably compact. Then $A$ is closed. For suppose it were not closed, then by the assumption $(\ast)$ on $X$ we'd have a sequence $a_n$ from $A$ without a cluster point in $A$. This contradicts $A$ being countably compact. So $A$ is closed and $X$ is $C$-closed.

The example $X= \omega_1+1$ is no contradiction. It is indeed not $C$-closed. The negation of $C$-closed in the alternative formulation ($\ast$) is:

There exists a subset $A$ of $X$ that is not closed, but such that all sequences from $A$ have a cluster point in $A$.

And indeed $A = \omega_1$ is as required. So $X$ is also not $C$-compact in the other definition ($\ast$).

I think it's mostly a matter of taking the negation of the alternative definition in the right way.


Terminology: "Open" and "closed" mean "open in $X$" and "closed in $X$".

For one direction: Let $X$ be $c$-closed and let $A\subset X$ be non-closed. Then $A$ is not countably compact, so let $C=\{C_k: k\in \Bbb N\}$ be an open cover of $A$ with no finite sub-cover.

For $k\in \Bbb N$ let $D_k=\cup_{j\leq k}C_j$ and $E=\{D_k: \neg [\exists k'<k\;(A\cap D_{k'}=A\cap D_k)]\;\}.$

We can enumerate $E$ as $E=\{E_n:n\in \Bbb N\}$ where $A\cap E_n\subsetneqq A\cap E_{n+1}$ for each $n\in \Bbb N.$ For each $n\in \Bbb N$ choose $a_n\in A\cap (E_{n+1}\backslash E_n).$

Any $a\in A$ belongs to $E_{F(a)}$ for some $F(a)\in \Bbb N$ but the set $\{a_n:n>F(a)\}$ is disjoint from $A\cap E_{F(a)}.$ So $F(a)$ is a nbhd of $a$ such that $\{n: a_n\in E_{F(a)}\}$ is finite. So $a$ is not a cluster point of the sequence $(a_n)_{n\in \Bbb N}.$

Addendum: Re the space $\omega_1+1$ with the $\epsilon$-order topology: If $(a_n)_{n\in \Bbb N}$ is any sequence of members of $A=\omega_1$ then $\{a_n:n\in \Bbb N\} \subset B=(\cup \{a_n:n\in \Bbb N\})+1\in \omega_1.$ The sub$space $ $ B $ is compact & Hausdorff so $(a_n)_{n\in \Bbb N}$ has a cluster point $a\in B\subset A.$


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.