I am attempting to solve exercise 1.10.1 from Topology and Geometry by Bredon:

If $X$ is a compact Hausdorff space then show that its quasicomponents are connected (and hence that its quasi-components coincide with its components).

The definition of the relation used for quasi-components is "$d(p) = d(q)$ for every discrete valued map $d$ on $X$."

He gives the hint:
If $C$ is a quasi-component, let $C = \cap C_{\alpha}$ where the $C_\alpha$ are the clopen sets containing $C$. If $C$ is disconnected, then $C = A \cup B, A \cap B = \emptyset, A,B$ closed. Let $f:X \to [0,1]$ be $0$ on $A$ and 1 on $B$. Put $U = f^{-1}([0,\frac{1}{2}))$ and apply the following result:
Let $X$ be a compact space and let $\{C_\alpha \mid \alpha \in A\}$ be a collection of closed sets, closed with respect to finite intersections. Let $C = \cap C_\alpha$ and suppose that $C \subset U$ with $U$ open. Then $C_\alpha \subset U$ for some $\alpha$.

I don't understand which collection of closed sets to use. If we use the clopen sets $C_\alpha$, and we have that $C \subset U$ but $U \subset X - B$, so $B = \emptyset$, and we don't need some $C_\alpha \subset U$. Furthermore, how would we show that $C \subset U$, it seems to more or less assumes $B = \emptyset$.


I think the hint is somewhat misleading. In fact, the clopen sets $C_\alpha$ containing $C$ are the correct choice, but the choice of $U$ is inadequate. So let us do it properly.

The quasi-component $C = C(x)$ of a point $x \in X$ is the intersection of all clopen subsets of $X$ containing $x$. In particular $C$ is a closed set. Let $\mathfrak C = \{ C_{\alpha} \mid \alpha \in A\}$ denote the set of clopen subsets of $X$containing $x$.

If $C$ is disconnected, then $C$ is the union of two non-empty disjoint closed sets $A, B \subset C$. W.l.o.g. assume $x \in A$. Note that $A,B$ are closed in $X$ since they are closed subsets of the closed $C$.

Since $X$ is normal, we find open neighborhoods $V$ of $A$ and $W$ of $B$ such that $V \cap W = \emptyset$. We have $C = \bigcap_\alpha C_\alpha \subset U = V \cup W$. We claim that $C_\alpha \subset U$ for some $\alpha$. In fact, $\bigcap_\alpha (C_\alpha \cap (X \setminus U)) = (\bigcap_\alpha C_\alpha) \cap (X \setminus U) = C \cap (X \setminus U) = \emptyset$. The sets $C_\alpha \cap (X \setminus U)$ are compact, hence the finite intersection property applies to show that there are finitely many $\alpha_i \in A$ such that $\bigcap_i (C_{\alpha_i} \cap (X \setminus U)) = (\bigcap_i C_{\alpha_i}) \cap (X \setminus U) = \emptyset$. But clearly $C^* = \bigcap_i C_{\alpha_i} \in \mathfrak C$ and $C^* \subset U$.

We conclude $$\overline {C_\alpha \cap V} \subset \overline C_\alpha \cap \overline V = C_\alpha \cap \overline V = (C_\alpha \cap U) \cap \overline V = C_\alpha \cap (U \cap \overline V) = C_\alpha \cap V, $$ thus $\overline {C_\alpha \cap V} = C_\alpha \cap V$. Hence $C_\alpha \cap V$ is also clopen. It contains $A$ and a fortiori $x$. Thus $C \subset C_\alpha \cap V \subset V$ which is impossible.

  • $\begingroup$ why are $A,B$ closed in $X$? @PaulFrost $\endgroup$ Jun 6 '20 at 16:22
  • 1
    $\begingroup$ @RubenBaeza $A,B$ are closed in $C$ and $C$ is closed in $X$. $\endgroup$
    – Paul Frost
    Jun 6 '20 at 16:52
  • $\begingroup$ @PaulFrost Maybe it is a dumb question but why $\bigcap_{\alpha} C_{\alpha} \subset U$ implies that $C_\alpha \subset U$ for some $\alpha$? $\endgroup$
    – Zanzag
    May 1 '21 at 17:00
  • $\begingroup$ @Zanzag See my update. $\endgroup$
    – Paul Frost
    May 1 '21 at 23:00

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.