I'm studying Intro to Topology by Mendelson.

The problem statement is,

Let $A,B\subset X$, $X$ a topological space. If $A$ is connected, $B$ open and closed, and $A\cap B\neq\emptyset$ then $A\subset B$.

My proof is,

By way of contradiction, suppose that $A$ is not a subset of $B$. Then there exists an $a\in A$ such that $a\in C(B)$. Consider the sets $P=A\cap B$ and $Q=A\cap C(B)$. Note that both $P$ and $Q$ are nonempty and open. Thus, we have that $A\subset A\cap B\cup A\cap C(B)$ and $P\cap Q=A\cap B\cap C(B)=A\cap\emptyset=\emptyset\subset C(A)$. Also, $P\cap A\neq\emptyset$ and $Q\cap A\neq\emptyset$. Therefore, $A$ is disconnected, which is a contradiction, since $A$ is assumed connected.

I'm not sure if this was even the right approach, but it's my best shot so far.

Thanks for any hints or feedback!

  • $\begingroup$ $A\cap B$ is open and closed as a subset of $A$ with the relative topology. So either $A\cap B=\emptyset$ or $A\cap B=A$. $\endgroup$ – egreg Jul 24 '13 at 9:21
  • $\begingroup$ This could surely made less indirect and also you could take $P=B$ and $Q=C(B)$ without intersecting them with $A$ first. $\endgroup$ – Hagen von Eitzen Jul 24 '13 at 9:23
  • $\begingroup$ @HagenvonEitzen, I didn't think of approaching it that way. I'll try that and see where I get. Thanks for the response. $\endgroup$ – Shant Danielian Jul 24 '13 at 9:29
  • $\begingroup$ @Hagen: Whether you can take $P=B$ and $Q=X\setminus B$ depends on how disconnected has been defined and on what’s been proved to be equivalent. $\endgroup$ – Brian M. Scott Jul 24 '13 at 9:29

Your argument is essentially correct, but the wording needs a bit more care in one place. When you say that the sets $P=A\cap B$ and $Q=A\setminus B$ are open, the default reading is ‘open in $X$’, which isn’t necessarily the case. You want to say that $P$ and $A$ are open in $A$. Specifically, $P$ and $Q$ are non-empty disjoint relatively open subsets of $A$ whose union is $A$, and therefore $A$ is not connected.

Added: If you’ve not already done so, you might find it useful for future reference to prove that a set $A$ in a space $X$ is disconnected iff it has a non-empty, proper subset that is clopen in the relative topology on $A$. Here you could have applied that result immediately: $B\cap A$ is clearly non-empty and relatively clopen in $A$, so if $A$ is connected ...

  • $\begingroup$ Would I already know by default that $P$ and $Q$ are open in $A$ or would I show that? It's not absolutely clear how that's the case, but I would have to think about it first. $\endgroup$ – Shant Danielian Jul 24 '13 at 9:35
  • $\begingroup$ @Shant: $B$ is open in $X$, so by the definition of the relative topology $B\cap A$ is open in $A$. Similarly, $X\setminus B$ is open in $X$, so ... ? $\endgroup$ – Brian M. Scott Jul 24 '13 at 9:37
  • $\begingroup$ Which would mean that $(X-B)\cap A$ is also relatively open in $A$. Thank you for helping connect the dots. $\endgroup$ – Shant Danielian Jul 24 '13 at 9:42
  • $\begingroup$ @Shant: Yep, you’ve got it. You’re welcome. $\endgroup$ – Brian M. Scott Jul 24 '13 at 9:45
  • $\begingroup$ @BrianM.Scott Since the set B is open and closed in X, I can conclude that the set B∩A is open and closed in A. But A is connected, then B∩A should be either empty or the whole set A. Is that right? $\endgroup$ – Melissa Oct 24 '17 at 17:01

It's simpler if you consider that $A\cap B$ is open and closed in the subspace $A$. Thus, since $A$ is connected, either $$ A\cap B=\emptyset\quad\text{or}\quad A\cap B=A $$

  • $\begingroup$ How do you know that $A\cap B$ is open and closed in the subspace $A$? Are you saying that both $A$ and $B$ are open and closed in the subspace $A$? Yet, $B$ is open and closed in $X$. Is it necessarily open in $A$ as well? $\endgroup$ – Shant Danielian Jul 24 '13 at 9:32
  • $\begingroup$ @ShantDanielian $A\cap B$ is open in $A$ by definition of relative topology. If $C$ is closed in $X$, then it's easy to see that $A\cap C$ is closed in $A$ (with respect to the relative topology). $\endgroup$ – egreg Jul 24 '13 at 9:38
  • $\begingroup$ Thank you egreg for your response, it makes more sense now. $\endgroup$ – Shant Danielian Jul 24 '13 at 9:43

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.