0

Attempt: Subspace of $R$ by removing a point has nontrivial clopen set, but Subspace of $R^n$ by removing a point has no clopen set. Since has clopen set is a topological invariant, the two subspace is not homeomorphism. So the original spaces are not homeomorphic.

But I cannot figure out why subspace of $R^n$ by removing a point has no nontrivial clopen set. Any help? Thanks.

The question not allows to specially use the definition of connected. Or it only allows to use that having no nontrivial clopen set is topological invariant. (Although having no nontrivial clopen set implies connected...)

BDN
  • 624
  • You certainly mean "not homeomorphic" and $n > 1$. – Paul Frost Jun 18 '18 at 23:00
  • Thanks, should be not. –  Jun 18 '18 at 23:02
  • I'm confused. In the body of your question you ask for help in showing that $\mathbb R^n$ minus a point is connected, but the title asks us not to use connected. How do you expect us to prove that something is connected without using the concept of connectedness? Or did you mean something else by "without using connected"? – bof Jun 18 '18 at 23:06
  • I mean do not specially use connected, since I think there are other way to show a certain space is connected without using clopen set. In this question it requires only to use clopen set. Or I should say only use clopen set to prove connected. –  Jun 18 '18 at 23:10
  • This is a very confusing question. – copper.hat Jun 19 '18 at 02:06
  • Inasmuch as "has no nontrivial clopen set" is the definition of what it means for a space to be connected, I can't imagine what it would mean to prove that something has no nontrivial clopen set without proving that it's connected. – bof Jun 19 '18 at 02:50
  • @bof I think that the question is asking to avoid the use of (path connected $\Rightarrow$ connected). – Michael Burr Jun 19 '18 at 02:56

1 Answers1

0

First, your question is somewhat misleading:

The condition that you're asking for is equivalent to connectedness. In particular, a space $X$ is disconnected if there exist nonempty open sets $U$ and $V$ so that $X=U\cup V$ and $U\cap V=\emptyset$. Moreover, if $W$ is a nontrivial clopen set, then $W$ and $W^c$ are both open, $X=W\cup W^c$, and $W\cap W^c=\emptyset$. Therefore, they form a separation of $X$. Hence, by asking for a nontrivial clopen set, you are asking for a separation and using the definition of disconnected.

What I think that you are asking is to avoid the implication path connected implies connected. So, to prove that $\mathbb{R}^n$ is connected without going through path connected.

Suppose that $\mathbb{R}^n\setminus\{0\}$ has a nontrivial clopen set $X$. Then, since $\mathbb{R}^n\setminus\{0\}$ has the induced topology, there is an open set $U$ in $\mathbb{R}^n$ and a closed set $D$ in $\mathbb{R}^n$ so that $U\cap \mathbb{R}^n\setminus\{0\}=X$ and $D\cap\mathbb{R}^n\setminus\{0\}=X$. We note that $U$ and $D$ only differ by a single point $\{0\}$ since their intersection with $\mathbb{R}^n\setminus\{0\}$ is the same. Moreover, the two sets must differ because $\mathbb{R}^n$ is connected and has non nontrivial clopen sets.

Let $x\in X$ and let $y\not\in X$. By perturbing $x$ and $y$ slightly, we may assume that the line $\ell$ though $x$ and $y$ does not include $0$ (since $X$ is clopen, there are small neighborhoods around $x$ in $X$ and around $y$ in in $X^c$). Then the induced topology from $\mathbb{R}^n$ to $\ell$ produces the standard topology on $\ell\simeq\mathbb{R}$. We observe that $U\cap \ell=D\cap \ell$, since their only difference is at $0$. Since $x\in U\cap\ell$ and $y\not\in U\cap \ell$, this defines a nontrivial clopen set in $\mathbb{R}$, which contradicts the fact that $\mathbb{R}$ is connected.

Michael Burr
  • 32,867
  • You can simplify your proof. The sets $U$ and $D$ are not needed. Take your line $\ell$ which avoids $0$. Then $\ell$ is homeomorphic to $\mathbb{R}$ and $\ell \cap X$ is a nontrivial clopen subset of $\ell$ which is impossible. However, it seems to me that the argument is just a hidden version of "path-connected" implies "connected" (the connectedness of real intervals is the essence of that implication). – Paul Frost Jun 19 '18 at 09:25
  • @PaulFrost I agree with both of your comments. If I suppress the sets $U$ and $D$, they are still there behind the scenes since all the topologies are induced. From the OP's question, I thought that a little bit more detail would be helpful for instructive purposes. – Michael Burr Jun 19 '18 at 12:53