Given topological spaces $X$ and $Y$, where $Y$ is connected, let $p \colon X \to Y$ be a quotient map. If, for each point $y \in Y$, the set $p^{-1}(\{y\})$ is connected, then how to prove that $X$ is connected also?

By the map $p$ being a quotient map is meant the following:

The map $p$ is surjective, and, for any subset $V$ of $Y$, the set $V$ is open in $Y$ if and only if $p^{-1}(V)$ is open in $X$.

$X$ is a union of disjoint connected sets, $X_i$ (the equivalence classes). If $X$ is disconnected and $U \cup V$ is a separation of $X$ (with $U$, $V$ disjoint and open in $X$), then there are two possible cases:

Case 1: One of the equivalence classes, say $X_i$, is split between $U$ and $V$: enter image description here

But then, $X_i \cap U$ and $X_i \cap V$ are disjoint and open in $X_i$, so they are a separation of $X_i$, contradicting the assumption that all $X_i$'s are connected.

Case 2: Every equivalence class is entirely contained in either $U$ or $V$: enter image description here

But then, $p(U)$ and $p(V)$ are disjoint. Since they are also open in $Y$, they are a separation of $Y$, contradicting the assumption that $Y$ is connected.

In sum, if $X$ is disconnected then either one of the equivalence classes is disconnected or the quotient space is disconnected.

Suppose $X$ were not connected. Let $X = U \cup V$, where both sets are open (and closed) in $X$, disjoint and non-empty. Then $f[U]$ is open and closed in $Y$, as $f$ is a quotient map and $U = f^{-1}[f[U]]$: it's clear that $U\subseteq f^{-1}[f[U]]$, but if the reverse inclusion would fail, we'd have a $p \in V$ with $f(p) = f(q)$ for some $q \in U$. But then $U,V$ would disconnect the connected set $f^{-1}[\{f(p)\}]$, which cannot be. Similarly $f(V)$ is open and closed. Both $f(U)$ and $f(V)$ are non-empty and they are disjoint, as otherwise again the inverse image of a point in their intersection would be a disconnected fibre (using $U$ and $V$ again). So this contradicts the connectedness of $Y$.

  • You wrote that $U=f^{-1}(f(U))$, this is true if $f$ is injective... however we are assuming $f$ is surjective (since it's a quotient map), the analogous result is that $U=f(f^{-1}(U))$....so this needs some cleaning up. – Squirtle Oct 28 '15 at 19:18
  • 1
    @Squirtle I show that $U = f^{-1}[f[U]]$ because all fibres are connected (using that $U$ and $V$ form an open partition of $X$). I don't use surjectivity or injectivity. If $A = f^{-1}[f[A]]$ for all $A$ then $f$ is 1-1. But we only need it for the specific $U$ and $V$. – Henno Brandsma Oct 28 '15 at 19:22
  • Thank you! That makes sense. – Squirtle Oct 28 '15 at 19:27

Let $f$ be any continuous map from $\rm X $ to $\{0,1\}$.

As $f$ is continuous, it induces a function $\overline{f} : \rm Y \to \{0,1\}$. Buy $\rm Y$ is connected so this function must be constant, let say $\overline{f} = 1$.

Now let $x$ be any element of $\rm X$. It lies in the fibre $\rm X_{p(x)}$ which is connected by assumption, so $f$ restricted to $\rm X_{p(x)}$ is constant. But $\overline{f}(\overline{x}) = 1$ so $f(x)$ must be equal to $1$.

We have proved that $f$ is in fact constant and so, $\rm X$ is connected.

  • Thank you for your answer, but would you please elaborate on how $f$ induces the continuos function $f$ bar? And do you mean "inverse image" of the image $p(x)$ when you use the word "fibre"? – Saaqib Mahmood Feb 13 '13 at 11:47
  • Yes, this is the meaning of fibre. And for $\overline{f}$ it is done in any book. – Damien L Feb 13 '13 at 11:54

Let $f$ be a continuous function from $X$ to $[2]=\{1,2\}$ with the discrete topology. The fibres $p^{-1}(y)$ are just the equivalence classes under the relation $x\sim x'\leftrightarrow p(x)=p(x')$. Since they are connected, we have $f(x)=f(y)$ whenever $x\sim y$. Hence $f$ induces a continuous function $\hat f:Y\to[2]$ defined by $\hat f(p(x))=f(x)$. Since $Y$ is connected, $\hat f$ is constant and so is $f$.
We have shown that an arbitrary continuous map $f:X\to[2]$ is constant. This implies that $X$ is connected.

  • But how come the map $f$ bar is continuous? Please show this rigorously. – Saaqib Mahmood Feb 13 '13 at 11:55
  • We have $\hat f\circ p=f$. If the composition $f$ is continuous then $\hat f$ is continuous, too, due to the fact that $Y$ carries the final topology with respect to $p$. – Stefan Hamcke Feb 13 '13 at 12:02

Your Answer

 

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.