The question:

Let $X$ be a space such that every continuous function $f:X\rightarrow\mathbb{R} $ does have the following property:

if $a<c<b$, $f(x) =a$, and $f(y) =b$, then there exists $z\in X$ such that $f(z) = c$. Prove that X is connected.

My answer:

All continuous functions have the IVP which means that they (the continuous functions) do not map to a two point discrete space which implies connectedness.

Is this correct?

  • $\begingroup$ "they do not map" ? What are "they"? $\endgroup$ Jun 21, 2016 at 6:31
  • $\begingroup$ the continuous functions $\endgroup$ Jun 21, 2016 at 6:31
  • $\begingroup$ Gary, are you there? $\endgroup$
    – Wlod AA
    Sep 28, 2019 at 5:22

3 Answers 3


I think what's missing in your proof is showing that, if $X$ is not connected, there is some continuous function that maps $X$ to a two point discrete space (hint: what can you say of the connected components of $X$?).

  • $\begingroup$ I'm using the definition "X is connected if any continuous function from X to {0,1} is constant". Since the IVP property doesn't it follow immediate that any such function would be constant?! $\endgroup$ Jun 21, 2016 at 7:06
  • $\begingroup$ With that definition and appropriate wording, I think your proof is fine. $\endgroup$ Jun 21, 2016 at 7:11

You are on the right track, but perhaps you should explain how your answer relates to functions from $X$ to the reals.

If $X=A\cup B$ where $A, B$ are each open and non-empty with $A\cap B=\emptyset$ , let $f(x)=0$ for $x\in A$ and $f(x)=1$ for $x\in B.$

If $U$ is an open subset of the reals then $f^{-1}U$ is one of $\emptyset, A,B,X$, so $f^{-1}U$ is open in $X,$ so $f$ is continuous.

Since $A\ne \emptyset \ne B$, there exists $x\in A$ and $y\in B,$ that is, $f(x)=0$ and $f(y)=1.$ Obviously no $z\in X$ satisfies $f(z)=1/2.$

So your condition on the set of continuous $f : X\to R$ implies that $X$ must be connected.... Note that the converse also holds, because jf $X$ is connected, then any continuous image of $X$ is connected.


Your conjecture is FALSE even in a much stronger version. I'll let you think about it but it's not completely trivial.

It's possible that space $\ X\ $ which has more than one point has no non-trivial connected components (i.e. only 1-point components), and still your Darboux kind of property may hold.


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.