I have this theorem that says:

$f$ is homeomorphism iff $A\subseteq X$ open iff $f(A)\subseteq Y$ is open.

Can I say it differently like this:

$f$ is homeomorphism iff $f^{-1}(B)\subseteq X$ open iff $B\subseteq Y$ is open?

If I say $f(A)=B$, then $f^{-1}(B)=f^{-1}(f(A))=A$.

I cannot say that there would be any problem, but I am not sure.

I do not understand why this theorem is usually stated like that (first way).

  • 2
    $\begingroup$ So any map between indiscrete topologies is an isomorphism? Maybe it has to be a bijection? $\endgroup$
    – Ilya
    Nov 13, 2015 at 10:26
  • $\begingroup$ Please fix your grammar! $\endgroup$ Nov 13, 2015 at 11:08
  • $\begingroup$ both ways are valid (for bijections). I don't see a reason to consider one better than the other. $\endgroup$ Nov 13, 2015 at 11:15

1 Answer 1


You need to know that $f$ is a bijection. Assuming this is the case throughout, the equivalence does hold.

Suppose that $f$ is a homeomorphism. Suppose $f^{-1}[B]$ is open in $X$. As $f$ is a bijection we know that $B = f[f^{-1}[B]]$ and $f$ is an open map as well, so $B$ is open as the image under $f$ of an open set. If $B$ is open in $Y$, continuity of $f$ will already give that $f^{-1}[B]$ is open in $X$. So the equivalence holds, assuming $f$ is a homeomorphism.

Suppose we know that for all $B \subseteq Y$, $B$ open iff $f^{-1}[B]$ open, and we already know $f$ is a bijection (!), then we conclude that $f$ is a homeomorphism: $f$ is continuous because of the implication $B$ open then $f^{-1}[B]$ open, and $f$ is open because if $O$ is open in $X$, we can write (by being a bijection!) as $f^{-1}[f[O]]$, so applying the implication the other side: we know for $B = f[O]$ that $f^{-1}[B] = O$ is open, so $B = f[O]$ is open.

We need the bijection condition because any map between discrete spaces would be a homeomorphism, while only bijections can be homeomorphisms in the first place.

A map $f$ such that for all $B \subseteq Y$: $B$ open iff $f^{-1}[B]$ open, is called a quotient map.

So the statement I just proved is just: a bijective map is a homeomorphism iff it is a quotient map. But there are many non-bijective quotient maps.


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.