Prove that $f$ is closed if and only if $f(\text{cl}(A)) \supseteq \text{cl}(f(A)).$ I can show the forward directions by saying:
Suppose $f$ is closed. Then, $f(\text{cl}(A)) = f(A) \cup f(\partial A)$ which is closed implying the inclusion of the smallest closed set containing $f(A)$ in $f(\text{cl}(A)).$

But I am struggling going in the other direction. I want to say take a closed set $B \subset X.$ $B = \text{cl}(A)$ for some open $A \subset X.$ But then I am stuck.

Thanks in advance


For the forward direction there’s no need to deal with the boundary of $A$: just observe that if $f$ is closed, then $f[\operatorname{cl}A]$ is a closed set containing $f[A]$ and must therefore contain $\operatorname{cl}f[A]$, since that’s the intersection of all closed sets containing $f[A]$.

For the other direction it’s probably easiest to suppose that $f$ is not closed and try to find a set $A\subseteq X$ such that $f[\operatorname{cl}A]\nsupseteq\operatorname{cl}f[A]$. If $f$ is not closed, there is a closed set $F\subseteq X$ such that $f[F]$ is not closed. Clearly $f[\operatorname{cl}F]=f[F]$, but $f[F]$ is not closed, so ... ?

  • $\begingroup$ on an intuitive level, it seems this theorem is saying closed maps expand the boundaries of a set. Am i correct in this intuition? $\endgroup$ – masszz Sep 29 '13 at 22:34
  • $\begingroup$ @sergey: In general I don’t find it useful to think in terms of boundaries at all. If anything it says that a closed map shrinks the closure (and hence the boundary). $\endgroup$ – Brian M. Scott Sep 29 '13 at 22:41
  • $\begingroup$ Another way to do the other direction: Let $A$ be closed. $\operatorname{cl} A = A$. Thus $f(A) \supset \operatorname{cl}(f(A))$. Hence $f(A)$ is closed. $\endgroup$ – Ayman Hourieh Sep 29 '13 at 22:56
  • $\begingroup$ @BrianM.Scott Brian doesn't it say that f takes a closed set and maps it something larger than the closure of the interior, hence EXPANDING the original closed set? $\endgroup$ – masszz Sep 30 '13 at 0:56
  • $\begingroup$ @sergey: It says that the closure of the image of $A$ may be smaller than the image of the closure of $A$, which I see as a kind of shrinkage. $\endgroup$ – Brian M. Scott Sep 30 '13 at 1:01

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.