On this post:

compact set always contains its supremum and infimum

People ask if compact set always contains its supremum and infimum. I know that it is true on $\mathbb{R}$ with the usual topology.

Pf: Let $K\subseteq \mathbb{R}$ be a nonempty compact set. Suppose not, by Heine Borel $K$ is closed, so the $\sup(K)$ lies in the open set. Take an open ball around the $\sup(K)$, then $\sup(K) -\epsilon$ is the new supremum, contradiction.

However, this requires Heine Borel. And in that post, people are basically saying, that since compactness = closed and bounded, the above proof always works, so compact set always contains supremum and infimum.

Does a compact set in general topological space necessarily contain sup and inf?

  • 1
    $\begingroup$ What do you mean by "sup and inf" in a general topological space? $\endgroup$ – Eric Wofsey Jul 30 '16 at 23:50
  • $\begingroup$ Might want to restrict your question to the order topology. $\endgroup$ – Hayden Jul 30 '16 at 23:51
  • 1
    $\begingroup$ @Hayden Most order topologies (other than those similar to $\mathbb R$) don't have many interesting compact sets though. The whole concept seems pretty useless outside of $\mathbb R$. $\endgroup$ – mathguy Jul 30 '16 at 23:58
  • $\begingroup$ @mathguy Oh, I don't know. The order topologies of well-ordered sets (ordinals) have compact sets which I think are kind of interesting. $\endgroup$ – bof Jul 31 '16 at 0:09
  • $\begingroup$ Let $X$ be3 a topological space, and let $\lt$ be a linear ordering of $X.$ If $\{x\in X:x\lt a\}$ is open for every $a\in X,$ then every compact nonempty subset of $X$ has a greatest element. Likewise, if $\{x\in X:x\gt a\}$ is open for every $a\in X,$ then every compact nonempty subset of $X$ has a least element. $\endgroup$ – bof Jul 31 '16 at 0:13

In a general topological space there is no order relation - and therefore no meaningful definition of sup and inf.

  • 2
    $\begingroup$ Really better as a comment. $\endgroup$ – Thomas Andrews Jul 31 '16 at 0:05
  • 3
    $\begingroup$ I initially posted it as a comment, but then I thought this is in fact "the" answer - pointing out that the question really doesn't make sense. What other answer is possible; that is, what WOULD constitute an answer rather than a comment? $\endgroup$ – mathguy Jul 31 '16 at 0:21

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.