Show that $(S, d)$ has Baire's property iff every set of first category has a dense complement.

A set is of first category if it is a countable union of nowhere dense sets. First Category

Baire's Lemma: Let $(X, \rho)$ be a complete metric space and $\{U_n\}_{n=1}^{\infty}$ a sequence of open dense sets in $X$. Then the set $\bigcap_{n =1}^{\infty} U_n$ is also dense.

Note that A is nowhere dense iff $(\overline{A})^c$ is open and dense. Suppose $\forall A$, $A$ is first category. That is, $A$ is a countable union of nowhere dense sets. i.e., $A_i$ is nowhere dense. $\forall A, \; A = \bigcup_{n \in \mathbb{N}}A_i \implies (A)^c = \bigcap_{n \in \mathbb{N}}(A_{i})^{c}$. But we don't know if this is a countable dense set! We know that $(\overline{A}_i)^c$ would be, but not just $(A)^c$.

If $A$ is nowhere dense, what can we say about the compliment of $A$?

Can anyone clear this up or provide advice?

More scratch work:

In my proposed proof, I said let all $A \subset S$ be of first category. i.e., each $A_i$ is nowhere dense, i.e, $(\overline{A}_i)^c$ is open and dense. So, $\forall A, (\overline{A}_i)^c = \bigcup (\overline{A}_i)^c \implies ((\overline{A}_i)^c)^c = \bigcap (((\overline{A}_i)^c)^c) = \bigcap (\overline{A}_i)$. Still though, what can I say about $\overline{A}_i$?


$\newcommand{\cl}{\operatorname{cl}}$Suppose first that $\langle S,d\rangle$ has the property that the intersection of countably many dense open sets is dense in $S$, and let $A$ be a first category set in $S$; we want to show that $S\setminus A$ is dense in $S$. Since $A$ is first category, there are nowhere dense sets $A_k$ for $k\in\Bbb N$ such that $A=\bigcup_{k\in\Bbb N}A_k$. For $k\in\Bbb N$ let $U_k=S\setminus\cl A_k$; each $U_k$ is a dense open subset of $S$, so $\bigcap_{k\in\Bbb N}U_k$ is dense in $S$. But

$$\bigcap_{k\in\Bbb N}U_k=\bigcap_{k\in\Bbb N}(S\setminus\cl A_k)=S\setminus\bigcup_{k\in\Bbb N}\cl A_k\subseteq S\setminus\bigcup_{k\in\Bbb N}A_k=S\setminus A\;,$$

so $S\setminus A$ is also dense in $S$.

Now suppose that the complement of every first category subset of $S$ is dense in $S$, and let $\{U_k:k\in\Bbb N\}$ be a family of dense open subsets of $S$. For each $k\in\Bbb N$ let $A_k=S\setminus U_k$; $A_k$ is closed and nowhere dense in $S$, so $A=\bigcup_{k\in\Bbb N}A_k$ is first category in $S$. Finally,

$$\bigcap_{k\in\Bbb N}U_k=\bigcap_{k\in\Bbb N}(S\setminus A_k)=S\setminus\bigcup_{k\in\Bbb N}A_k=S\setminus A\;,$$

which is dense in $S$, as desired.

A good book with much information on such topics is John C. Oxtoby, Measure and Category: A Survey of the Analogies between Topological and Measure Spaces, 2nd edition. (The first edition is also good.)

  • $\begingroup$ More detail than my answer so i'll put it as the 'Accepted' one. Question though, how did you find this post? It is a month old. $\endgroup$ – Ozera Jan 20 '15 at 22:08
  • $\begingroup$ @Ozera: I’m a set-theoretic topologist, so from time to time I check the general-topology tag to see whether I’ve missed anything interesting. Apparently I missed your question when it was posted. $\endgroup$ – Brian M. Scott Jan 20 '15 at 22:19
  • $\begingroup$ Ah, I see. A non-relevant question, from your profile information, you are a Professor at Cleveland State University. I am currently applying to summer REU programs. Is CSU holding a (math) REU program in the summer? I could not find any information about it. $\endgroup$ – Ozera Jan 20 '15 at 22:30
  • $\begingroup$ @Ozera: Retired now, so not always up to date on what’s going on, but in this case I do know: it’s not. (I don’t think that it ever has done.) I did find a list of all the REU math programs here; so far as I know, it’s current. $\endgroup$ – Brian M. Scott Jan 20 '15 at 22:33
  • $\begingroup$ That is the same list I am using at the moment. There is also a list provided by AMS which is nice. Thanks for information. I suspect that i'll see more of your answers to my analysis/topology questions in the future. $\endgroup$ – Ozera Jan 20 '15 at 22:41

A nowhere dense set has empty interior. Therefore, the complement of a nowhere dense set is dense. Hence, I think my original solution is fine for one direction. If anyone else has comments to add, please do so.


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.