Formally stated, Let $S_1, S_2, \cdots$ be closed sets and assume that $\bigcup_{j} S_j = \mathbb{R}$. Prove that at least one of the sets $S_j$ has nonempty interior.

After doing some searching before I asked this question, I found it to be a special case of Baire Category Theorem. (ZF) If $\mathbb{R}^k$ is a countable union of closed sets, then at least one has a nonempty interior.

However this is my first real analysis class, so we haven't learned anything about category sets/nowhere dense sets yet. This is left as a starred textbook problem of the section perfect set, so I am wondering if it can be solved with more fundamental techniques.

The hint from the textbook is to use an idea from the proof that perfect sets are uncountable. (The textbook I am using is Real Analysis and Foundation, Third Edition, by Steven Krantz)

  • $\begingroup$ I guess your title should say "nonempty" rather than "empty"? $\endgroup$ – MPW Oct 31 '15 at 15:47
  • 1
    $\begingroup$ This isn't "measure theory." This is about nowhere dense sets, 1st category sets, 2nd category sets, residual sets, and of course the Baire category theorem. Find out about these. $\endgroup$ – B. S. Thomson Oct 31 '15 at 15:48
  • $\begingroup$ In the "related" section, this turns up: math.stackexchange.com/questions/88632/… — I guess that should be helpful. $\endgroup$ – celtschk Oct 31 '15 at 15:55

Let $F_n$ be closed sets with empty interior ($n=1,2,3,\dots$). Construct recursively closed intervals $J_n$ of positive length ${} \le 1/n$ with $J_1 \supset J_2 \supset J_3 \supset \dots$ and $J_n \cap F_n = \varnothing$. Then $\bigcap J_n$ is nonempty, but disjoint from $\bigcup F_n$.

  • $\begingroup$ This is nice, but you should use the word compactness when you say that $\cap J_n$ is non-empty. $\endgroup$ – Zach Stone Oct 31 '15 at 17:07
  • $\begingroup$ @ZachStone: Since it is a "hint" I was hoping taiweic would come up with things like that. $\endgroup$ – GEdgar Oct 31 '15 at 17:09
  • $\begingroup$ I am sorry I am afraid I didn't get it after staring at it for a long time. Can you possibly give out more hint/explanation as for how to construct recursively intervals $J_n$ of positive length $\leq \frac{1}{n}$ with $J_1 \subset J_2 \subset J_3 \subset \cdots$ and $J_n \cap F_n = \emptyset$? $\endgroup$ – LKSR Oct 31 '15 at 22:35

If none would have empty interior then they would have been nowhere dense thus $\mathbb R$ becomes a set in first category which is false since $\mathbb R$ is complete.


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.