We have the classic result that an open subset of $\mathbb{R}$ is an at-most-countable union of disjoint open intervals.

In trying to prove this, I found the following line of reasoning, and I would like to know if its correct, because it's quite different from the other proofs I have read.

Let S be an open subset of R.

Thus for every $a$ in $S$, $a$ is contained in some open ball $B$ which is contained in $S$.

In particular $a$ is contained in an open interval which is contained in $S$.

For each $a$, let's associate the maximal such interval $I_a=(\alpha, \beta)$ where $\alpha$ and $\beta$ live in $\mathbb{R} \bigcup \{\pm\infty\}$

$\alpha=\inf\{y: (y,a] \subset S\}$

$\beta=\sup\{y: [a,y) \subset S\}$

Now we have that $S$ is the union of $I_a$ over all $a$ in S. Furthermore if $I_{a_1}$ and $I_{a_2}$ are distinct, then they must be disjoint (otherwise we would contradict the maximality of at least one of them).

Hence $S$ is a disjoint union of open intervals.

Finally, each such interval contains a rational number which is not contained in any other such interval; if there were uncountably many intervals in the union, there would be uncountably many rationals, a contradiction.

Thus $S$ is an at-most-countable union of disjoint open intervals.


Your proof works. Perhaps the only thing which needs a bit more elaboration is why such "maximal" intervals exist. The only proper way to think about maximality is with respect to set inclusion, in other words for $a \in S$, $I_a$ is the unique open interval which contains every open interval contained in $S$ and containing $a$. So we define $I_a$ as follows: $$I_a = \bigcup_{I \text{ is an open sub-interval of $S$ and $a \in I$}} I$$

We know that the union of arbitrarily many connected sets (i.e. intervals) which have at least one point in common is also connected, so $I_a$ is an interval. It is open as the union of open sets.

  • $\begingroup$ Does defining $I_a$ by its infimum and supremum as I did above work? $\endgroup$ – Joshua Benabou Jan 22 '17 at 16:14
  • $\begingroup$ @JoshuaBenabou yes. $\endgroup$ – user384138 Jan 22 '17 at 16:20

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.