Ok so I was goin over the structure theorem proof for open sets in my book, and there seems to be one point where I am completely lost.

Theorem: Structure theorem for open set For a nonempty set $S$ in $\mathbb{R}$, the following are equivalent

$1)$ $S$ is an open set

$2)$ every element of $S$ is an interior point of $S$ i.e. $S^o = S$.

$3)$ S is a countable union of pairwise disjoint open intervals.

Proof of the theorem: First we prove that $1$ $\implies$ $2$, that is S open $\implies$ $S = \bigcup_{\alpha} (a_\alpha, b_\alpha)$. $\forall x \in S$, $x \in (a_\alpha, b_\alpha)$ for at least one $\alpha$. Let $r=\min(x-a_\alpha, b_\alpha-x)$, then $r>0$, and $(x-r,x+r) \subseteq (a_\alpha, b_\alpha) \subseteq S \implies x \in S^o$, thus $S=S^o$. Next we show that $2 \implies 3$. Suppose $S^o=S$, if $x \in S$, then $(x-r,x+r) \subseteq S$ for some $r.0$. Let $I_x$ be the largest interval containing $x$ such that such that $I_x \subseteq S$. Here $I_x$ is an interval from $m_x$ to $M_x$ where $m_x= \inf \{m:(m,x] \subseteq S\}$ and $M_x= \sup \{M:[x,M) \subseteq S\}$.

Now if $a = M_x$ or $m_x$ is a real number then $a\notin S$ (otherwise $(a-r',a+r')\subseteq S$ for some $r'>0$, then $I_x \bigcup (a-r',a+r')$ is a larger interval in $S$ containing $x$ which is a contradiction). For $x,y\in S$, if $I_x \bigcap I_y \ne \varnothing$, then $I_y=I_x$. For if it was not then $I_x \bigcup I_y$ is a larger interval in $S$ containing $x$ and $y$). So $S$ is the union of these pairwise disjoint open intervals.

In each of these open pairwise disjoint intervals, we can choose a rational number by the density of $\mathbb{Q}$. For different pairwise disjoint open intervals the rational numbers are different. Since $\mathbb{Q}$ is countable there are countably many open pairwise disjoint intervals.


$1)$ Why is that $M_x$ and $m_x$ cannot be a real number? Because any number is considered real, so does that mean $M_x$ and $m_x$ cannot take any value?, this really confuses me.

$2)$ Why do we choose a rational number $\mathbb{Q}$ for the pairwise disjoint open intervals?

Basically I dont get it at all after the part of the proof where it starts: Now if $a = M_x$ or $m_x$ is a real number then $a\notin S$ ....

Can anyone please help me clarify this confusion for me? I really need to understand this part so I can go further on with other theorems.


1) Here $M_x , m_x$ cannot be real number means they can be $\infty$ or -$\infty$. For example, if $S=\mathbb R$ then $M_x=\infty$ and $m_x=-\infty$ for any $x\in S$.

2) Instead of $\mathbb Q$ you can take any countable dense subset of $\mathbb R$. But $\mathbb Q$ is one of the easiest example of such subset, so we generally consider $\mathbb Q$. As $\mathbb Q$ is dense in $\mathbb R$, every open interval contains at least one rational number. Now as $\mathbb Q$ is countable, open subintervals of $S$ are countable because we have surjective map from $\mathbb Q$ to open subintervals of $S$.

| cite | improve this answer | |
  • $\begingroup$ Thank you for the reply, but I wonder why cant they be a real number?, so it means it can only be $\infty$? What does taking $ \mathbb{Q}$ actually really mean? Could you like give a numerical example? $\endgroup$ – Aurora Borealis Nov 24 '17 at 7:01
  • $\begingroup$ @AuroraBorealis I have edited my answer. Take a look at it. $\endgroup$ – Mayuresh L Nov 24 '17 at 11:01

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.