Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $E$ be a subset of a metric space $(S,d)$. Prove that:
A point is in the boundary of $E$ if and only if it belongs to the closure of both $E$ and its complement.

Here is what I thought:
I'm first trying to understand what I need to prove.
The boundary of $E$ is the set $E^- -E ^{\circ}$.
It a point belongs to both $E^-$ and $(S-E)^-$, then it belongs to $E^- \cap (S-E)^-$.
Therefore I think I need to prove that:
$$E^- -E ^{\circ}=E^- \cap (S-E)^-$$

And now I'm not sure if there any set theory rules I could use.
Intuitive I would say that I need to prove: $S-E ^\circ =(S-E)^-$. Is this correct ?
And I can kind of feel that this last statement is right, but I can't prove this rigoursly.

Could anybody help me how I can prove this statement more rigourisly ?

share|improve this question
Do you mean $E^-$ to be $\bar{E}$ = \bar{E}? or $\overline{E}$ = \overline{E} –  amWhy Apr 3 '13 at 15:49
This really comes down to element chasing. Let $x$ be in the closer of $E$ but not in the interior of $E$, then it's certainly in the closure of $E$. Can you show that it's also in the closure of the complement of $E$? (Hint: it's not in the interior of $E$ so all neighbourhoods of $x$ intersect the complement of $E$). This gives you one inclusion. Try the other inclusion using similar element chasing methods. –  Daniel Rust Apr 3 '13 at 15:51
@amWhy In my book: Elementary Analysis by Ross, they denote the closure of the set $E$ with $E^-$. What is the standard notation for this ? –  Kasper Apr 3 '13 at 15:55
Kasper, that's fine, I was just checking. Often closure of a set $E$ is denoted $\overline{E}$. –  amWhy Apr 3 '13 at 16:11

2 Answers 2

up vote 4 down vote accepted

You're absolutely right!

Note that if a point $x$ is interior to $E$, then it certainly isn't in $S-E$. In fact there is some $r>0$ such that any point of $S-E$ is separated from $x$ by a distance of at least $r$, so $x$ isn't a limit point of $S-E$, either. Hence, $E^o$ and $(S-E)^-$ are disjoint, so that $S-E^o$ is contained in $(S-E)^-$.

On the other hand, suppose $x\in(S-E)^-$. If $x\in S-E$, then $x\notin E$, so $x\notin E^o\subset E,$ and so $x\in S-E^o$. Otherwise, $x$ is a limit point of $S-E$, so for all $r>0$ there is a point $y_r\in S-E$ such that $0<d(x,y_r)<r$. Hence, no open ball around $x$ is disjoint from $S-E$, so no open ball around $x$ is contained in $E$, so $x\notin E^o$, and so $x\in S-E^o$. Thus, we have the other inclusion.

share|improve this answer
thanks ! ${}{}{}{}{}{}$ –  Kasper Apr 3 '13 at 16:08

Yes. $S-E^\circ=\overline{S-E}$ is correct.

$x\notin E^\circ \iff $ there is no whole ball around $x$ within $E$ $\iff$ all balls around $x$ intersect $S-E$ $\ \iff x\in\overline{S-E}$.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.