Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Question: Let $U$ denote a subset of $\mathbb{R}$. We say that $U$ is an open set when for every $x \in U$ there is an $\epsilon > 0$ for which $U_{\epsilon}(x) = (x - \epsilon, x +\epsilon)$ is contained in $U$. Show:

(i) For $a,b \in \mathbb{R}$ the interval $(a,b)$ is open, but the interval $[a,b]$ is not.

(ii) If $U,W \subset \mathbb{R}$ are open, so are $U \cup W$ and $U \cap W$.

(iii) If $(U_{i})_{i \in I}$ is a set of open subsets from $\mathbb{R}$, $\cup_{i \in I}U_{i}$ is also open.

(iv) If $(U_{i})_{i \in I}$ is a set of open subsets from $\mathbb{R}$, $\cap_{i \in I}U_{i}$ is not necessarily open.

My attempt: I think that with (i) i can conceptually understand everything:

$\forall x \in (a,b): a< x - \frac{|a-x|}{2} < x < x + \frac{|b-x|}{2} < b$

If $x> \frac{|b-a|}{2}$ let $\epsilon = \frac{|b-x|}{2}$

If $x< \frac{|b-a|}{2}$ let $\epsilon = \frac{|a-x|}{2}$

$\Rightarrow \forall x \in (a,b) \exists \epsilon > 0 : U_{\epsilon}(x) = (x - \epsilon, x +\epsilon) \Rightarrow$ the interval (a,b) is open.


max$[a,b] = b \Rightarrow \forall x \in [a,b]: b \geq x$

If $\epsilon > 0, b \ngeq b+ \epsilon \Rightarrow$ there exists an $x$ for which $U_{\epsilon}(x) = (x - \epsilon, x +\epsilon)$ is not contained in $U \Rightarrow$ the interval is not open

With (ii) and (iii) i think i understand, but i am not sure how to start stating that they are true mathematically. to be sure, $\cup_{i \in I}U_{i}$ denotes several unions in a row, correct? i'm even reading a proof which goes:

Given $v \in U \cap W$, there exists an open ball $B$ of radius $r$ centered at $v$ contained in $U$, and there exists an open ball $B'$ of radius $r'$ centered at $v$ contained in $W$. Let $\delta =$ min$(r, r')$. Then the open ball of radius $\delta$ centered at $v$ is contained in $U \cap W$, which is therefore open.

Unfortunately i'm having trouble interpreting this and am wondering if there is another way to put it or approach it.

With (iv) i am not even sure intuitively why it should be true. i can't think of a situation where it wouldn't be true, i guess it has to do with whether the sets are finite or not, but i don't see how this would make a difference (especially given that openness is apparently preserved with infinite unions).

If anyone can help provide suggestions, corrections, hints, etc i appreciate it, thanks!

share|cite|improve this question
up vote 4 down vote accepted

Intuitively, a set $U$ is open if every point in the set can be completely surrounded by points that are all in the set; here, "surrounded" means that an entire little interval centered at $x$ is completely contained in $U$, which means, in a sense, that $x$ is really inside $U$. A set is open if every point in the set is really "inside" the set. Compare with $[a,b]$ (which you show is not open), where one might say that although $b$ is in the set, it is "barely" in the set: move just a bit to the right, and you fall outside $[a,b]$.

That said, on to your work:

Your idea for the first part of (i) is okay, but the way you describe $\varepsilon$ is not. For example, say $x=1000$, $a=999.9$, and $b=1000.5$. Then $|b-a|/2 = .3$, and $x\gt |b-a|/2$, so you take $\varepsilon = 0.3$. But $(x-0.3,x+0.3)$ is not contained in $(a,b)$, because $x-0.3 = 999.7\lt a$.

What you really want to say is that $\varepsilon=\min(|x-b|/2, |x-a|/2)$.

For the second part, you've got it right, but there is no need to say "there exists an $x$ ...", just say that for every $\varepsilon\gt 0$, $(b-\varepsilon,b+\varepsilon)$ is not contained in $[a,b]$.

For (ii); not quite: $\cup_{i\in I} U_i$ is not "several unions in a row", because the set may not something that you can do "one after another"; maybe there is one for every real number! No, $\cup_{i\in I}U_i$ just means "the union of all the $U_i$"; this is the collection of all points that are in at least one of the $U_i$.

I think you mean $U\cap W$ when you said $V\cap V$. But you've go the idea right for the intersection. Conceptually: you know that you can find an open interval around $x$ that "works" for $U$, and one that works for $W$; to find one that works for both, just take the smaller one.

As for (iv), it's not true: you are being asked to show it is not true. Can you think of a way to get a closed interval as an intersection of open ones?

Added: Okay, here's a hint for (iv): for every positive integer $n$, $\frac{1}{n}$ is stricly larger than $0$; but the infinimum of $\{ \frac{1}{n}\mid n\in\mathbb{Z}_{\gt 0}\}$ is not strictly larger than $0$.

share|cite|improve this answer
thanks for the pointers, i edited the part where i had incorrectly written $V \cap V$. ok that makes a lot more sense when it comes to (i)! as for (iv) unfortunately i still can't think of a way to get a closed interval as an intersection of open ones.... – ghshtalt Nov 25 '10 at 7:48
@user3711: Think of the argument you give for finite intersections of open sets (you take the smallest $\epsilon$ among the available $\epsilon$'s). When and why will this argument fail for infinite intersections? Now use this fact to construct a counter example. – user17762 Nov 25 '10 at 8:04

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.