Here is Prob. 5, Sec. 26, in the book Topology by James R. Munkres, 2nd edition:

Let $A$ and $B$ be disjoint compact subspaces of the Hausdorff space $X$. Show that there exist disjoint open sets $U$ and $V$ containing $A$ and $B$, respectively.

First of all, here is Lemma 26.1:

Let $Y$ be a subspace of (a topological space) X. Then $Y$ is compact (relative to the subspace topology that $Y$ inherits from $X$) if and only if every covering of $Y$ by sets open in $X$ contains a finite subcollection covering $Y$.

And, here is Lemma 26.4:

If $Y$ is a compact subspace of the Hausdorff space $X$ and $x_0$ is not in $Y$, then there exist disjoint open sets $U$ and $V$ of $X$ containing $x_0$ and $Y$, respectively.

I think I am clear on the proof of either of Lemmas 26.1 and 26.4. So I'll be using these in my proof of Prob. 5, Sec. 26, which is as follows:

Since $A$ and $B$ are disjoint, therefore for each point $a \in A$, there exist disjoint open sets $U_a$ and $V_a$ of $X$ containing $a$ and $B$, respectively, by Lemma 26.4.

As the collection $$ \left\{ \ U_a \colon \ a \in A \ \right\}$$ is a covering of $A$ by sets open in $X$, so by Lemma 26.1 there is a finite subcollection of this collection that also covers $A$; let $U_{a_1}, \ldots, U_{a_n}$ be this finite subcollection.

Now let us put $$ U \colon= \bigcup_{i=1}^n U_{a_i} \qquad \mbox{ and } \qquad V \colon= \bigcap_{i=1}^n V_{a_i}. \tag{1} $$ Here $V_{a_1}, \ldots, V_{a_n}$ are the open sets that correspond to the sets $U_{a_1}, \ldots, U_{a_n}$, respectively, as in the first paragraph of this proof.

Then both the sets $U$ and $V$ as defined in (1) here are open sets of $X$; moreover the set $U$ contains $A$ by our choice of the sets that $U$ is composed of.

As each set $V_a$ in the first paragraph contains $B$, so does each of the sets $V_{a_i}$ in (1) above; therefore $B$ is contained in $V$.

Finally, if $u \in U$, then $u \in U_{a_k}$ for some $k = 1, \ldots, n$, and thus this point $u$ would not be in the corresponding set $V_{a_k}$, and hence $u$ would not be in the set $V$ in (1) above.

Conversely, if $v \in V$, then $v$ is in each of the sets $V_{a_1}, \ldots, V_{a_n}$, and therefore $v$ is in none of the sets $U_{a_1}, \ldots, U_{a_n}$, which implies that $v$ is not in $U$.

Thus the sets $U$ and $V$ are disjoint.

Is this proof correct? If so, then is it clear in each and every one of its steps? If not, then where are the issues?

  • 2
    $\begingroup$ I cannot find any mistakes in it. Redundancy: if you have proved that an element of $U$ cannot be an element of $V$ then you are ready. It is not necessary anymore to prove also that an element of $V$ cannot be be an element of $U$. $\endgroup$ – drhab Sep 10 '18 at 13:46

Your proof is correct.

Note that the proof idea is exactly the same as for lemma 26.4 where we can replace one point in a Hausdorff space by a compact set and we separate by a union-and-intersection-trick because we get a finite union (from the finite subcover) and a finite corresponding intersection (so that makes this intersection open, as we need; an infinite intersection of open sets can have empty interior).

The proof of the empty intersection could be shorter: suppose we'd have $x \in U = \cup _{i=1}^n U_{a_i}$ and also $x \in V= \cap_{i=1}^n V_{a_i}$.

Then $x \in U_{a_j}$ for some fixed $j \in \{1,\ldots,n\}$ (definition of union). But then by being in the intersection $V$ we also know that $x \in V_{a_j}$ for that same $j$, and this immediately contradicts $U_{a_j} \cap V_{a_j} =\emptyset$ which is how we chose the corresponding pairs of $U_a$'s and $V_a$'s in the first place.
So immediately $U \cap V = \emptyset$ is clear.

You should also note the similarity to the proof of the tube lemma and its generalisations. It's a very common technique in compactness proofs.

  • $\begingroup$ Following is the link to a post of mine on the generalisation of the tube lemma. You've contributed to that post too. math.stackexchange.com/questions/1278681/… $\endgroup$ – Saaqib Mahmood Sep 11 '18 at 8:38
  • 1
    $\begingroup$ @SaaqibMahmood that generalisation is called Wallace's theorem (in Engelking's book) and it also holds for any number of factors: any product of compacts in a product of spaces that sits in an open set has a basic open set sitting between it and that open set. So products of compacts behave like points. Just as compact sets behave like points wr.t. separation axioms: in a Hausdorff space we can separate two disjoint compact sets as well (but not two closed disjoint ones in general). $\endgroup$ – Henno Brandsma Sep 11 '18 at 8:45

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.