Let $X$ be an uncountable set and let $\tau = \lbrace \phi \rbrace \cup \lbrace U \subset X : U^c$ is countable $\rbrace$.

1) i) Prove the countable intersection of open sets in $\tau$ remains open.

ii) Is this result true in usual $\mathbb{R}$?

  • i) Let $A_i \in \tau$, how can I write the countable intersection ? is it like that $(\bigcap A_i)^c$ ?

If it is correct, then

$(\bigcap A_i)^c =\bigcup A_i^c$

The countable union of countable set is a countable set, and then it closed, thus it's complement which is $\bigcap A_i$ open set.

  • ii) NO, $\cap (\frac{-1}{n},\frac{1}{n})=\lbrace 0 \rbrace$.

2) Show that the finite intersection of open sets is non empty. Is this result true in usual $\mathbb{R}$?

Let $A, B\in \tau$, and suppose that $A \cap B=\phi$, then $A \subset B^c$, since $B \in \tau$, then $B^c$ is countable, and hence $A$ also is countable, But it is not necessarily to be that, so our supposition is not correct.

3) Is $\overline{\mathbb{Q}} = \mathbb{R}$? for the topology $\tau$. what about $\mathbb{Q}^c$ and $[0,1]$.

No, $\mathbb{Q}$ is closed, since it is countable, then $\overline{\mathbb{Q}}=\mathbb{Q}$.

but what about for $\overline{\mathbb{Q}^c}$?, the closure is the smallest closed set which is contains $\mathbb{Q}^c$, in this topology, it will be the smallest countable set which is contains $\mathbb{Q}^c$, and this is impossible, since $\mathbb{Q}^c$ is not countable, the same thing with $[0,1]$.

Is that true, please?

"sorry I don't speak English well".


The essence of the argument you gave is OK.

Suppose that $U_n$, $n \in \mathbb{N}$ are open sets in the co-countable topology. If one of them is $\emptyset$, then the intersection is $\emptyset$ too, hence open, so we can assume that the $U_n$ all have countable complement.

Then $(\bigcap_{n \in \mathbb{N}} U_n)^\complement = \bigcup_{n \in \mathbb{N}} U_n^\complement$ by de Morgan, so the complement of $\bigcap_{n \in \mathbb{N}} U_n$ is a countable union of countable sets, hence countable. So $\bigcap_{n \in \mathbb{N}} U_n$ is open.

Your example for the usual topology is fine. Maybe prove that $\{0\}$ is not open, as you claim?

b) If $U$ and $V$ are both non-empty open (so countable complements) and suppose $U \cap V = \emptyset$, then indeed $U^\complement \cup V^\complement = \mathbb{R}$ and the reals would be countable, contradiction.

In the usual topology $(0,1)$ and $(2,3)$ are disjoint non-empty open sets.

c) Indeed $\mathbb{Q}$ is closed, being countable, and so equals its own closure in this topology.

$[0,1]$ and $\mathbb{Q}^\complement$ are dense: the only closed sets are the countable sets and $\mathbb{R}$ so any uncountable set has only one containing closed set.


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.