I am struggling with this old problem, which is also posted here:

Let $X$ satisfy countable chain condition(abbreviated as CCC) and $X$ has a regular $G_\delta$-diagonal. Then the cardinality of $X$ is at most $\mathfrak c$.

$X$ has a regular $G_\delta$-diagonal iff there is a collection of open sets of $X^2$, say $\lbrace U_n: n\in N\rbrace$, such that $\Delta=\bigcap\lbrace \overline{U_n}: n \in N\rbrace$, where $\Delta=\lbrace(x,x): x \in X\rbrace$.

Note that the question is answered; however I hope to get new proof.

By certain effort, If $|X|>\mathfrak c$, we can get an uncountable closed discrete subset $S$ of $X$, and for any point $x \in X$, there exists an open set $U_x$ such that $\overline{U_x} \cap S$ has at most one point.

I would like to know whether the following conjecture is right, wrong, or neither:

Let $X$ be a Hausdorff space. If $S \subset X$ is an uncountable closed discrete subset of $X$, and for any point $x \in X$, there exists an open set $U_x$ such that $\overline{U_x} \cap S$ has at most one point. Then could we obtain an uncountale collection of disjoint open sets in $X$?

Thanks for your any help.


Assuming you want $U_x$ to be a neighbourhood of $x$, your condition on $S$ is equivalent to $S$ being closed discrete, whenever $X$ is a regular space. That's because $X$ is regular if and only if for every point $x \in X$ and every open neighbourhood $U$ of $x$ there is an open set $V$ such that $x \in V \subset \overline{V} \subset U$. Now, there are plenty of regular ccc spaces with uncountable closed discrete subspaces. For example, let $\kappa$ be any uncountable cardinal and $X=\{x \in 2^\kappa: (\exists \beta < \kappa)(x(\beta)=1 \wedge (\forall \gamma>\beta)(x(\gamma)=0)) \}$ with the topology induced from the usual topology on $2^\kappa$. $X$ has the ccc as a dense subspace of the ccc space $2^\kappa$. For $\alpha < \kappa$ define a point $x_\alpha \in X$ as follows: $x_\alpha(\beta)=1$ for $\beta \leq \alpha$ and $x_\alpha(\beta)=0$, for $\beta > \alpha$. Then $D=\{x_\alpha: \alpha < \kappa \}$ is a closed discrete subspace of $X$ of cardinality $\kappa$. To see that, note that $\{[\sigma] \cap X: \sigma \in Fn(\kappa,2)\}$ is a base for $X$, where $Fn(\kappa,2)$ denotes the set of all finite partial functions from $\kappa$ to $2$ and $[\sigma]:=\{x \in 2^\kappa: \sigma \subset x\}$. We have $[\{(\alpha,1), (\alpha+1,0)\}] \cap D=\{x_\alpha\}$ for every $\alpha < \kappa$ and hence $D$ is discrete. Moreover, $D$ is closed because if $x \in X \setminus D$, then there are ordinals $\alpha, \beta < \kappa$, $\alpha < \beta$ such that $x(\alpha)=0$ and $x(\beta)=1$ and hence $[\{(\alpha,0), (\beta,1)\}]\cap X$ is an open neighbourhood of $x$ which misses $D$.

  • 1
    $\begingroup$ I don't think the set $D$ is closed. If $\gamma$ is a limit ordinal, $y_{\gamma}(\beta)=1$ for $\beta<\gamma$, and $y_{\gamma}(\beta)=0$ for $\beta\geq\gamma$, then $(x_{\alpha})_{\alpha<\gamma}\rightarrow y_{\gamma}$. $\endgroup$ – Joseph Van Name May 6 '13 at 0:33
  • $\begingroup$ You're right, Joseph. Thanks for pointing that out! I think that to fix that it suffices to redefine $X$ to be the set of all $x \in 2^\kappa$ such that there is $\beta < \kappa$ with $x(\beta)=1$ and $x(\gamma)=0$ for every $\gamma > \beta$. I edited my answer accordingly. $\endgroup$ – Santi Spadaro May 6 '13 at 7:25

I claim that every discrete space of non-measurable cardinality can be embedded as a closed subspace of some completely regular space satisfying the countable chain condition. Let $S$ be an uncountable discrete space whose cardinality is below the first measurable cardinal. Since discrete spaces of non-measurable cardinality are realcompact, the space $S$ is embeddable as a closed subspace of some product $\mathbb{R}^{I}$ for some $I$. However, the space $\mathbb{R}^{I}$ satisfies the countable chain condition. We can give $\mathbb{R}$ a Borel probability measure. If we take the infinite product measure, then we get a measure $\mu$ on $\mathbb{R}^{I}$ such that $\mu(U)>0$ for each non-empty open set $U$. Therefore there can only be countably many disjoint open sets in $\mathbb{R}^{I}$ since $\mu(\mathbb{R}^{I})=1$.

  • $\begingroup$ what is meaning of non-measurable cardinality? $\endgroup$ – Paul May 4 '13 at 11:27
  • $\begingroup$ A non-measurable cardinal is a cardinal $\kappa$ such that there is no non-principal $\sigma$-complete ultrafilter on $\kappa$. Said differently, a non-measurable cardinal is a cardinal below the first measurable cardinal(if it exists). Being something called a large cardinal axiom, the existence of a measurable cardinal is independent from the standard axioms of set theory, and the first measurable cardinal is extremely large if it exists, so the non-measurability of $S$ is not too great of a restriction. $\endgroup$ – Joseph Van Name May 4 '13 at 12:33

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.