Small cardinals and the pseudocompactness of hyperspaces of subspaces of \(\beta\omega\) (Q725652): Difference between revisions

In what follows, $[\omega]^\omega$ denotes the family of all infinite subsets of the set $\omega$ of all natural numbers. We assume the reader is familiar with the definitions of the small cardinals $\mathfrak{p}, \mathfrak{t}$ and $\mathfrak{b}$, as defined in [\textit{A. Blass}, in: Handbook of set theory. In 3 volumes. Dordrecht: Springer. 395--489 (2010; Zbl 1198.03058)]; in fact, Malliaris and Shelah have recently proved that $\mathfrak{p} = \mathfrak{t}$ [\textit{M. Malliaris} and \textit{S. Shelah}, Proc. Natl. Acad. Sci. USA 110, No. 33, 13300--13305 (2013; Zbl 1301.03041)]. The small cardinal $\mathfrak{h}$ is the smallest cardinal $\kappa$ for which there is a collection of $\kappa$ open dense families whose intersection is empty (a family $\mathcal{A}$ of infinite subsets of $\omega$ is said to be a \textit{open dense family} if for every $X \in [\omega]^\omega$ there exists $A \in \mathcal{A}$ such that $X \supseteq A$ and, for every $A,B \in [\omega]^\omega$, if $A \in \mathcal{A}$ and $B \setminus A$ is finite then $B \in \mathcal{A}$). It is well-known that $\aleph_1 \leqslant \mathfrak{t} \leqslant \mathfrak{h} \leqslant \mathfrak{b} \leqslant \mathfrak{c}$, and that each one of the possible strict inequalities is consistent. We also assume the reader is familiar with \textit{MAD families} (i.e., maximal almost disjoint families of infinite subsets of $\omega$), as well as with Mrówka-Isbell spaces of the form $\Psi(\mathcal{A})$, where $\mathcal{A}$ is a MAD family. As usual, the family of all free ultrafilters over $\omega$ will be denoted by $\omega^*$, whereas the family of all ultrafilters over $\omega$ is denoted by $\beta \omega$; recall that $\beta \omega$ (endowed with a well-known topology) is usually identified with the Stone-Čech compactification of the discrete set $\omega$. \par Given a topological space $X$, its \textit{Vietoris hyperspace} is the set $\text{CL}(X)$ of all non-empty closed subsets of $X$ endowed with the topology generated by the sets of the form $A^+ = \{F \in \text{CL}(X): F \subseteq A\}$ and $A_{-} = \{F \in \text{CL}(X): F \cap A \neq \emptyset\}$, where $A \subseteq X$ is open. \par If $(B_n: n < \omega)$ is a sequence of subsets of a topological space $X$ and $p \in \omega^*$, a point $x \in X$ is said to be a \textit{$p$-limit} of $(B_n: n < \omega)$ if for every neighbourhood $V$ of $x$ one has $\{n < \omega: V \cap B_n \neq \emptyset\} \in p$. If $\kappa$ is a cardinal and $M \subseteq \omega^*$, we say that a topological space $X$ is \textit{ $(\kappa,M)$-pseudocompact} if for every family $\{(V^\alpha_n: n \in \omega): \alpha < \kappa\}$ of sequences of open subsets of $X$ there is $p \in M$ such that for every $\alpha < \kappa$ there is a $p$-limit of $(V^\alpha_n: n \in \omega)$ in $X$. \par In previous works due to Ginsburg, Hrušák, Hernandéz-Hernandéz and Tomita, among others, the relationship between the pseudocompactness of spaces of the form $X^\omega$ and $\text{CL}(X)$ (for $X$ being either a subspace of $\beta \omega$ or a Mrówka-Isbell space) has been investigated. There are several results which are known from those mentioned works, for instance: \begin{itemize} \item[(i)] if every power of a space $X$ is countably compact then so is $\text{CL}(X)$, and if $\text{CL}(X)$ is countably compact (pseudocompact) then so is every finite power of $X$; \item[(ii)] under $\mathfrak{h} < \mathfrak{c}$, there is a MAD family $\mathcal A$ such that $\text{CL}(\Psi(\mathcal A))$ is not pseudocompact, but under $\mathfrak{p} = \mathfrak{c}$ one has that for every MAD family $\mathcal A$ the hyperspace $\text{CL}(\Psi(\mathcal A))$ is pseudocompact; \item[(iii)] it is not true in $\mathbf{ZFC}$ that the pseudocompactness of $X^\omega$ implies the pseudocompactness of $\text{CL}(X)$; and (iv) if $X$ is a homogeneous Tychonoff space such that $\text{CL}(X)$ is pseudocompact then $X^\omega$ is pseudocompact. \end{itemize} In the paper under review, the authors investigate whether the countable compactness of $X^\omega$ implies the pseudocompactness of $\text{CL}(X)$. The main results of the paper are the following: \begin{itemize} \item[(i)] There exists $X \subseteq \beta \omega$ of cardinality $\mathfrak{c}$ such that $\omega \subseteq X$ and $X^\kappa$ is countably compact for every $\kappa < \mathfrak{t}$, but $\text{CL}(X)$ is not pseudocompact. \item[(ii)] There exists $X \subseteq \beta \omega$ of cardinality $2^\mathfrak{h}$ such that $\text{CL}(X)$ is not pseudocompact, $\omega \subseteq X$ and $X^\kappa$ is countably compact for every $\kappa < \mathfrak{h}$. \item[(iii)] If $X$ is a subspace of $\beta \omega$ which contains $\omega$ and $\text{CL}(X)$ is pseudocompact, then for every $\kappa < \mathfrak{h}$ one has that $X$ is $(\kappa,\omega^*)$-pseudocompact. It follows that, for every subspace $X$ of $\beta \omega$ containing $\omega$, if $\text{CL}(X)$ is pseudocompact then all powers of $X$ are pseudocompact. \item[(iv)] There exists a space $X$, $\omega \subseteq X \subseteq \beta \omega$, such that $\text{CL}(X)$ is pseudocompact and $X$ is not $(\mathfrak{b},\omega^*)$-pseudocompact. \item[(v)] If $\omega \subseteq X \subseteq \beta \omega$ and $\text{CL}(X)$ is pseudocompact then $\text{CL}(X)$ is $(\kappa,\omega^*)$-pseudocompact for every $\kappa < \mathfrak{h}$. \end{itemize} The paper finishes by posing two questions. Is it consistent that $\mathfrak{t} < \mathfrak{h} < \mathfrak{c}$ and there is a space $X$ such that $\omega \subseteq X \subseteq \beta \omega$, $X^\mathfrak{h}$ is countably compact and $\text{CL}(X)$ is not pseudocompact? Is there a space $X$ such that $\omega \subseteq X \subseteq \beta \omega$, $\text{CL}(X)$ is pseudocompact but $X$ is not $(\mathfrak{h},\omega^*)$-pseudocompact?