Anti-Urysohn spaces (Q330042)

In this very well-written paper the authors introduce and study the following two classes of infinite Hausdorff topological spaces: 1) A space \(X\) is called \textit{anti-Urysohn} (or \(AU\) space, in short) if and only if any two non-empty regular closed subsets of \(X\) intersect. 2) A space \(X\) with at least two non-isolated points is called \textit{strongly anti-Urysohn} (or \(SAU\) space, in short) if and only if any two infinite closed subsets of \(X\) intersect. In Section 2 of the paper under review, as a consequence of a general lemma, for every infinite cardinal \(\kappa\), the authors construct an \(AU\) space \(X\), with \(|X|=\kappa\) and \(d(X)=\log(\kappa)\), such that the intersection of any family of fewer than \(\mathrm{cf}(\kappa)\) many non-empty regular closed sets in \(X\) is non-empty. Since every \(AU\) space is connected, the case \(\kappa=\omega\) gives one more example of a countable connected Hausdorff space. The space in the authors' example has the stronger property (compared with the previously known examples) that the intersection of any finitely many non-empty regular closed subsets of that space is non-empty. In addition, after showing that if \(\kappa\geq\omega\), then \(|X|\leq 2^{2^\kappa}\) whenever \(X\) is a locally \(\kappa\), connected space, the authors prove a general theorem from which it follows that for every infinite cardinal \(\lambda\), with \(\kappa<\lambda\leq 2^{2^\kappa}\), there is a locally \(\kappa\), \(AU\) space \(X\) such that \(|X|=\lambda\). Therefore for every infinite \(\lambda\leq 2^{\mathfrak c}\) there exists a locally countable \(AU\) space of cardinality \(\lambda\). (We recall that a space \(X\) is locally \(\kappa\) if every point of \(X\) has a neighborhood of cardinality \(\leq\kappa\).) In Section 3 the authors prove (in \(ZFC\)) that if \(X\) is an \(SAU\) space then: (1) \(X\) is countably compact; (2) every compact subset of \(X\) is finite; (3) any infinite closed subset of \(X\) is uncountable; (4) any countable intersection of infinite closed subsets of \(X\) is infinite; (5) \(X\) contains an \(SAU\) subspace of cardinality \(\leq\mathfrak c\); and (6) \(\mathfrak s\leq |X|\leq 2^{2^{\mathfrak c}}\), where \(\mathfrak s\) is the splitting number. In contrast to the case of \(AU\) spaces, the authors show only consistently that \(SAU\) spaces exist. In particular, they prove the following: (1) if \(\mathfrak r=\mathfrak c\), where \(\mathfrak r\) is the reaping number, then there exists a separable, crowded, locally countable \(SAU\) space of cardinality \(\mathfrak c\); (2) after adding \(\lambda> \omega\) Cohen reals to any ground model, then in the extension there exist \(SAU\) spaces of cardinality \(\kappa\) for every \(\omega_1\leq\kappa\leq\lambda\); (3) if \(GCH\) holds and \(\kappa\leq \lambda\) are uncountable regular cardinals, then there is a \(CCC\) generic extension in which \(\mathfrak s=\kappa\), \(\mathfrak c=\lambda\), and there exists an \(SAU\) space of cardinality \(\mu\), for every \(\mathfrak s \leq\mu\leq \mathfrak c\). At the end of the paper the authors ask the following questions: (1) Is there an \(SAU\) space in \(ZFC\)? (2) Is it consistent that there is an \(SAU\) space of cardinality greater than \(\mathfrak c\)? Is it consistent that there exists a locally countable \(SAU\) space of cardinality greater than \(\mathfrak c\)? (3) Does the existence of an \(SAU\) space imply the existence of a crowded \(SAU\) space?