Supremum vs. maximum: \(\lambda\)-sets (Q886299)

A~set~\(X\) of reals is a~\(\lambda\)-set if every countable subset of~\(X\) is relatively~\(G_\delta\) in~\(X\). Let \(\mathcal A\) be an almost disjoint family of infinite subsets of~\(\omega\). The \(\Psi\)-space \(\Psi(\mathcal A)\) associated to~\(\mathcal A\) is \(\omega\cup\mathcal A\) where the points of~\(\omega\) are isolated and the basic neighbourhoods of \(A\in\mathcal A\) are of the form \(\{A\}\cup A\setminus F\) where \(F\subseteq\omega\) is finite. The space \(\Psi(A)\) is a first countable, separable, locally compact Moore space. If \(\Psi(\mathcal A)\) is a~pseudonormal space, then \(\mathcal A\)~is a~\(\lambda\)-set as a~subset of~\(\mathcal P(\omega)\), and if there is a~\(\lambda\)-set of cardinality~\(\kappa\), then there is an almost disjoint family~\(\mathcal A\) of cardinality~\(\kappa\) such that \(\Psi(\mathcal A)\) is a~pseudonormal space. The paper under review deals with a~problem of \textit{P.~Nyikos} [``On first countable, countably compact spaces. III. The problem of obtaining separable noncompact examples'', in: J. van Mill et al. (eds.), Open problems in topology. North-Holland, Amsterdam, 127--161 (1990; Zbl 0718.54001)], who asked: ``Is there a~pseudonormal \(\Psi\)-space of cardinality~\(\mathfrak d\)? More generally, what is the maximum cardinality of a~pseudonormal \(\Psi\)-space?'' The author shows that in the Cohen model the answer for the first part of the question is ``no''. Next he proves that in the absence of inaccessible cardinals there is a~\(\lambda\)-set of maximal cardinality, and assuming the existence of a~strongly inaccessible cardinal it is consistent that \(\mathfrak c\)~is a~limit cardinal and for every \(\kappa<\mathfrak c\) there is a~\(\lambda\)-set of size~\(\kappa\), yet there is no \(\lambda\)-set of size~\(\mathfrak c\). That is, the maximum cardinality of a~\(\lambda\)-set may not be attained.