Characterizations of \(\mathbb{N}\)-compactness and realcompactness via ultrafilters in the absence of the axiom of choice (Q6971562)

The aim of the paper under review is to give choice-free proofs of some topological theorems which had been originally proved using the Axiom of Choice (AC). The authors work in the Zermelo-Fraenkel system of axioms ZF, which does not include AC.\N\NRecall that a topological space \(X\) is \textit{\( \mathbb N\)-compact} if there exists a non-empty set \(J\) such that \(X\) is homeomorphic to a closed subspace of \( \mathbb N^J\). \textit{Realcompactness} is defined in an analogous way with reference to \(\mathbb R\).\N\NThe authors present a proof in ZF of the Herrlich-Chew theorem, which states that a zero-dimensional \(T_1\) space \(X\) is \(\mathbb N\)-compact if and only if every clopen ultrafilter with the countable intersection property in \(X\) is fixed (recall that a filter is \textit{fixed} if the intersection of all its members is nonempty). A variation is also provided, assuming CMC, the countable axiom of multiple choice.\N\NAgain assuming CMC, the authors prove Hewitt's theorem, stating that a Tychonoff space \(X\) is realcompact if and only if every \(z\)-ultrafilter with the countable intersection property in \(X\) is fixed. It is stated as an open problem whether CMC is necessary in the proof of Hewitt's theorem; however, a modification of the theorem is introduced and proved in ZF.\N\NSeveral applications of the above results are presented in ZF, possibly assuming CMC. For example, among many other applications, the authors prove that every zero-dimensional Lindelöf \(T_1\)-space is \(\mathbb N\)-compact and, assuming CMC, that every Tychonoff Lindelöf space is realcompact. Moreover, after a review, the authors present new results about \(E\)-completely regular and \(E\)-compact spaces, for \(E\) an arbitrary topological space. A topological space \(X\) is \textit{\(E\)-completely regular} (\textit{\(E\)-compact}) if there exists a non-empty set \(J\) such that \(X\) is homeomorphic to a subspace (a closed subspace) of \(E^J\). The final section of the paper presents a list of open problems.\N\NThe paper is clearly written and largely self-contained; the new proofs are original and might be of interest also to topologists not particularly concerned with avoiding the use of AC.