Co-analytic spaces, \(K\)-analytic spaces, and definable versions of Menger's conjecture (Q2215683)

In the paper under review, a completely regular topological space is said to be: \textit{analytic} if it is a continuous image of the space \(\mathbb{P}\) of irrationals; \textit{Lusin} if it is an injective continuous image of a closed subspace of \(\mathbb{P}\); \textit{\(K\)-analytic} if it is a continuous image of a Lindelöf Čech-complete space; \textit{\(K\)-Lusin} if it is an injective continuous image of a Lindelöf Čech-complete space; \textit{co-analytic} if its remainder in some compactification is analytic; \textit{co-\(K\)-analytic} if its Stone-Čech remainder is \(K\)-analytic. The first result of the paper establishes that co-analytic spaces are co-\(K\)-analytic. The above definitions extend, for beyond metrizable spaces, a number of notions of Descriptive Set Theory (and that we usually see at work on Polish spaces). The spirit of the paper lies in the discussion, in the realm of topological spaces, of a pattern that has been studied by analysts for more than a hundred years: \textit{``simply defined sets of reals have a nice property, e.g. Borel sets are measurable, but using the Axiom of Choice, we can construct a complicated non-measurable set''\,}. Rather than discussing measurability, the investigation is in the field of \textit{Selection Principles}: more precisely, the focus is on \textit{Menger spaces} (a topological space is said to be a Menger space if whenever \(\{\mathcal{U}_n: n < \omega\}\) is a sequence of open covers, there exist finite \(\mathcal{V}_n\) such that \(\mathcal{V}_n \subseteq \mathcal{U}_n\) for all \(n < \omega\) and \(\bigcup\limits_{n < \omega} \mathcal{V}_n\) is an open cover). Menger's conjecture was that Menger subsets of \(\mathbb{R}\) should be \(\sigma\)-compact. \textit{W. Hurewicz} [Math. Z. 24, 401--421 (1925; JFM 51.0454.02)] has shown that Menger analytic metrizable spaces are \(\sigma\)-compact, and later \textit{A. V. Arkhangel'skij} [Sov. Math., Dokl. 33, 396--399 (1986; Zbl 0606.54013); translation from Dokl. Akad. Nauk SSSR 287, 525--528 (1986)] has generalized the preceding result by showing that Menger analytic spaces are \(\sigma\)-compact. Menger's conjecture was disproved in \(\mathbf{ZFC}\) by \textit{A. W. Miller} and \textit{D. H. Fremlin} [Fundam. Math. 129, No. 1, 17--33 (1988; Zbl 0665.54026)], and they also showed that, modulo a large cardinal, it is undecidable whether Menger co-analytic sets of reals are \(\sigma\)-compact. In a previous work, the author and \textit{S. Tokgöz} have shown [Topology Appl. 220, 111--117 (2017; Zbl 1423.54052)] that \(\mathbf{CD}\), the \textit{Axiom of Co-Analytic Determinacy} -- which follows from the existence of a measurable cardinal -- implies Menger co-analytic sets of reals are \(\sigma\)-compact. And, indeed, \(\mathbf{PD}\), the \textit{Axiom of Projective Determinacy}, implies Menger projective sets of reals are \(\sigma\)-compact (see also [\textit{F. D. Tall}, Topology Appl. 158, No. 18, 2556--2563 (2011; Zbl 1242.54009); \textit{F. D. Tall} and \textit{S. Tokgöz}, ibid. 220, 111--117 (2017; Zbl 1423.54052)]). Recall that projective sets come from closed sets via a finite process of taking complements and projections. In the paper under review, the author shows that \(\mathbf{CD}\) implies every metrizable Menger co-\(K\)-analytic space is \(\sigma\)-compact, and uses this result to show that \(\mathbf{CD}\) implies every Menger co-\(K\)-analytical group is \(\sigma\)-compact. After hearing about the latter result, \textit{S. Tokgöz} has shown in [Turk. J. Math. 42, No. 1, 12--20 (2018; Zbl 1424.03021)] that under \(\mathbf{V = L}\) there is a Menger co-analytical group of reals which is not \(\sigma\)-compact, and thus one concludes that it is undecidable (modulo large cardinals) whether Menger co-analytic topological groups are \(\sigma\)-compact. A large number of related results are established in the paper. The author also poses some questions. For instance: Are Menger \(K\)-Lusin spaces \(\sigma\)-compact? The author shows that if one strengthens Menger to Rothberger, then a positive answer is obtained (a topological space is Rothberger if whenever \(\{\mathcal{U}_n: n < \omega\}\) is a sequence of open covers, there is a cover \(\{U_n: n < \omega\}\) with \(U_n \in \mathcal{U}_n\) for all \(n < \omega\)). Are productively Lindelöf co-analytic spaces \(\sigma\)-compact? Productively Lindelöf spaces are those spaces \(X\) whose product \(X \times Y\) is Lindelöf whenever \(Y\) is a Lindelöf space. Is it consistent that co-analytic Menger spaces are \(\sigma\)-compact? In \(\mathbf{ZFC}\), is there a co-\(K\)-analytic Menger space which is not \(\sigma\)-compact?