Approximate limits of \(m\)-compact spaces (Q1917171)

Recall that a space \(X\) is initially \(m\)-compact provided that for every open cover \({\mathcal U}\) of \(X\), if \(|{\mathcal U}|\leq m\) then \({\mathcal U}\) has a finite subcover; that \(X\) is strongly \(m\)-compact provided that for every filter base \({\mathcal F}\) on \(X\), if \(|{\mathcal F}|\leq m\) then there is a compact subset \(K\) of \(X\) such that \({\mathcal F}|K\) is a filter base. In this article the author studies approximate inverse systems in the sense of \textit{S. Mardešić} [Fundam. Math. 142, 241-255 (1993; Zbl 0810.54013)]. An approximate inverse system is a collection \(\mathbb{X}= \{X_a,p_{ab},A\}\), where \((A,\leq)\) is a directed preordered set, \(X_a\), \(a\in A\), is a topological space and \(p_{ab}:X_b\to X_a\), \(a\leq b\), are mappings such that \(p_{aa}= \text{id}\) and the following condition is satisfied: for each \(a\in A\) and each normal open cover \({\mathcal U}\) of \(X_a\) there is an index \(b\geq a\) such that \((p_{ac} p_{cd},p_{ad})\) are \({\mathcal U}\)-near, whenever \(a\leq b\leq c\leq d\), i.e., for each \(y\) there exists \(U\in{\mathcal U}\) with \(p_{ac}p_{cd}(y), p_{ad}(y)\in U\). Some theorems obtained by the author are the following. (3.5) Let \(\mathbb{X}=\{X_a,p_{ab},A\}\) be an approximate inverse system of nonempty initially \(|A|\)-compact Tikhonov spaces. If \(\prod X_a\) is initially \(|A|\)-compact, then \(X=\lim \mathbb{X}\) is a nonempty initially \(|A|\)-compact Tikhonov space. (3.16)--(3.18) Let \(\mathbb{X}=\{X_a,p_{ab},A\}\) be an approximate inverse system of normal spaces such that \(\prod X_a\) is initially \(|A|\)-compact. Then the following hold: (i) \(X=\lim \mathbb{X}\) is normal; (ii) \(\lim\beta \mathbb{X}\approx \beta\lim \mathbb{X}\); and (iii) the approximate map \({\mathbf p}=\{p_a:\lim \mathbb{X}\to X_a\), \(a\in A\}\) is an approximate resolution. Recall that a space \(X\) is said to satisfy condition \((1)_m\) provided that for every filter base \({\mathcal F}\) on \(X\), if \(|{\mathcal F}|\leq m\), then there exist a filter base \({\mathcal G}\) on \(X\) such that \(|{\mathcal G}|\leq m\) and a compact subset \(K\) of \(X\) such that \({\mathcal G}\) is finer than both \({\mathcal F}\) and the filter base of all open sets containing \(K\). This definition is due to \textit{J. E. Vaughan} [with the reviewer, Trans. Am. Math. Soc. 196, 177-189 (1974; Zbl 0296.54005)]. Using properties of \((1)_m\) obtained in the preceding article, the author proves several results such as (3.19) Let \(\mathbb{X}=\{X_a,p_{ab},A\}\) be an approximate inverse system of nonempty Tikhonov spaces. If each \(X_a\), \(a\in A\), satisfies \((1)_m\), \(m\geq|A|\), then \(X=\lim \mathbb{X}\) is a nonempty initially \(|A|\)-compact space. The author provides a proof that if each \(X_a\) is pseudocompact and \(r:X\to \mathbb{X}=\{X_a,p_{ab},A\}\) is an approximate resolution of a Tikhonov space \(X\), then \(X\) is pseudocompact. The article concludes with an example which shows that the concepts of approximate inverse limits and approximate resolution are different concepts in the class of countably compact spaces. Correction: in the author's reference no. 16, the name ``J. E. Vaughan'' should be listed as one of the authors of that article.