Characters of countably tight spaces and inaccessible cardinals (Q386181)

A topological space \(X\) is countably tight (CT for short) at some \(x\in X\) if whenever \(A\subseteq X\) and \(x\in\overline{A}\), there exists a countable \(B\subseteq A\) such that \(x\in\overline{B}\); and it is indestructibly countably tight (ICT for short) at \(x\) if, in addition to being countably tight at \(x\), it remains so after forcing with the tree \(2^{<\omega_1}\) (equivalently, after forcing with any \(\sigma\)-closed forcing notion). \textit{M. Scheepers} [ibid. 161, 407--432 (2014; Zbl 1337.03071)] has found a characterization of indestructible countable tightness in terms of games: consider the game, of length \(\omega_1\), where in the \(\alpha\)-th inning player I plays a set \(A_\alpha\) such that \(x\in\overline{A_\alpha}\), and player II responds by picking a point \(x_\alpha\in A_\alpha\); and in the end player II wins if and only if \(x\in\overline{\{x_\alpha\mid \alpha<\omega_1\}}\). Then the space \(X\) is ICT at \(x\) if and only if (in addition to being CT at \(x\)) player I does not have a winning strategy in this game. From this topological characterization, it is quite easy to see that if \(X\) is CT at some \(x\in X\) and \(\chi(x,X)\leq\omega_1\) then \(X\) is ICT at \(x\), and Scheepers [loc. cit.] has proved that it is consistent that the converse of this statement does not hold (i.e. that there can consistently exist spaces \(X\) that are ICT at some \(x\in X\) satisfying \(\chi(x,X)>\omega_1\)). In the paper that occupies us in this review, the author provides the consistency that this converse statement holds, relative to an inaccessible cardinal: if there exists an inaccessible cardinal, then it is consistent that every space \(X\) which is ICT at some \(x\in X\) must satisfy \(\chi(x,X)\leq\omega_1\). The model witnessing this consistency statement is the one obtained by forcing with \(\mathrm{Col}(\omega_1,<\kappa)\) where \(\kappa\) is a (strongly) inaccessible cardinal. The author also proves the following: if there is an inaccessible cardinal, then it is consistent that every CT space \(X\) of size \(\omega_1\) has character distinct from \(\omega_2\) (although it is possible that the character of such a space is \(\omega_3\)). The forcing notion used for this result is \(\mathrm{Col}(\omega_1,<\kappa)\times\mathrm{Fn}(\kappa^+,2,\omega_1)\) (this is, after collapsing \(\kappa\) to \(\omega_2\), we add \(\kappa^+=\omega_3\) subsets to \(\omega_1\)), with \(\kappa\) a (strongly) inaccessible. On the other hand, if \(\omega_2\) is not inaccessible in the constructible universe \(L\) then there exists an ICT space (which is Tychonoff) \(X\) satisfying \(|X|=\omega_1\) and \(\chi(X)=\omega_2\). This space is of the form \(X=M\cap C_p(Y)\), where \(M\) is an elementary submodel, of size \(\omega_1\), of \(H_\theta\) for some large enough \(\theta\), and \(Y\) is a particular nicely behaved LOTS that \textit{R. R. Dias} and \textit{F. D. Tall} [ibid. 160, No. 18, 2411--2426 (2013; Zbl 1295.54026)] have obtained from a Kurepa tree. Thus, putting together the three results mentioned above, the author obtains that the following two statements are equiconsistent with the existence of an inaccessible cardinal (keep in mind that the author takes ``inaccessible'' to mean ``strongly inaccessible'', which is relevant when looking at the proofs but of course is not relevant in terms of consistency strength): (1) Every ICT space \(X\) of size \(\omega_1\) satisfies \(\chi(X)\leq\omega_1\), and (2) \(2^{\omega_1}>\omega_2\) and there is no CT space of size \(\omega_1\) and character \(\omega_2\).