P-filters and Cohen, random, and Laver forcing (Q2217238)

A subset \(A\) of a topological space \(X\) is called a P-set in \(X\) if, for every \(G_{\delta}\)-set \(H\) in \(X\) such that \(A\subseteq H\), the set \(A\) is contained in the interior of \(H\). A point \(x\) of a topological space \(X\) is called a P-point of \(X\) if the singleton \(\{x\}\) is a P-set. The symbol \(\omega^{\ast}\) denotes the remainder of the Čech-Stone compactification \(\beta\omega\) of the countable discrete space \(\omega\). It is known that ZFC+CH implies that \(\omega^{\ast}\) cannot be covered by nowhere dense P-sets. It is also known that, in some models of ZFC, \(\omega^{\ast}\) can be covered by nowhere dense P-sets. To shed more light on the problem of when \(\omega^{\ast}\) can be covered by nowhere dense P-sets, the author uses the following concept of a P-filter. For sets \(a,b\), it is said that \(a\) is almost included in \(b\) if \(a\setminus b\) is finite. A set \(a\) is a pseudointersection of a family \(\mathcal{A}\) if \(a\) is infinite and almost included in every member of \(\mathcal{A}\). A filter (base) \(\mathcal{F}\) on \(\omega\) is called a P-filter (base) on \(\omega\) if every non-empty countable subfamily of \(\mathcal{F}\) has a pseudointersection which belongs to \(\mathcal{F}\). A filter is called nowhere dense if it is a free filter which has no pseudointersection. A useful fact is that the statement that \(\omega^{\ast}\) can be covered by nowhere dense P-sets is the Stone dual of the statement that every ultrafilter on \(\omega\) has a nowhere dense P-subfilter. In this article, the author answers several interesting questions about P-filters in the Cohen, random, and Laver forcing extensions of models of CH. In what follows, it is assumed that the ground model of ZFC satisfies CH. Concerning P-points of \(\omega^{\ast}\), it is known that if \(\aleph_1\)-many Cohen reals are added to the ground model, then P-points of \(\omega^{\ast}\) exist in all random real forcing extensions of the model. It is unknown if P-points of \(\omega^{\ast}\) exist in all models obtained by adding more than \(\aleph_1\)-many random reals to the ground model. The author proves that if the ground model satisfies also \(\square_{\omega_1}\), then \(P\)-points of \(\omega^{\ast}\) exist in the extensions obtained by adding \(\aleph_2\)-many random reals. Answering a question posed in [\textit{P. Nyikos}, Arch. Math. Logic 59, No. 7--8, 1009--1026 (2020; Zbl 1461.03048)], the author proves that, in the standard Laver extension of the ground model, every ultrafilter on \(\omega\) has a nowhere dense P-subfilter, so, in Laver's model, \(\omega^{\ast}\) can be covered by nowhere dense P-sets. However, if the ground model satisfies \(\square_{\omega_1}\)and \(\aleph_2\)-many Cohen reals are added to it, then \(\omega^{\ast}\) cannot be covered by nowhere dense P-sets in the extension. Furthermore, the author solves another problem posed by P. Nyikos. The problem concerns scales in \(\omega^{\omega}\) and pre-orders on \(\omega^{\omega}\) determined by ultrafilters on \(\omega\). Namely, if \(\mathcal{U}\) is an ultrafilter on \(\omega\), then the pre-order \(<_{\mathcal{U}}\) on \(\omega^{\omega}\) is defined as follows: for all \(f,g\in\omega^{\omega}\), \(f<_{\mathcal{U}}g\) if and only if \(\{n\in\omega: f(n)<g(n)\}\in\mathcal{U}\). A scale in \(\omega^{\omega}\) is a \(<^{\ast}\)-unbounded set of monotone increasing functions in \(\omega^{\omega}\) which is well-ordered by \(<^{\ast}\). A filter \(\mathcal{F}\) on \(\omega\) is called non-meager if, for each strictly increasing \(f\in\omega^{\omega}\), there exists \(F\in\mathcal{F}\) such that the set \(\{n\in\omega: F\cap[f^{n}(0), f^{n+1}(0)]=\emptyset\}\) is infinite. A (descending) tower is a subset of \([\omega]^{\aleph_0}\) that is well-ordered by reverse almost inclusion and has no pseudointersection. Every tower is a nowhere dense P-filter base. The author observes that if \(\mathcal{U}\) is an ultrafilter on \(\omega\) such that there exists a scale in \(\omega^{\omega}\) which is not \(<_{\mathcal{U}}\)-unbounded, then \(\mathcal{U}\) has a non-meager P-subfilter with a base that is a tower. Next, the author proves that if \(\mathfrak{b}=\mathfrak{d}\), then there exists an ultrafilter \(\mathcal{U}\) on \(\omega\) such that \(\mathcal{U}\) has no non-meager P-subfilter having a base which is a tower. In consequence, \(\mathfrak{b}=\mathfrak{d}\) implies that there exists an ultrafilter \(\mathcal{U}\) on \(\omega\) such that every scale in \(\omega^{\omega}\) is \(<_{\mathcal{U}}\)-unbounded. This is a nice solution to a problem posed in [\textit{P. Nyikos}, loc. cit.].