Representing free Boolean algebras (Q4344952)

The author proves, assuming the Proper Forcing Axiom, a reflection theorem for ideals of countable sets: if \(\mathcal I\) and \(\mathcal J\) are such ideals and \(\mathcal I\) is a \(P\)-ideal then either they can be countably separated --- meaning that there are countably many sets \(X_n\) such that for each pair \(\langle I,J\rangle\) in \(\mathcal I\times J\) there is an \(n\) such that \(I\subseteq^* X_n\) and \(J\cap X_n=^*\emptyset\) --- or there are \(\aleph_1\)-sized subfamilies \(\mathcal I'\) of \(\mathcal I\) and \(\mathcal J'\) of \(\mathcal J\) that cannot be countably separated. In fact the families \(\mathcal I'\) and \(\mathcal J'\) can be chosen with considerable structure: they can be enumerated in type \(\omega_1\) such that either they form a Hausdorff-Luzin-type gap --- meaning that for every \(\alpha<\omega_1\) and every \(m<\omega\) there are only finitely many \(\beta<\omega_1\) such that \(I_\alpha\cap J_\beta\) has size \(m\) or less --- or they form what the author calls a finite-to-one graph type family --- this means the \(J_\beta\) are pairwise disjoint and that each \(I_\alpha\) contains the graph of a finite-to-one function \(g_\alpha:\alpha\to\omega\), where \(\bigcup\mathcal J'\) is identified with \(\omega_1\times\omega\). This result allows the author to give unified proofs, from PFA, of statements that were previously shown to be consistent by collapsing a supercompact cardinal: 1) if \(\kappa>\aleph_1\) is regular then \(\kappa^*\setminus U(\kappa)\) is \(C^*\)-embedded in \(\beta\kappa\) and {} 2) a countably tight space is first-countable if{}f every \({}\leq\aleph_1\)-sized subspace is first-countable. The last two sections of the paper contain information about \(\omega^*\). First the author shows that Martin's Axiom and the Open Colouring Axiom together imply that there is a dense \(C^*\)-embedded subspace in \(\omega^*\) of size \(\mathfrak c\); the same is shown to hold in the model obtained by adding \(2^{\mathfrak c}\) Cohen reals. The second result about \(\omega^*\) is that in models obtained by iterated Miller forcing \(\omega^*\setminus\{p\}\) is \(C^*\)-embedded for some but not all \(p\): it is if{}f \(p\) is not a \(P\)-point.