Cardinal invariants of monotonically normal spaces (Q1361362): Difference between revisions

The author presents a fairly comprehensive study of the behaviour of the more familiar cardinal functions in the class of monotonically normal spaces. Of note are: the equality of density and hereditary density, the inequality \(\tau(X)\leq c(X)\), the equality \(d(\beta X)=d(X)\) and the fact that \(\pi w(X)=d(X)=hd(X)=h\pi w(X)\) --- the last fact being valid for subspaces of compact monotonically normal spaces. In the course of the investigation the author also shows that calibre \((\kappa^+,\kappa^+,2)\) is equivalent to density at most \(\kappa\) and that all other forms of calibre \((\kappa^+,\lambda,\mu)\) are equivalent to cellularity at most \(\kappa\). Reviewer's remark: Although the author ``vigorously defends the validity and merit of his `forcing and absoluteness' proof'' of \(d(X)=hd(X)\), it must be said that this proof contains a gap. This gap occurs in his claim that when one forces with a tree the generic branch will meet all levels of the tree. A counterexample to this claim can be found by making the tree rather wide: consider the tree obtained by joining the tree \(2^{<\alpha}\) for \(\alpha <\omega_1\) to a common root; every generic branch will be countable, yet the tree has \(\omega_1\)-many levels. The author's tree however has the \(\kappa^+\)-cc for a certain \(\kappa\) and every generic branch can be shown to be of size at most \(\kappa\). This readily gives a maximal antichain \(A\) and for each \(t\in A\) an ordinal \(\alpha_t<\kappa^+\) such that the subtree of elements comparable with \(t\) is of height \(\alpha_t\). As \(A\) has size at most \(\kappa\) this tells us that \(T\) has at most \(\kappa^+\) many levels and so is of size \(\kappa\) --- which was to be shown.