On a refinement of anti-Souslin tree property (Q1375731): Difference between revisions

A tree has been said to be anti-Souslin if every uncountable subset contains an uncountable antichain. The notion of uncountable antichain has been refined, for \(\omega_1\)-trees, to stationary and cub antichain. More generally, a subset \(A\) of a tree is said to be stationary or cub if the level projection of \(A\) is a stationary or cub subset of \(\omega_1\). This paper introduces classes of trees, \(XY\) for \(X, Y \in \{ U,S,C\}^2\), where a tree is in \(XY\) if every \(X\) set contains a \(Y\) antichain. As you would guess, \(U\) denotes uncountable, \(S\) denotes stationary, and \(C\) denotes cub. The relationship with the familiar notions of \({\mathbb Q}\)-embeddable and \({\mathbb R}\)-embeddable is also investigated. The starting point is Baumgartner's early results that \(\diamondsuit\) implies that neither of the ZFC implications ``\({\mathbb Q}\)-embeddable implies \({\mathbb R}\)-embeddable'' and ``\({\mathbb R}\)-embeddable implies \(UU\)'' is reversable and Shelah's result that \(\diamondsuit^*\) implies there is an \({\mathbb R}\)-embeddable tree with no stationary antichain (and so is not \(CS\)). The main new results are that \(\diamondsuit\) implies ``\(CS \) does not imply \({\mathbb R}\)-embeddable'' and \(\diamondsuit^*\) implies that ``\({\mathbb R}\)-embeddable plus \(CS\) does not imply \(SS\)''. The author poses the question of whether \(SS\) implies \({\mathbb R}\)-embeddable.