Projective Hausdorff gaps (Q2436617): Difference between revisions

Let \(A,B\subseteq [\omega]^\omega\). We say that a pair \((A,B)\) is a \textit{gap} if for every \(a\in A\), \(b\in B\) the intersection \(a\cap b\) is finite and there is no \(c\in[\omega]^\omega\) such that \(\forall a\in A\,a\subseteq^*c\) and \(\forall b\in B\,b\subseteq^*\omega\setminus c\). A set \(A\subseteq [\omega]^\omega\) is \(\sigma\)\textit{-directed} if for every countable collection \(\{a_n:n\in\omega\}\subseteq A\) there exists \(a\in A\) such that \(a_n\subseteq^*a\) for all \(n\in\omega\). If \((A,B)\) is a gap and both \(A\) and \(B\) are \(\sigma\)-directed, we will call it a \textit{Hausdorff gap}. In [Fundam. Math. 150, No. 1, 55--66 (1996; Zbl 0851.04002)], \textit{S. Todorčević} proved that there cannot be a Hausdorff gap \((A,B)\) if \(A\) is analytic. In the paper under review, the author extends this result to the next level of the projective hierarchy. Namely, he proves that the statements ``There cannot be a Hausdorff gap \((A,B)\) if \(A\) is coanalytic'' and ``There cannot be a Hausdorff gap \((A,B)\) if \(A\) is \(\mathbf{\Sigma}^1_2\)'' are equivalent. In the second part of the paper, the author considers real-valued games corresponding to Hausdorff gaps and shows that assuming the axiom of real determinacy \(\text{AD}_{\mathbb{R}}\) for a projective pointclass \(\boldsymbol{\Gamma}\) there are no Hausdorff gaps \((A,B)\) if \(A\in\boldsymbol{\Gamma}\).