On extremally disconnected topological groups (Q2502961): Difference between revisions

An ultrafilter~\(p\) on a~set~\(X\) of cardinality~\(\kappa\) is said to be partially selective if there is \(E\subseteq[X]^2\) such that \(\{x\in X:\{y\in X:\{x,y\}\in E\}\in p\}\in p\) and for every partition \(\{A_\alpha:\alpha<\kappa\}\) of~\(X\), either \(A_\alpha\in p\) for some~\(\alpha\) or there there exists \(A\in p\) such that \([A\cap A_\alpha]^2\cap E=\emptyset\) for all~\(\alpha\). For \(E=[X]^2\) we obtain a~definition of a~selective ultrafilter. Let us recall that a~set \(D\subseteq X\) is a~strongly discrete subset of~\(X\) if there is a~disjoint system of open sets \(U_x\), \(x\in D\), such that \(x\in U_x\). In the paper under review the author deals with a~well-known difficult question whether there is in ZFC a~nondiscrete extremally disconnected topological group. He proves that if \(G\) is an extremally disconnected topological group and \(p\) a~converging nonprincipal ultrafilter on~\(G\) containing a~strongly discrete subset of~\(G\), then \(p\)~is partially selective. Every nonprincipal partially selective ultrafilter on a~countable set can be mapped onto a~\(P\)-point in~\(\omega^*\). Therefore it is consistent with ZFC that there is no extremally disconnected topological group containing countable discrete nonclosed subsets.