\(L\)-spaces and the \(P\)-ideal dichotomy (Q624199): Difference between revisions

\textit{S. Todorčević} [Partition problems in topology. Contemporary Mathematics, 84. Providence, RI: American Mathematical Society (AMS) (1989; Zbl 0659.54001)] proved that his principle (\(\mathcal K\)) implies that there are no strong \(L\)-spaces of countable tightness; more precisely, \((*)\): a regular space with all finite powers Lindelöf contains no \(L\)-subspaces. The authors derive \((*)\) from the conjunction of \(\mathfrak{p}>\aleph_1\) and a weak version of the \(P\)-ideal dichotomy [cf. \textit{V. A. Lyubetskij}, Dokl. Math. 56, No.~3, 851--854 (1997; Zbl 0958.03040)]; rather than a decomposition into countably many sets orthogonal to the ideal, it asserts the existence of an uncountable such set. To show that this is not covered by Todorčević's result, they construct a model in which \(\mathfrak{p}>\aleph_1\) and in which the \(P\)-ideal dichotomy holds for \(\aleph_1\)-generated ideals but in which all Aronszajn trees are special.