A conjecture of Gleason on the foundations of geometry (Q386210)

In [Algebr. Topol. Foundations Geom., Proc. Colloq. Utrecht, August 1959, 39--44 (1962; Zbl 0107.02702)], \textit{A. M. Gleason} posed a conjecture on the possibility of topologizing a group acting on a topological space so that a continuous action would ensue. The conjectured conditions on the group~\(G\) and the space~\(M\) were: (1)~\(M\)~is a Polish space and \(G\)~is a group of homeomorphisms that acts transitively on~\(M\), and (2)~there are \(n\in \mathbb N\) and \(\langle m_1,\dots,m_n\rangle\in M^n\) such that \(g\mapsto\langle gm_1,\dots,gm_n\rangle\) is injective and for all \(m\in M\) the subset \(\{\langle gm_1,\dots,gm_n,gm\rangle:g\in G\}\) of \(M^{n+1}\) is analytic. The authors disprove the conjecture but show that some modifications of~(2) do lead to positive results: one should require that \(\{\langle gm_1,\dots,gm_n\rangle:g\in G\}\) is a \(G_\delta\)-set and that \(\{\langle gm_1,\dots,gm_n,gm\rangle:g\in G\}\) is Borel. In this case one can even take a point in~\(M^{\mathbb N}\) rather than in a finite power. Examples show that the conditions are not necessary.