No label defined (Q4520125)

The author investigates actions (by pseudohomeomorphisms -- i.e, category preserving bijections) of countable groups on perfect Polish spaces (equivalently, Borel-measurable equivalence relations with countable classes, on such spaces), as introduced by \textit{D. Sullivan, B. Weiss} and \textit{J. D. M. Wright} [Trans. Am. Math. Soc. 295, 795-809 (1986; Zbl 0617.46069)], and cocycles defined on them. All statements are modulo first category sets. The results are topological analogs of almost everywhere statements [see \textit{S. I. Bezuglyj} and \textit{V. Ya. Golodets}, Math. USSR, Izv. 29, 1-18 (1987; Zbl 0636.28009); \textit{V. Ya. Golodets} and \textit{S. D. Sinel'shchikov}, J. Funct. Anal. 121, No. 2, 455-485 (1994; Zbl 0821.28010); Publ. Res. Inst. Math. Sci. 23, No. 5, 737-769 (1987; Zbl 0656.46053)] and concern existence of ergodic cocycles, weak equivalence of all ergodic cocycles, and outer conjugacy conditions for actions of groups selected from the normalizer of an ergodic group action. More precisely, let \(\Gamma\) denote an ergodic (i.e., point transitive) countable group action on a perfect Polish space \(X\), let \(G\) denote a countable group. Then Theorem 3.9: There exists an ergodic \(G\)-valued cocycle over \(\Gamma\). Theorem 4.1: Any two ergodic \(G\)-valued cocycles \(\phi\), \(\psi\) over \(\Gamma\) are weakly equivalent, i.e., \[ \phi(y,\gamma y) = \psi(\Theta y, \Theta\gamma y) \] for any \(\gamma\in \Gamma\), \(y\in Y\), where \(\Theta\) is an element of the normalizer of \(\Gamma\) (full group-preserving pseudohomeomorphism) which is a homeomorphism on a dense \(G_\delta\) \(\Gamma\)-invariant subset \(Y\subset X\). Theorem 5.3: Let \(\pi_1\) and \(\pi_2\) be two actions of \(G\) by elements of the normalizer of \(\Gamma\). Then these actions are outer conjugate (i.e., \(\pi_1(g)y = \Theta^{-1}\pi_2(g)\tau(g)\Theta y\) for each \(y\in Y\), where \(Y\) and \(\Theta\) are as in the previous theorem, and \(\tau(g)\) is an element of the full group of \(\Gamma\)) if only if \(\pi_1(g)\) belong to the full group of \(\Gamma\) for exactly the same elements \(g\in G\) as \(\pi_2(g)\) do.