On strictly ergodic models which are not almost topologically conjugate (Q1178337)

This paper settles a question raised by S. Glasner and D. Rudolph. Let \((X,{\mathcal B},\mu,T)\) be a topological dynamical system, where \(X\) is a compact metric space, \(T\) is a homeomorphism of \(X\) and \(\mu\) is a \(T\)- invariant Borel probability measure on \(X\). Two such systems are almost topologically conjugate (ATC) iff one can find in both systems residual invariant Borel sets of full measure such that the subsystems defined by these two sets are topologically conjugate. A well-known theorem (due to R. Jewett and W. Krieger) says that given any such ergodic system, one can always find a metrically isomorphic, uniquely ergodic topological model for it. On the other hand M. Denker and M. Keane showed that given any such ergodic system that is not strictly ergodic, it is always possible to find another topological dynamical system that is metrically isomorphic but not finitary isomorphic to the given one. S. Glasner and D. Rudolph developed this result further by showing that in fact one can find uncountably many systems metrically isomorphic but not ATC to the given one. They are also pairwise not ATC to each other. The authors improve this result further by showing that one can in fact choose such an uncountable family so that each system in it is strictly ergodic. This answers affirmatively a question of S. Glasner and D. Rudolph. A key result in these construction is a perturbation theorem for measurable partitions which may be of independent interest to ergodic theorists.