PermalinkSubmitted by Mike Boyle on Thu, 07/27/2017 - 16:15

Various versions of ``conjugacy between topology and measure theory'' have been studied in the case of a mixing shift of finite type $S$ with a specified invariant Markov measure $\mu$. At one extreme, a conjugacy $\phi: (S,\mu ) \to (S,\mu')$ is only required to satisfy $\phi_* (\mu ) = \mu'$ with $\phi S = S'\phi$ on a set of full $\mu$ measure. At the other extreme, $\phi$ is also required to be a homeomorphism. Intermediate requirements on $\phi$ were studied by Keane and Smorodinsky, Parry, Schmidt, Marcus and Tuncel, Gomez ... For a brief review of this with references, and the still-open ``Good finitary conjecture'', see the discussion of Problem 9 in [1] (also available as ``OPSD'' at http://www.math.umd.edu/~mboyle/open/ ).

In another direction, one can consider homeomorphisms $S$ and $S'$ with respect to all invariant Borel probabilities (not just one), and ask for a topological or Borel conjugacy between the complements of sets of measure zero for all nonatomic invariant Borel probabilities. Hochman [2] proved striking universality results in this direction (for application of this work to $C^+$ diffeomorphisms, see [3]).

Finally, following Shelah and Weiss, one can consider $S$ and $S'$ as Borel systems, neglecting wandering sets; see [4] for dramatic progress on this long stalled study.

## Tags

## Comments

## FL

I understand: Notions of similarity intermediate between topological conjugacy and measurable isomorphism a.e.. right?

## Entropy-conjugacy refers to

Entropy-conjugacy refers to Bowen's paper [1].

See [2] for recent work.

See Problem 99

## References

## Various versions of `

Various versions of ``conjugacy between topology and measure theory'' have been studied in the case of a mixing shift of finite type $S$ with a specified invariant Markov measure $\mu$. At one extreme, a conjugacy $\phi: (S,\mu ) \to (S,\mu')$ is only required to satisfy $\phi_* (\mu ) = \mu'$ with $\phi S = S'\phi$ on a set of full $\mu$ measure. At the other extreme, $\phi$ is also required to be a homeomorphism. Intermediate requirements on $\phi$ were studied by Keane and Smorodinsky, Parry, Schmidt, Marcus and Tuncel, Gomez ... For a brief review of this with references, and the still-open ``Good finitary conjecture'', see the discussion of Problem 9 in [1] (also available as ``OPSD'' at http://www.math.umd.edu/~mboyle/open/ ).

In another direction, one can consider homeomorphisms $S$ and $S'$ with respect to all invariant Borel probabilities (not just one), and ask for a topological or Borel conjugacy between the complements of sets of measure zero for all nonatomic invariant Borel probabilities. Hochman [2] proved striking universality results in this direction (for application of this work to $C^+$ diffeomorphisms, see [3]).

Finally, following Shelah and Weiss, one can consider $S$ and $S'$ as Borel systems, neglecting wandering sets; see [4] for dramatic progress on this long stalled study.

## References

## Add a new comment