Abundance of wild historic behavior (Q2660145)

For a dynamical system \((X,f)\), a point \(x\in X\) is said to have historic behavior or to be irregular if the sequence \(\frac{1}{n_{k}} \sum_{j=1}^{n_{k}} \delta_{f^{j}(x)}\) does not converge in the weak* topology where \(\delta_{y}\) is the Dirac measure at point \(y.\) From Birkhoff's ergodic theorem, the irregular set is not detectable from the point of view of any invariant measure. However, the irregular set may have strong dynamical complexity in the sense of Hausdorff dimension, Lebesgue positive measure, topological entropy, topological pressure, distributional chaos and residual property. Here the authors prove that a kind of stronger historical behavior, called wild historical behavior, is a topologically generic subset (countable intersection of open and dense subsets) for wide classes of dynamical models. They consider a set function \[\tau_{x}(A)=\lim \sup _{n \rightarrow \infty} \frac{1}{n} \sum_{j=1}^{n} \delta_{f^{j}(x)}(A)\] that will serve as a pre-measure to obtain a Borel measure \(\eta_{x}\) through a classic well-known construction of Carathéodory. They prove that a point \(x\in X\) has historic behavior if and only if \(\eta_{x}\) is not a probability measure. A point \(x\in X\) is said to have wild historic behavior in \(\Lambda\) or to be a wild historic point if \(\eta_{x}\) gives infinite mass to every open subset of a compact invariant subset \(\Lambda\) for the dynamics. It is clear that a wild historic point has historic behavior. The paper contains several interesting results. For example, they prove that the set of points with wild historic behavior is a topologically generic subset in: (1) Every mixing topological Markov chain with a denumerable set of symbols (either one-sided or two-sided); (2) Every open continuous transitive and positively expansive map of a compact metric space; (3) Each local homeomorphism defined on an open dense subset of a compact space admitting an induced full branch Markov map; (4) Suspension semiflows, with bounded roof functions, over the local homeomorphisms of the previous item; (5) The stable set of any basic set \(\Lambda\) of either an Axiom A diffeomorphism, or an Axiom A vector field; (6) The support of an expanding measure for a \(C^{1+}\) local diffeomorphism away from a non-flat critical/singular set on a compact manifold; (7) The support of a non-atomic hyperbolic measure for a \(C^{1+}\) diffeomorphism, or a \(C^{1+}\) vector field of a compact manifold. They also use properties of the measure \(\eta_{x}\) to deduce some features of the involved dynamical system, like existence of heteroclinic connections from the existence of open sets of historic points. A well-known example of heteroclinic connections is given by ``Bowen eyes''. Theorem B gives a sufficient condition for a diffeomorphism to exhibit a heteroclinic attractor as in the example of \textit{R. Bowen} [Am. J. Math. 95, 429--460 (1973; Zbl 0282.58009)].