Periodic points classify a family of Markov shifts (Q2566165): Difference between revisions

Given a finite abelian group \(G\), there is an associated two-dimensional Markov shift defined by the natural shift action on the set of all arrays \(x=(x_{(s,t)})\in G^{\mathbb{Z}^2}\) with the property that \(x_{(s,t+1)}=x_{(s,t)}+x_{(s+1,t)}\) for all \((s,t)\in\mathbb{Z}^2\). This is an example of a two-dimensional Markov shift with zero topological entropy. The reviewer [J. Number Theory 71, No. 1, 1--11 (1998; Zbl 0935.28008)] conjectured that this system \(\Sigma(G)\) would exhibit the rigidity property that \(\Sigma(G_1)\) and \(\Sigma(G_2)\) would have the same number of points of each period (a period being a subgroup of finite index in \(\mathbb Z^2\)) if and only if \(G_1\) and \(G_2\) are isomorphic groups, and proved this under the assumption that the order of either group is not divisible by a high power of \(2\) nor by the square of any Wieferich prime exceeding \(4\times10^{12}\); the notorious primes \(1093\) and \(3511\) had to be handled by hand in the argument. In this paper all the arithmetic conditions on the order of the group are removed using Teichmüller systems to study the polynomial rings arising, giving a proof of the conjecture in general and raising the possibility that a similar rigidity phenomena might become provable for any algebraic \(\mathbb Z^d\) action of low enough entropy rank.