Strongly independent matrices and rigidity of \(\times A\)-invariant measures on \(n\)-torus (Q6058279)

The authors generalize some results related to the Fustenberg's conjecture about ergodic invariant measures of non-lacunary multiplicative semigroup actions of positive integers over the circle \(\mathbb{T}\), to integer matrix actions over the \(n\)-torus. In order to do this they introduce the notion of strongly independent matrices over a field. Roughly speaking, an \(n\)-tuple of \(n\times n\) matrices over a field is said to be \emph{strongly independent} if for every fixed non-zero vector \(v\) the vectors obtained from the action of these matrices over \(v\) are linearly independent. Moreover, a single \(n\times n\) matrix is said to be strongly independent if the \(n\)-tuple comprised of the identity together with the first \(n-1\) powers of \(B\) is strongly independent. Lemma 3.1 in the paper presents a criterion to determine whenever an \(n\)-tuple of matrices is strongly independent. Using this criterion, the authors show that the strong independence of a matrix is equivalent to the irreducibility of its characteristic polynomial (see Theorem 1.4). A consequence of this result is that algebraically closed fields do not have strongly independent \(n\times n\) matrices for \(n\geq 2\) (see Theorem 1.5). The main result of the paper is Theorem 1.6 which states three different sufficient conditions for a Borel probability measure on \(\mathbb{T}^n\) that is invariant under the action of an integer matrix to be either finitely supported or the Lebesgue measure. Each condition involves some dynamical property of the measure (ergodicity, weak mixing, strong mixing) together with some conditions about the existence of tuples of integer matrices that are strongly independent over the rationals with additional properties. As a corollary the authors are able to determine the existence of an abelian multiplicative semigroup of integer \(n\times n\) matrices and a matrix \(B\) in this subgroup such that the Lebesgue measure is the unique non-atomic Borel probability measure on the torus that is invariant under the action of each matrix in the subgroup and ergodic for the action of \(B\) (see Corollary 1).