Sets of \(k\)-recurrence but not \((k+1)\)-recurrence (Q877530)

Inspired by \textit{H. Furstenberg's} proof [J. Anal. Math. 31, 204--256 (1977; Zbl 0347.28016)] of \textit{E. Szemerédi's} theorem [Acta Arith. 27, 199--245 (1975; Zbl 0303.10056)] that every set of integers with positive density contains arbitrarily long arithmetic progressions, define a set of \(k\)-recurrence to be a subset \(S\) of the integers such that for every probability preserving dynamical system \((X,K,\mu,T)\), if \(A\in K\) and \(\mu(A)> 0\), then there exists \(n\in S\) such that \(\mu(A\cap T^{-n} A\cap T^{-2n} A\cap\cdots\cap T^{-kn}A)> 0\). The authors answer positively a question of \textit{V. Bergelson} [Pollicott, Mark (ed.) et al., Ergodic theory of \(\mathbb Z^d\) actions. Proceedings of the Warwick symposium, Warwick, UK, 1993-94. New York: Cambridge University Press. Lond. Math. Soc. Lect. Note Ser. 228, 1--61 (1996; Zbl 0846.05095)] by proving the following nice result: for any irrational \(\alpha\), the set \[ S_k= \{n\in\mathbb N;\;\{n^k\alpha\}\in [1/4,3/4]\} \] is a set of \((k-1)\)-recurrence but not a set of \(k\)-recurrence. Please note that reference by \textit{T. Ziegler} appeared in [J. Am. Math. Soc. 20, No. 1, 53--97 (2007; Zbl 1198.37014)].