Borel sets which are null or non-\(\sigma\)-finite for every translation invariant measure (Q2369429)

The paper investigates aspects of measurability for the set \(X_1\) of Liouville numbers, and for set \(X_n\) of non-Besicovich-Eggelston numbers. Recall: (1) \( x \in X_1 \Leftrightarrow x\) is irrational, and \( \forall \theta>0\) \( \exists p,q \in \mathbb N\) \((q>1)\) such that \(| x-p/q| < q^{-\theta}\). (2) \(x \in X_n\) (for \(n \in \mathbb N\;(n \neq 1)) \Leftrightarrow \) For the expansion \(0.x_1 x_2 x_3 \dots\) of the fractional part of \(x\) with respect to the basis \(n\), we have that there exists \(c \in \{0,1,\dots,n-1\}\) such that the limit \(\lim_{k \to \infty} \frac{1}{k} \text{card} \{i: x_i =c\), for \(0<i \leq k\}\) does not exist. The following summarizes the main results of the paper. Theorem 1. For each \(n \in \mathbb N\) we have that \(X_n\) is `immeasurable'. That is, for every translation-invariant measure \(\mu\) on \(\mathbb R\) we either have \(\mu(X_n)=0\) or otherwise \(X_n\) is not \(\sigma\)-finite set with respect to \(\mu\) (that is, \(X_n\) can not be written as a union of sets of finite \(\mu\)-measure). In order to prove this theorem, the authors derive the following useful criterion for checking immeasurability. Theorem 2. If \(X \subset \mathbb R\) is a \(G_\delta\)-set of Lebesgue measure zero such that \(\{ x \in \mathbb R: X+x \subset X\}\) is dense in \(\mathbb R\), then \(X\) is immeasurable. As an immediate application of Theorem 1 the authors obtain the result that for the Hausdorff measure \({\mathcal H}_\varphi\) with respect to some gauge function \(\varphi\), we have \({\mathcal H}_\varphi (X_1) \in \{0,\infty\}\), for any arbitrary gauge function \(\varphi\). Note that this result was also obtained recently by \textit{L. Olsen} [Manuscr. Math. 116, No. 2, 157--172 (2005; Zbl 1095.28012)] and \textit{L. Olsen} and \textit{D. L. Renfro} [ibid. 119, No. 2, 217--224 (2006; Zbl 1126.28007)]. Here, the reviewer likes to remark that the latter result can be viewed as being a folklore result, since it can be derived via a straight forward combination of well known techniques in Diophantine harmonic analysis developed by Khintchine, Jarnik, Besicovitch, and Nevanlinna.