On a theorem of Hildebrand (Q2420498): Difference between revisions

\textit{A. Hildebrand} [Mich. Math. J. 38, No. 2, 241--253 (1991; Zbl 0727.11036)] proved the following theorem: For any fixed \(k \in \mathbb{Z}^+\), if \(f: \mathbb{Z}^+ \rightarrow \mathbb{C}\) is a completely multiplicative function (i.e., \(f (mn)= f (m) f (n)\) for all \(m, n \in \mathbb{Z}^+ \)) taking its values in the \(k\)-th roots of unity, then the set of \(a\in \mathbb{Z}^+\) fulfilling \( f (a) + f (a+1)= 1\) is nonempty. The paper under review proves the following stronger statement: Let \(H\leq\mathbb{Q}^+\) be a (multiplicative) subgroup such that \(\mathbb{Q}^+/H\) is cyclic of finite order. Let \(H^*=H\cap \mathbb{Z}^+\). Then \(H^*\) is nonempty, as it is a ``large set'', called IP-set. A set is an IP-set, if it contains all finite sums of an infinite sequence of positive integers.