On automatic subsets of the Gaussian integers (Q505037)

\textit{A. Cobham} [Math. Syst. Theory 6, 164--192 (1972; Zbl 0253.02029)] introduced the notion of what is now called a \(k\)-automatic sequence. Such a sequence \((a_n)_{n \geq 0}\) is computed as follows: on input \(n\) expressed in base-\(k\), a finite automaton processes the digits, moving from state to state, until the input is completely read. Then, a function is applied to the last state reached to get \(a_n\). A set is called \(k\)-automatic if its characteristic sequence is \(k\)-automatic. \textit{J.-P. Allouche} et al. [Fractals 3, No. 4, 663--677 (1995; Zbl 0879.11010)] generalized this notion to, among other things, the Gaussian integers. In this paper, this idea is taken further. Using a mix of combinatorial and topological ideas, the authors prove that the set \(\{ a^n \, : \, n \geq 0 \}\) is not \(b\)-automatic if \(a, b \in \mathbb{Z}[i]\) are multiplicatively independent. A corollary answers a question from the paper of Allouche et al.~mentioned previously. The authors also prove that the set \(\mathbb{Z}\) is \(b\)-automatic if and only if \(b^j \in \mathbb{N}\) for some \(j \geq 1\).