Primitive substitutive numbers are closed under rational multiplication (Q1292623): Difference between revisions

Is the sum of two ``deterministic'' real numbers a ``deterministic'' real number? And what happens if deterministic is replaced by a more restrictive condition? Lehr proved that the set of real numbers whose base \(r\) expansion is \(q\)-automatic is a \(\mathbb{Q}\)-vector space [\textit{S. Lehr}, Theor. Comput. Sci. 108, 385-391 (1993; Zbl 0768.11013)]. This set is not an algebra [\textit{S. Lehr}, \textit{J. Shallit} and \textit{J. Tromp}, Theor. Comput. Sci. 163, 193-210 (1996; Zbl 0874.11029)]. In the paper under review the authors study the set of real numbers whose base \(r\) expansion is ultimately primitive substitutive (i.e., has a tail which is the letterwise image of an infinite fixed point of a primitive morphism). They prove that this set is closed under multiplication by a rational number, but not under addition. They actually give two real numbers whose base 10 expansions are fixed by the same primitive (constant length) morphism, but whose sum has no primitive substitutive tail. Note that in Ref. [4] ``Theoret. Comp. Sys.'' should be replaced by ``Theor. Comput. Sci''.