On a generalization of Artin's conjecture (Q959033)

The authors study a conjecture proposed by Golomb [\textit{S. W. Golomb} and \textit{P.-F. Lee}, IEEE Trans. Inform. Theory 53, 768--774 (2007; Zbl 1145.11334)]. It states that for integers \(a > 1\) squarefree and \(r > 0\), the set \(P(a,r)\) of primes \(\{p \equiv 1\bmod r\}\) such that \(a\) has order \((p-1)/r\) in \((\mathbb Z/p \mathbb Z)^*\) is infinite and has a positive natural density. As will be pointed out below, a corrected version of this conjecture was established already in 1977 by H. Lenstra, assuming the Generalized Riemann Hypothesis (GRH). In case \(r=1\) and \(p\) is in \(P(a,1)\), then \(a\) is said to be a primitive root modulo \(p\). In case \(r>1\) the reviewer coined the term `near primitive root' [\textit{P. Moree}, J. Number Theory 83, No. 1, 155--181 (2000; Zbl 1101.11326)]. Artin's primitive root conjecture is the case \(r=1\). C. Hooley established the latter conjecture assuming GRH in 1967 and showed that the set \(P(a,1)\) has a natural density which is a positive rational number times the Artin constant \(A\). The authors extend Hooley's argument and show that if \(a\) is even and \(r\) is odd, then Golomb's conjecture is true. However, using only the law of quadratic reciprocity it is easy to see that Golomb's conjecture is false if \(a\equiv 1\pmod4\), \(r\) is odd and \(a|r\). In this case the set of primes \(P(a,r)\) is finite (e.g. if \(a=5\) and \(r=5\)). By work of \textit{H. W. Lenstra} [Invent. Math. 42, 201--224 (1977; Zbl 0362.12012)], it is known that under GRH the set \(P(a,r)\) is finite iff \(a\equiv 1\pmod4\), \(r\) is odd and \(a|r\). Also he showed that if \(P(a,r)\) is infinite, then under GRH the set has a positive density which is a rational number times the Artin constant. An explicit formula for the density was provided by \textit{L. Murata} [Arch. Math. 57, No. 6, 555--565 (1991; Zbl 0755.11029)]. Instead of taking \(a>1\) to be a squarefree integer, one can also consider the case where \(a\) is a rational number. Here matters become considerably more difficult, but again the answer, assuming GRH, is implicit in the literature and due to Lenstra (ibid.), qualitative version, and \textit{S. S. Wagstaff} (quantitative version) [Acta Arith. 41, 141--15 (1982; Zbl 0496.10001)]. For the details the reader is referred to the Max-Planck-Institut für Mathematik preprint ``On Golomb's near-primitive root conjecture'' (MPIM 2009-106) by the reviewer. Thus to sum up, the existent literature has far more to offer concerning Golomb's conjecture than the present paper.