Simultaneous insolvability of exponential congruences (Q2672006)

Let \(a\) and \(b\) be fixed integers. The two-variable Artin conjecture states that the set of primes \(p\) such that the exponential congruence \(a^x\equiv b\bmod{p}\) is solvable for some integer \(x\) has a natural density. In the case \(a\) and \(b\) are multiplicatively independent this can be proven under the Generalized Riemann Hypothesis (GRH), otherwise unconditionally. In the first case the density equals the Stephens constant \(S=\prod_q(1-q/(q^3-1))\approx 0.57596\), where the product is over all primes \(q\), multiplied by a computable rational number, see [\textit{P. Moree} and \textit{P. Stevenhagen}, J. Number Theory 85, No. 2, 291--304 (2000; Zbl 0966.11042)], in the second case the density is a rational number. \par Here systems of congruences \(a_i^x\equiv b_i\bmod{p}\), for \(1\le i\le n\) are studied. In an earlier paper the author [Proc. Am. Math. Soc. 149, No. 9, 3651--3668 (2021; Zbl 1479.11171)] showed under GRH that if \(a_i,b_i>1\), then there are infinitely many primes such that all of these exponential congruences are solvable. In the present paper he gives a necessary and sufficient condition for the infinitude of primes \(p\) such that none of these equations is solvable. In case this set of primes is infinite, then their lower density is positive. This result is unconditional. \par The proof starts by relating the problem to divisibility questions of the multiplicative order. Then techniques used in the study of Artin's primitive root conjecture are employed. In particular, the relevant set of primes are expressed as a countable disjoint union of sets of primes with suitable Artin symbols and the Chebotarev's density theorem is put in action. The relevant number fields involve higher roots of the numbers \(a_1,\ldots,a_n\). The author considers their Galois groups and their Kummer theory. \par The author also shows that given integers \(a_1,\ldots,a_n\) with \(|a_i|>1\) and non-zero integers with \(m_1,\ldots,m_n\), there are infinitely many primes \(p\) such that \(m_i\) divides the multiplicative order modulo \(p\) of \(a_i\) for all \(i\). In addition, he shows that this set of primes has a density that is positive. For \(i=1\) this problem has been well-researched, see for example, [the reviewer, Funct. Approximatio, Comment. Math. 33, 85--95 (2005; Zbl 1228.11152)].