Artin's primitive root conjecture is a famous unsolved problem in number theory. It says that for almost any integer "a" (with a few obvious exceptions like perfect squares), there are infinitely many prime numbers p for which a is a "primitive root" modulo p. Being a primitive root modulo p means that repeatedly multiplying a by itself generates all the nonzero remainders when dividing by p. Despite being stated nearly a century ago, this conjecture remains unproven, though partial results exist assuming other unproven hypotheses like the Generalized Riemann Hypothesis.
The central tool developed in this paper is a generalization of something called the "large sieve inequality." The classical large sieve is a powerful analytic technique that controls how evenly sequences of numbers are distributed across different arithmetic progressions. It works well in "abelian" settings, meaning situations involving simpler, commutative symmetry groups. Artin's conjecture, however, is naturally connected to more complex "non-abelian" symmetry groups, where the order of operations matters. The authors formulate what a non-abelian large sieve inequality should look like and show that if this inequality holds, then Artin's primitive root conjecture follows as a consequence.
Since the non-abelian large sieve inequality is itself not yet proven, the authors also work to provide evidence for it. Using a mathematical technique called duality, they derive unconditional results, meaning results that do not rely on any unproven assumptions, that point toward the truth of the non-abelian large sieve inequality. This creates a pathway: proving the new inequality would automatically prove Artin's conjecture, and the unconditional results make the inequality more plausible. The paper therefore connects two open problems in a way that focuses future research efforts.