Chebyshev's bias in Galois extensions of global function fields (Q555288): Difference between revisions

Chebyshev's bias is usually referred to Chebyshev's observation that there seem to be less primes in the arithmetic progression 1 modulo \(4\) than in the arithmetic progression \(3\) modulo \(4\). In the paper under review, the authors study the Chebyshev bias in a finite, not necessarily abelian, Galois extension of global function fields. They show that when the extension is geometric and a certain hypothesis called \textit{Linear Independence (LI)} is satisfied, the fewer squares a conjugacy class of the Galois group has, the more primes there are whose Frobenius conjugacy class is equal to the given conjugacy class. The LI is a function field version of Rubinstein and Sarnak's \textit{Grand Simplicity Hypothesis}, and states that \(\pi\) and the arguments of the zeros of the Artin \(L\)-functions associated to the irreducible representations of the Galois group are linearly independent over the field of rational numbers. Most of the results are similar to those obtained by \textit{M. Rubinstein} and \textit{P. Sarnak}, [``Chebyshev's bias'', Exp. Math. 3, No. 3, 173--197 (1994; Zbl 0823.11050)] for the number field case as well as the work by the first author [``Chebyshev's bias in function fields'', Compos. Math. 144, No. 6, 1351--1374 (2008; Zbl 1233.11099)] in the case of the polynomial rings over finite fields. The authors also give, under (LI) and following the technology of Rubinstein and Sarnak, the necessary and sufficient conditions for a certain limiting distribution to be symmetric. They provide examples when LI holds, as well as examples when LI does not hold. In the last section of the paper they study the case when the Galois extension is a scalar field extension and describe the complete picture of the prime number race in this particular case.