Enhanced Koszul properties in Galois cohomology (Q2183187)

The Galois cohomology ring \(H^*(K, \mathbb{F}_p)\) of the field \(K\) (assumed to contain a primitive \(p\)-th root of unity, where \(p\) is a prime) is a quadratic algebra, that is, it is generated by elements of degree 1 subject to relations of degree 2. This is a consequence of the Milnor-Bloch-Kato conjecture, now a difficult theorem by Voevodsky and Rost. It is natural to wonder whether it is in fact a Koszul algebra -- a question first asked by Positselski. In this paper, the authors investigate several stronger conditions -- universal Koszulity, strong Koszulity, and so on. They show in particular that universal Koszulity is a property which is stable under certain basic operations on algebras (direct sums and twisted extensions). This blends very well with the elementary type conjecture. The latter describes groups of the form \(G_K(p)\), that is, the largest pro-\(p\) quotient of the absolute Galois group of a field \(K\) as above. (The cohomology of this group is just \(H^*(K, \mathbb{F}_p)\).) According to the conjecture, such groups can be obtained, when they are finitely generated, from free groups and Demushkin groups using free products and semidirect products, and as a result, their cohomology rings are obtained from basic pieces using direct sums and twisted extensions. Thus it follows from the results of this paper that, if the elementary type conjecture is true, then the mod \(p\) cohomology of a field \(K\) is Koszul if it is finitely generated as an algebra. It is worth pointing out that the authors have included some very welcome general material on Koszulity, making the paper very accessible.