Invariants of upper motives (Q2439222)

The article roughly belongs to the mathematics around and inspired by Rost's degree formula. It follows the notion of \(p\)\textit{-equivalence }(for a chosen prime number \(p\)), as introduced by \textit{N. A. Karpenko} and \textit{A. S. Merkurjev} [Adv. Math. 205, No. 2, 410--433 (2006; Zbl 1119.14041)], and studies it for general proper smooth varieties over a field. It is a coarser equivalence than ordinary birational equivalence. To any smooth projective variety we may also attach its Chow motive, in the sense of classical Grothendieck pure motives, but with the homomorphisms formed from the Chow groups of correspondences, as opposed to the value of a Weil cohomology theory. Chow motives form an additive category where, thanks to the hugely bigger homomorphism groups as opposed to the category of schemes, it often happens that motives decompose into direct summands, e.g., projective homogeneous varieties over an algebraically closed field just become direct sums of Tate motives. Over an arbitrary field the situation is more subtle, and inspired by the classical works of Merkurjev, Suslin, and notably \textit{M. Rost} [Doc. Math., J. DMV 1, 319--393 (1996; Zbl 0864.14002)], studying the direct summand decomposition of the Chow motive can decode the arithmetic data of the projective homogeneous variety in a very useful and elegant way. This has been used with great effect in the study of quadratic forms for example. Following earlier work of \textit{N. A. Karpenko} [J. Reine Angew. Math. 677, 179--198 (2013; Zbl 1267.14009)], when considering the Chow motive of a variety with \(\mathbb{F}_{p}\) coefficients, one can isolated a notion of \textit{upper motive} (summands pinned down by a multiplicity one projector). Now, at least for all varieties where over the algebraic closure the Chow motive simplifies to Tate motives, there is a quite close connection between agreement of the upper motive summands and \(p\)-equivalence. The author now considers a certain ideal, generalizing index considerations (the precise set up is an axiomatic homology theory and sufficiently general to accomodate for example algebraic cobordism), and proves a strong result in the direction that this ideal \textit{only} depends on the upper motive of the variety. This generalizes several earlier results in the literature.