Motivic complexes over finite fields and the ring of correspondences at the generic point (Q1038556)

Already in the 1960s Grothendieck understood that one could obtain an almost entirely satisfactory theory of motives with rational coefficients over a finite field when one assumes the Tate conjecture. The main result of the the paper under review is a similar result for motivic complexes. In a first section the statements of the Tate conjecture and of the generalised Tate conjecture are recalled and proven to be equivalent. The construction of ``the'' category of motivic complexes over a finite field \(k\) proceeds then in essentially two steps. In a first step a semisimple Tannakian category of motives \(\mathcal{M}_{\mathrm{num}}(k)\) over \(k\) is constructed. This is essentially done by modifying the commutativity constraint in Grothendieck's category of pure motives (for numerical equivalence), using the Weil conjectures proven by Deligne. The group of Weil \(q\)-numbers can be identified with a subgroup of the fundamental group \(\pi(\mathcal{M}_{\mathrm{num}}(k))\), and it is shown that equality holds if and only if the Tate conjecture holds over \(k\). The second step of the construction consists of showing that the bounded derived category \(\mathcal{D}^{\mathrm{b}}(\mathcal{M}_{\mathrm{num}}(k))\) is essentially the only candidate for a good triangulated category of motivic complexes. This category indeed has the hoped for properties if and only if the Tate conjecture holds over \(k\) and numerical equivalence coincides with rational equivalence (always for rational coefficients). In a last section, the (conditional) results are used to produce an explicit description of Beilinson's algebra of correspondences at the generic point \(A(X)\), for a smooth projective variety \(X\) over \(k\). Reviewer's remark: The reader might wish to consult the appendix (12p.) in the version of the paper provided on the first author's webpage, which fills in some details and proofs.