No label defined (Q4379987)

These lectures on mixed motives consist of an introduction and four sections: (1) The derived category of motives over a field; (2) The Tannakian category of mixed Tate motives over a field; (3) Motives with \({\mathbb Z}/n{\mathbb Z}\)-coefficients and étale cohomology, and (4) Chow polylogarithms. The introduction sets the stage for the foundations that should underly a good theory of mixed motives (over a field \(F\)): they should form an abelian category comprising the conjecturally semi-simple category of pure motives, and the category itself should be the heart of a larger triangulated category \({\mathcal D}(F)\), the derived category of mixed motives over \(F\). More generally, for any base scheme \(S\) one should have the derived category of mixed motivic sheaves \({\mathcal D}(S)\) over \(S\). For \(p:X\rightarrow\text{ Spec}(F)\) smooth projective one defines the motivic cohomology \(H^i_{\mathcal M}(X,{\mathbb Q}(j)):=\text{ Ext}^i_{{\mathcal D}(X)}({\mathbb Q}(0)_X,{\mathbb Q}(j)_X)\simeq\text{ Ext}^i_{{\mathcal D}(F)}({\mathbb Q}(0)_F,{\mathcal R}p_*{\mathbb Q}(j)_X)\), and one of the core conjectures of the theory is whether Bloch's higher Chow groups \(\text{CH}^r(X,n)\otimes{\mathbb Q}\) map isomorphically onto \(H^{2r-n}_{\mathcal M}(X,{\mathbb Q}(r))\). In (1) Voevodsky's work on a candidate for \({\mathcal D}(F)\) is discussed. It is important to note that this does not depend on any conjecture. For a perfect field \(k\) one defines an additive category \(\text{SmCore}(k)\) with objects smooth schemes of finite type over \(k\) and as morphisms \(X\rightarrow Y\) the free abelian group on closed subvarieties \(Z\subset X\times Y\) such that \(pr_1:Z\rightarrow X\) is finite and surjective onto a connected component of \(X\). The category of effective geometric motives \(DM^{eff}_{\text{ gm}}\) is then defined as the Karoubian (i.e. pseudo-abelian) envelope of a suitable localization of the homotopy category \({\mathcal H}^b(\text{SmCore}(k))\) of bounded complexes of objects in \(\text{SmCore}(k)\). One gets rid of the superscript `eff' by inverting the Tate object as usual. To use sheaf-theoretic methods one needs to use the Nisnevich topology. The upshot is an embedding of the highly mysterious category \(DM^{eff}_{\text{ gm}}\) as a full subcategory of bounded above Nisnevich sheaves (with transfer) and homotopy invariant cohomology. In (2) the Tannakian category \(\mathfrak{MTM}\) of mixed Tate motives over a field \(k\) is described. A graded pro-Lie algebra \({\mathcal L}={\mathcal L}_{-1}\oplus{\mathcal L}_{-2}\oplus\ldots\) over \({\mathbb Q}\) is constructed such that \(\mathfrak{MTM}\) can be identified with the category of finite-dimensional \({\mathbb Q}\)-representations of \({\mathcal L}\). Using algebraic cycles one constructs an augmented graded differential graded \({\mathbb Q}\)-algebra \(A^{\bullet}=\oplus_{n\geq 0}A^{\bullet}_n\), its bar construction \(B(A^{\bullet})\) and sets \(\chi_{\text{ Mot}}:=H^0(B(A^{\bullet}))\). One conjectures that \(A^{\bullet}\) is a \(K(\pi,1)\) in the sense of Sullivan. Let \({\mathcal C}\) denote the category of finite-dimensional graded corepresentations of \(\chi_{\text{ Mot}}\). The main result can be stated: (i) Objects of \({\mathcal C}\) admit mixed Tate Hodge and \(\ell\)-adic realizations. These arise from homomorphisms of Hopf algebras \(\chi_{\text{ Mot}}\rightarrow\chi_{\acute{e}t}\) (resp. \(\chi_{\text{ Betti}}\)), where \(\chi\) on the right is the Hopf algebra associated to the Tannakian category of mixed Tate \(\ell\)-adic (resp. Hodge) realizations; (ii) \(\text{ Ext}^i_{\mathcal C}({\mathbb Q},{\mathbb Q}(1))\simeq k^{\times}\otimes{\mathbb Q}\); (iii) \({\mathcal C}\) contains polylogarithm corepresentations with Hodge realizations as given by \textit{A. Beilinson} and \textit{P. Deligne} [``Interprétation motivique de la conjecture de Zagier'', Proc. Symp. Pure Math. 55, Pt. 2, 97-121 (1994; Zbl 0799.19004)]. Section (3) is concerned with the relation between motives with \({\mathbb Z}/n{\mathbb Z}\)-coefficients and étale cohomology. A key role is played by Suslin's and Voevodsky's singular homology \(H_{\bullet}^{sing}(X,{\mathbb Z}/n{\mathbb Z})\). For \(X\) a variety defined over an algebraically closed field of characteristic \(0\), this is isomorphic to \(H^{\acute{e}t}_{\bullet}(X,{\mathbb Z}/n{\mathbb Z})\). The central result here is: For an equidimensional quasi-projective variety \(X\) over an algebraically closed field of characteristic zero, one has \(\text{CH}^i(X,n;{\mathbb Z}/m{\mathbb Z})\simeq H^{2(d-i)+n}_{\acute{e}t}(X,{\mathbb Z}/m{\mathbb Z}(d-i))^*\), where \(i\geq d=\dim(X)\) and the RHS means the dual of cohomology with compact support. When \(X\) is smooth, this amounts to \(\text{CH}^i(X,n;{\mathbb Z}/m{\mathbb Z})\simeq H^{2i-n}_{\acute{e}t}(X,{\mathbb Z}/m{\mathbb Z}(i))\). The proof uses Voevodsky's \(h\)- and \(qfh\)-topologies. The final section interpretes polylogarithms in terms of the graded pro-Lie algebra of section (2). It recalls work of Beilinson and Deligne [loc. cit.]. It closes with a result on Goncharov's currents on Chow varieties of projective spaces in terms of Zagier's polylogarithms. For the entire collection see [Zbl 0882.00032].