Motivic bivariant characteristic classes (Q2445391)

The non expert person wishing to understand the historical motivations for the present paper should first consult the clear exposition by the second author in the review for [\textit{S. Yokura}, Publ. Res. Inst. Math. Sci. 30, No. 4, 603--610 (1994; Zbl 0836.14006)]. Such a reader might also find desirable to know what is the object of a bivariant theory of W. Fulton and R. MacPherson. The basic idea is to associate to each morphism \(f:X\to Y\) in a convenient category an abelian group given either by constructible functions, cycles, homology \(\dots\) It is required that in the concrete cases of identity \(\mathrm{id}:X\to X\) and \(X\to \mathrm{pt}\) respectively, one obtains the standard contravariant and covariant theories. Furthermore the machinery comes with theories of characteristic classes (or similar objets) parametrized by the space \(Y\). After this preparation one can peruse with profit the content of the abstract: ``Let \(K_0(\mathcal V/X)\) be the relative Grothendieck group of varieties over \(X\in Obj(\mathcal V)\), with \(\mathcal V=\mathcal V^{(\mathrm{qp})}_k\) (resp. \(\mathcal V=\mathcal V^{\mathrm{an}}_c\)) the category of (quasi-projective) algebraic (resp. compact complex analytic) varieties over a base field \(k\). We constructed the motivic Hirzebruch class transformation \({T_y}_*: K_0(\mathcal V /X) \to H_*(X) \otimes \mathbb Q[y]\) in the algebraic context for \(k\) of characteristic zero, with \(H_*(X)=\mathrm{CH}_*(X)\) (resp. in the complex algebraic or analytic context, with \(H_*(X)=H^{\mathrm{BM}}_{2*}(X)\)). It ``unifies'' the well-known three characteristic class transformations of singular varieties: MacPherson's Chern class, Baum-Fulton-MacPherson's Todd class and the \(L\)-class of Goresky-MacPherson and Cappell-Shaneson. In this paper we construct a bivariant relative Grothendieck group \(\mathbb K_0(\mathcal {V}/X \to Y)\) for \(\mathcal V=\mathcal V^{(\mathrm{qp})}_k\) (resp., \(\mathcal V^{\mathrm{an}}_c\)) so that \(\mathbb K_0(\mathcal V/X \to \mathrm{pt})=K_0(\mathcal V/X)\) in the algebraic context with \(k\) of characteristic zero (resp., complex analytic context). We also construct in the algebraic context (in any characteristic) two Grothendieck transformations \(m \mathcal C_y=\Lambda_y^{\mathrm{mot}}: \mathbb K_0(\mathcal V^{\mathrm{qp}}/X \to Y) \to\mathbb K_{alg}(X \to Y)\otimes \mathbb Z[y]\) and \(T_y: \mathbb K_0(\mathcal V^{\mathrm{qp}}/X \to Y) \to \mathbb H(X \to Y) \otimes \mathbb Q[y]\) with \(\mathbb K_{alg}(f)\) the bivariant algebraic \(K\)-theory of \(f\)-perfect complexes and \(\mathbb H\) the bivariant operational Chow groups (or the even degree bivariant homology in case \(k=\mathbb C\)). Evaluating at \(y=0\), we get a ``motivic'' lift \(T_0\) of Fulton-MacPherson's bivariant Riemann-Roch transformation \(\tau :\mathbb K_{{\mathrm{alg}}} \to \mathbb H \otimes \mathbb Q\). The covariant transformations \(\mathcal C_y: \mathbb K_0(\mathcal V^{\mathrm{qp}}/X \to \mathrm{pt}) \to G_0(X)\otimes \mathbb Z[y]\) and \(T_{y*}: \mathbb K_0(\mathcal V^{\mathrm{qp}}/X \to pt) \to H_*(X) \otimes \mathbb Q[y]\) agree for \(k\) of characteristic zero with our motivic Chern- and Hirzebruch class transformations defined on \(K_0(\mathcal V^{\mathrm{qp}}/X)\). Finally, evaluating at \(y=-1\), for \(k\) of characteristic zero we get a ``motivic'' lift \(T_{-1}\) of Ernström-Yokura's bivariant Chern class transformation \(\gamma: \tilde{\mathbb F}\to \mathrm{CH}\).''