Grothendieck's pairing on component groups of Jacobians. (Q1608575): Difference between revisions

Let \(R\) be a discrete valuation ring with field of fractions \(K\). Let \(A_K\) be an abelian variety over \(K\) with dual variety \(A_K'\). Then, with regard to their respective Néron models \(A\) and \(A'\), there is a bilinear pairing of the component groups of these Néron models \[ \langle,\rangle: \Phi_A \times \Phi_{A'} \to\mathbb{Q}/ \mathbb{Z}, \] which has been introduced by \textit{A. Grothendieck} [SGA 7, I, Lect. Notes Math., 288, Springer Verlag (1972; Zbl 0237.00013)] thirty years ago. Grothendieck himself conjectured that this pairing is perfect. In the meantime, Grothendieck's conjecture has been proved in various special cases, but also series of counter-examples in other cases have been found. In the paper under review, the authors investigate the case of the Jacobian \(J_K\) of a smooth proper curve \(X_K\) admitting a \(K\)-rational point. Their main result is an explicit formula for the pairing \(\langle,\rangle\), which is then used to prove Grothendieck's conjecture for certain types of such Jacobians, on the one hand, and disprove it for particular Jacobians in the case where the residue field of the ground ring \(R\) is imperfect. The method of proof is purely geometric and based on the invention of a certain pairing attached to a symmetric matrix, which leads to a more practical description of Grothendieck's pairing. Many explicit examples, at the end of the paper, demonstrate the power of the authors' approach.