Field change for the Cassels-Tate pairing and applications to class groups (Q6573191)

From MaRDI portal





scientific article; zbMATH DE number 7881679
Language Label Description Also known as
default for all languages
No label defined
    English
    Field change for the Cassels-Tate pairing and applications to class groups
    scientific article; zbMATH DE number 7881679

      Statements

      Field change for the Cassels-Tate pairing and applications to class groups (English)
      0 references
      0 references
      0 references
      16 July 2024
      0 references
      Establishing the triviality of the Tate-Shafarevich group \(\Sha(A)\) of an abelian variety \(A\) over a global field being a distant reality, Cassels and Tate considered the subset \(\Sha(A)_{div}\) of divisible elements. When \(A\) is an elliptic curve over a number field, Cassels first defined in 1962 a non-degenerate alternate pairing (which was generalized by Tate in the same year to \(A\) of higher dimensions using the -- at that time newly defined -- Tate duality): \N\[\N\Sha(A)/\Sha(A)_{\operatorname{div}} \times \Sha(A^{\vee})/\Sha(A^{\vee})_{\operatorname{div}} \rightarrow \mathbb{Q}/\mathbb{Z},\N\]\Nwhere \(A^{\vee}\) is the abelian variety dual to \(A\). When we consider global fields \(F\) of characteristic \(p>0\), Tate constructed a non-degenerate alternating bilinear pairing: \N\[\N\Sha(A)_{p\operatorname{-div}}/\Sha(A)_{\operatorname{div}} \times \Sha(A^{\vee}_{p\operatorname{-div}})/\Sha(A^{\vee})_{\operatorname{div}} \rightarrow \mathbb{Q}/\mathbb{Z},\N\]\Nwhere \(\Sha(A)_{p\operatorname{-div}}\) denotes the subset of \(p\)-divisible elements in \(\Sha(A)\). This Cassels-Tate pairing has been generalize over the years and, if \(A\) is identified with its dual by a principal polarization defined over \(F\), this pairing is known to be antisymmetric but not necessarily alternating. This last result was proved in [Ann. Math. (2) 150, No. 3, 1109--1149 (1999; Zbl 1024.11040)] by \textit{B. Poonen} and \textit{M. Stoll}.\N\NThe Cassels-Tate pairing can also be written for Selmer groups, and the present authors had observed in an earlier work that an analogue of this can be defined more generally, for any short exact sequence of finite Galois modules satisfying certain local conditions, and having order that is not a multiple of \(p\). In that work, the authors had defined a category SMod\(_F\) of so-called `Selmerable' modules. The objects of this additive category are pairs \((M,W)\) where \(M\) is a finite \(G_F\)-module with order, not a multiple of \(p\), and \(W\) is an open, compact subgroup of the (restricted direct) product \(\prod'_v H^1(G_v,M)\) where the local Galois group \(G_v\) is viewed as a subgroup of \(G_F\), and the restriction in the product is with respect to the classes in the unramified cohomology. The subgroup \(W\) represents the `local conditions' mentioned above. The morphisms are the \(G_F\)-homomorphisms from \(M\) to \(M'\) which map \(W\) into \(W'\). Associating to each object \((M,W)\) of SMod\(_F\), its Selmer group \(\mathrm{Sel}(M,W) :=\mathrm{Ker} \bigg( H^1(G_F,M) \rightarrow \prod'_v H^1(G_v,M)/W \bigg)\), we have a functor \((M,W) \mapsto \mathrm{Sel}(M,W)\) from SMod\(_F\) to the category of abelian groups. They obtain a more general theory of the Cassels-Tate pairing that is sufficiently flexible to compare pairings resulting from different short exact sequences. As a result, they were able to give a new, completely algebraic, proof of Poonen-Stoll's result on the failure of alternate-ness of the pairing in general abelian varieties of dimension \(>1\).\N\NIn the present paper, the authors build on the previous one with a view to applications to the study of class groups as well as to the distribution of Selmer groups in twist families. The earlier work did not address how to move between the categories SMod\(_K\) and SMod\(_F\) when \(K/F\) is a finite extension. One of the main results of the present paper shows that the Cassels-Tate pairing behaves well under a `restriction of scalars' functor from SMod\(_K\) to SMod\(_F\). The paper also describes how to apply this generality to the description of the distribution of class groups.\N\NNeedless to say, this paper contains strong results with potential applications to the resolution of diverse problems including natural questions on the distribution of class groups that embody Cohen-Lenstra-Martinet heuristics.
      0 references
      Cassels-Tate pairing
      0 references
      Cohen-Lenstra-Martinet
      0 references
      Selmer groups
      0 references
      category of selmerable modules
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references