Metric groups attached to skew-symmetric biextensions (Q969612)

Let \(k\) be a field of characteristic \(p>0,\) and let \(G\) be a perfect, commutative, connected, unipotent \(k\)-group scheme. Let \(E\) be a skew-symmetric biextension of \(G\times G.\) It is well-known that such biextensions correspond to homomorphisms \(f:G\rightarrow G^{\ast}\). Set \(A=\pi_{0}\left( \ker f\right) .\) This gives rise to a map \(q:A\rightarrow \mathbb{Q}_{p}/\mathbb{Z}_{p}\) which is a nondegenerate quadratic form. This gives \(A\) the structure of a metric group, denoted \(\left( A,q\right) .\) The main result of this work proves a two-part conjecture of Drinfeld concerning \(A\). First, it is shown that \(A\) has cardinality an even power of \(p\). For the second, one can consider equivalence classes of metric groups over \(k\), two such groups being equivalent if they have isomorphic metric subquotients. These classes have the structure of an abelian group called the Witt group of metric groups. The class of \(\left( A,q\right) \) is equal to the class \(\left( \mathbb{F}_{p^{2}},i\circ\text{Nm}_{\mathbb{F}_{p^{2} }/\mathbb{F}_{p}}\right) ,\) where \(i:\mathbb{F}_{p}\rightarrow\mathbb{Q} _{p}/\mathbb{Z}_{p}\) is given by \(i\left( 1\right) =1/p.\) As an application it is shown that a category of certain perverse objects is equivalent to the modular category corresponding to the metric group \(A\). The proof of this result is accomplished by first considering the case of the perfectization of the additive group, i.e. \(\mathbb{G}_{a,\text{p}} =\)Spec\(\left( k\left[ x,x^{1/p},x^{1/p^{2}},\dots\right] \right) .\) The homomorphisms \(\mathbb{G}_{a,\text{p}}\rightarrow\mathbb{G}_{a,\text{p}} ^{\ast}\) are made explicit -- indeed, by identifying \(\mathbb{G}_{a,\text{p} }^{\ast}\) with \(\mathbb{G}_{a,\text{p}}\) the skew-symmetric homomorphisms can be viewed as elements of \(k\left\{ \tau,\tau^{-1}\right\} \) where \(\tau\) is the Frobenius map..These are the maps which correspond to biextensions which are skew-symmetric. A result of Drinfeld, Gelakia, Nikshych, and Ostrik completes the result in this special case. For the more general version with \(\dim G>1\) one reduces to \(\mathbb{G}_{a,\text{p}}\) inductively by considering closed subgroups of \(G\) isomorphic to \(\mathbb{G}_{a,\text{p}}.\)