A generalization of the Oort conjecture

From MaRDI portal
Publication:2403294



Abstract: The Oort conjecture (now a theorem of Obus-Wewers and Pop) states that if k is an algebraically closed field of characteristic p, then any cyclic branched cover of smooth projective k-curves lifts to characteristic zero. This is equivalent to the local Oort conjecture, which states that all cyclic extensions of kt lift to characteristic zero. We generalize the local Oort conjecture to the case of Galois extensions with cyclic p-Sylow subgroups, reduce the conjecture to a pure characteristic p statement, and prove it in several cases. In particular, we show that D_9 is a so-called local Oort group.


The Oort conjecture (now a theorem of Obus-Wewers and Pep) states that if \(k\) is an algebraically closed field of characteristic \(p\), then any cyclic branched cover of smooth projective curves lifts to characteristic zero. This is equivalent to the local Oort conjecture, which states that all cyclic extensions of \(kt\) lift to characteristic zero. More precisely, the local-lifting problem can be stated as follows: let \(k\) be an algebrically closed field of characteristic \(p\), \(\Gamma\) a finite group and \(kz/ks\) a \(\Gamma\)-Galois extension. Is it true that there exists a DVR \(R\) aof characteristic zero with residue field \(k\) and a \(\Gamma\)-Galois extension \(RZ/RS\) that reduces to \(kz/ks\)? For cyclic groups \(\Gamma\) the local-lifting problem, and hence the similar global-lifting problem, has a positive answer. For other groups it is convenient to look at the so-called KGB obstruction, which vanishes if the desired lift exists (roughly speaking, the KGB obstruction vanishes if, for a local \(\Gamma\)-extension, there is a \(\Gamma\)-extension of a certain chareacteric zero power series ring for which the different behaves in the same way as in the local \(\Gamma\)-extension) The paper adresses this question for groups \(\Gamma\) of the shape \(P \rtimes {\mathbb Z}/m{\mathbb Z}\), where \(P\) is a \(p\)-group and \(p\nmid m\). The main theorem gives a positive answer to this question under a number of hypotheses, namely: (i) \(\Gamma\) has no KGB obstruction; (ii) the upper ramification breaks of \(P\) satisfy a certain inequality; (iii) an ``isolated differential data criterion, which can be expressed also algebraically in terms of the invertibility of a ``Vandermode-like matrix, is satisfied. There is also a conjecture, motivated by the known results in the subject, which states that for local \(\Gamma\)-extensions where \(\Gamma\) has the form \({\mathbb Z}/p^n{\mathbb Z}\rtimes {\mathbb Z}/m{\mathbb Z}\), the KGB obstruction is the only obstruction to lifting.











This page was built for publication: A generalization of the Oort conjecture

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2403294)