Skip to main content
12 events
when toggle format what by license comment
Aug 3, 2017 at 15:49 comment added Ariyan Javanpeykar @PiotrAchinger Let Et be the stack of finite etale morphisms. (So objects are $(S,f:X\to S)$ with $S$ a scheme and $f$ finite etale.) This is an algebraic stack locally of finite type. Let Et$_d$ be the substack of finite etale morphisms of degree $d$. This is of finite type. You are right that $\mathrm{Hom}(\mathbb A^1, \mathrm{Et})$ is not representable by an algebraic stack. (Your argument shows that Artin's axioms fail. Morally, this is because $\mathbb A^1$ is not proper.)
Jun 26, 2017 at 16:15 vote accept John Z.
Jun 25, 2017 at 22:10 history edited YCor CC BY-SA 3.0
fixed typo in title
Jun 25, 2017 at 21:02 comment added Piotr Achinger @AriyanJavanpeykar The situation in characteristic $p$ is quite perplexing: any $f\in A$ gives rise to the Artin-Schreier $\mathbb{Z}/p$-covering $t^p - t = f$ of $X={\rm Spec} A$, and $f$ and $g$ yield the same covering iff $f-g = h^p - h$ for some $h\in A$. In particular, unlike in char. 0, one can easily have "families" of coverings, e.g. $t^p-t=xy$, as a family of coverings of $\mathbb{A}^1$ parametrized by $\mathbb{A}^1$. On the other hand, by the unique lifting property of etale maps, every such family is formally constant. This seems to mean that there is no reasonable moduli problem.
Jun 25, 2017 at 20:55 answer added Piotr Achinger timeline score: 7
Jun 25, 2017 at 20:49 comment added Ariyan Javanpeykar @PiotrAchinger You're right. What I wrote only works when the scheme is generically proper over $A$. My bad.
Jun 25, 2017 at 20:44 comment added Piotr Achinger @AriyanJavanpeykar Why is the stack of finite type? Here $X$ is affine, and if we are in characteristic $p$, we probably do not even have such a stack, and even if we did, the moduli problem would be unbounded...
Jun 25, 2017 at 20:43 comment added John Z. Yes, I do assume that Galois coverings are finite etale coverings, and variety means integral scheme of finite type.
Jun 25, 2017 at 20:18 comment added Ariyan Javanpeykar ....In particular, any $k$-point of characteristic $p>0$ (with $k$ the residue field of a maximal ideal of $A$) lifts to characteristic zero (and thus to a $\mathbb C$-point). So, if you have a finite etale cover of $X_k$, it will "come" from a finite etale cover of $X$ (via specialization).
Jun 25, 2017 at 20:17 comment added Ariyan Javanpeykar When you say "variety" over $S$, do you mean "scheme"? Also, is your Galois covering (finite) etale? If yes, for $d\geq 1$, let $M_d$ be the stack of finite etale covers of $X$ over $A$. This is a finite type stack over $A$. Like "everything" over $A$, it is generically flat over the function field $K(A)$ of $A$. Therefore, by spreading out (see Rydh's appendix in arxiv.org/abs/0904.0227), replacing Spec $A$ by a dense open if necessary, we may and do assume that $M_d$ is flat over $A$....
Jun 25, 2017 at 19:29 review First posts
Jun 25, 2017 at 19:31
Jun 25, 2017 at 19:26 history asked John Z. CC BY-SA 3.0