Let $\pi:X\to\mathcal X$ be a presentation of an Artin stack $\mathcal X$ of finite type over a field $k,$ and let $f:Y\to X$ be a finite \'etale covering. Does there exist a finite \'etale covering $Y'\to X$ factoring through $Y,$ such that $Y'$ can be given descent structure, i.e. there exists an isomorphism $pr_1^*Y'\cong pr_2^*Y'$ over $X\times_{\mathcal X}X$ satisfying cocycle condition, so that $Y'\to X$ descend to a finite \'etale covering $\mathcal Y\to\mathcal X$?
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
I don't think so (finite etale covers cannot be localized in smooth topology in the sense that you describe). Say, $\mathcal{X}$ is a point, and $X$ is a smooth variety with non-trivial fundamental group, say, an elliptic curve (or $\mathbb{A}^1-\{0\}$). Then $\pi$ is a presentation. Let $f:Y\to X$ be a non-trivial finite etale cover, say, the cover of the elliptic curve by an isogeneous elliptic curve. Then your question becomes: `is there a trivial (i.e., lifted from $\mathcal{X}$) cover $Y'$ of $X$ with a map to $Y$? This is of course not true.
-
$\begingroup$ Thanks. I wanted to know if any locally constant constructible sheaf on the stack $\mathcal X$ can be trivialized by a finite etale cover (which is true for schemes). That's why I asked that descent question. I was told that the answer to this lcc sheaf question is Yes, because one can take the espace etale of the sheaf... $\endgroup$– shenghaoCommented Dec 19, 2009 at 6:22
-
$\begingroup$ I see. Well, the original question is curious by itself, while, as you explain, the answer to the lcc sheaf question is essentially tautological. $\endgroup$– t3sujiCommented Dec 19, 2009 at 14:26