Let an algebraic group $G$ act on a complex variety $X$ such that there is a good enough quotient $X/G$ (for example, $G$ acts on a vector space $V$ linearly and $X=V_{ss}$ is a variety of semi-stable points). Let $E$ be a vector bundle on $X$ with a $G$-action commuting with the $G$-action on $X$. I believe that if $G$ acts effectively then there is a quotient vector bundle $E/G$ on $X/G$. Is it true? Could you give me references? Also, does all line bundles on $X/G$ arise this way?
$\begingroup$
$\endgroup$
5
-
8$\begingroup$ No, this is false. You need that the stabilizer of a closed point $x$ acts trivially on $E_x$. For a precise statement, see §2 of Drézet-Narasimhan Groupe de Picard des variétés de modules de faisceaux semi-stables sur les courbes algébriques, Invent. Math. 97 (1989), 53-94. $\endgroup$– abxCommented Oct 2, 2016 at 12:52
-
$\begingroup$ @abx Seems like you could just add that as an answer. $\endgroup$– Ben Webster ♦Commented Oct 2, 2016 at 13:24
-
$\begingroup$ @abx, thank you, and add this as an answer, please. $\endgroup$– evgenyCommented Oct 2, 2016 at 13:44
-
$\begingroup$ related : mathoverflow.net/questions/194955/line-bundle-descends, where @abx gives the same answer $\endgroup$– NielsCommented Oct 2, 2016 at 20:22
-
$\begingroup$ Related: MSE:1990550, there is a link to an article by Knop, Kraft, Vust with very clear explanation. $\endgroup$– evgenyCommented Oct 29, 2016 at 20:07
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
I put my comment as an answer: the necessary and sufficient condition for $E$ to be the pull back of a vector bundle on $X/G$ is that the stabilizer of any closed point $x$ with a closed orbit acts trivially on $E$. This is a lemma of Kempf, well explained in §2 of Drézet-Narasimhan Groupe de Picard des variétés de modules de faisceaux semi-stables sur les courbes algébriques, Invent. Math. 97 (1989), 53-94.