1
$\begingroup$

If $$f:\mathscr{X} \to \mathscr{Y}$$ is a map between (possibly ineffective) orbifolds (in the sense of differentiable stacks, or orbifold groupoids), does it follow that $f$ induces a map between their underlying effective orbifolds? At first, I thought it should always be true, but now that I think about it, you might need $f$ to be open (in this case, I can prove it). Is this known? (by this I mean, is it known to always hold, even without this open assumption?) Note, this really has nothing to do with the differentiable structure, so you may as well ask this for proper etale topological stacks, in fact, I doubt properness plays a role.

$\endgroup$
3
  • $\begingroup$ Remind me the definition of an effective groupoid.... $\endgroup$
    – David Roberts
    Nov 13, 2010 at 23:55
  • $\begingroup$ @David: In terms of orbifold groupoids, by using the fact that it is etale and proper, you can show that for every point $x \in G_0$ in the groupoid, there exists a neighborhood $U$ on which $G_x$ acts, where $G_x$ is the stabilizing group (i.e. Hom_G(x,x)). (This is done by taking a local inverse of the source map that goes through $g$ and applying the target map). $G$ is effective if each of these actions are faithful. $\endgroup$ Nov 14, 2010 at 1:10
  • $\begingroup$ Even better, look at this other question I asked: mathoverflow.net/questions/39531/… $\endgroup$ Nov 14, 2010 at 1:50

1 Answer 1

2
$\begingroup$

Not true:
Take X to be a non-trivial Z/2-gerbe on S^2, take Y to be a faithful vector bundle over X, and take f to be the inclusion of the zero section.

The effective quotient of X is S^2, and it has no map back to X. In particular, it has no map to Y that's compatible with f.

$\endgroup$
1
  • $\begingroup$ Thanks Andre. In fact, I realized this is not true later, but forgot to answer my own question on here. I give some sufficient conditions for a map to descend however in the preprint you are reading. I call these conditions "etale invariance" $\endgroup$ Nov 23, 2010 at 23:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.