Let $C$ be a 1-dimensional complex manifold whose universal covering is provided by the half-plane $\mathcal{H}=\{z \in \mathbb{C} \mid \operatorname{Im}z>0\}$. The symmetric product $C^{(n)} = C^n / S_n$ is a complex manifold. My question is: is the universal covering of $C^{(n)}$ provided by $\mathcal{H}^{(n)}$? Or is it some different quotient of $\mathcal{H}^n$?
-
2$\begingroup$ The projection from the symmetric power of the half-plane to the symmetric power of $C$ is not a covering space projection. $\endgroup$– Tom GoodwillieCommented Jul 7, 2023 at 11:23
-
$\begingroup$ If $C$ is $CP^1$, then the nth symmetric product is $CP^n$, which is simply connected, and so is its own universal cover. $\endgroup$– Nicholas KuhnCommented Jul 8, 2023 at 0:20
1 Answer
In fact, the universal cover of $C^{(n)}$ will not be $\mathcal H^n$ once $n \gg 0$. Indeed, if $C$ is a compact Riemann surface of genus $g \geq 2$ (so the universal cover is $\mathcal H$) with a base point $x$, then the Riemann–Roch theorem implies that the map \begin{align*} C^{(n)} &\to \operatorname{Jac}_C = \operatorname{Pic}^0_C\\ (x_1,\ldots,x_n) &\mapsto [x_1 + \ldots + x_n - nx] \end{align*} is a $\mathbf P^{n-g}$-bundle if $n > 2g-2$ (see e.g. [CS86, Ch. VII, Rmk. 5.6(c)]). So the universal cover of $C^{(n)}$ is a $\mathbf P^{n-g}$-bundle over $\mathbf C^g$, which is very far from $\mathcal H^n$ (e.g. it is not Kobayashi hyperbolic since $C^{(n)}$ is not Brody hyperbolic).
References.
[CS86] G. Cornell, J. H. Silverman (eds.), Arithmetic geometry. Springer-Verlag, 1986. ZBL0596.00007.
-
$\begingroup$ Thank you for the answer! I just edited my question. Also, I found the paper "Fixed point classes on symmetric product spaces" by Zhao which states that the universal covering of $C^{(n)}$ is given by the quotient of $\mathcal{H}^n$ by a normal subgroup of the semi-direct product $\pi_1(C)^n \rtimes \mathrm{S}_n$. $\endgroup$– KuSiCommented Jul 9, 2023 at 9:36
-
$\begingroup$ If you would like to ask a different question, you can open a new question. It is considered bad form on this site to change the question after an answer has been posted — it is impossible to hit a moving target. $\endgroup$ Commented Jul 9, 2023 at 19:21
-
$\begingroup$ You are absolutely right. I have opened a new question at mathoverflow.net/q/450455/505150. $\endgroup$– KuSiCommented Jul 9, 2023 at 20:45