8
$\begingroup$

I've been wondering about the following "finiteness statement" concerning etale covers for a while.

Let $K$ be a field of characteristic zero, not necessarily algebraically closed. A variety over $K$ is a smooth quasi-projective geometrically connected scheme over $K$.

Let $X$ be a variety, and let $d$ be an integer.

Question. Are there only finitely many varieties $Y$ over $K$ such that there exists a finite etale morphism $Y\to X$ of degree $d$ (up to $K$-isomorphism)?

This is true if $K$ is algebraically closed. In fact, by a standard Lefschetz principle argument, we may and do assume $K$ is the field of complex numbers. Then, the statement follows from GAGA and the fact that the topological fundamental group of $X$ is finitely generated.

In general, there might be problems with twists, but I'm not sure.

Note that if $\dim X = 0$ then the only variety $Y$ satisfying the above hypotheses is $Y= $ Spec $K$, because we require our varieties to be geometrically connected (over $K$).

If it helps, in my set-up, we may assume $K$ is finitely generated over $\mathbf Q$, or that $K$ is even a number field.

$\endgroup$
3
  • 7
    $\begingroup$ There is indeed a problem with twists. For example, consider the degree $2$ covers of $\mathbb{A}^1 - \{0\}$ given by adjoining a square root of $cX$ to $K(X)$, where $X$ is the standard coordinate on $\mathbb{A}^1$ and $c \in K$. $\endgroup$
    – naf
    Sep 12, 2013 at 13:07
  • $\begingroup$ Is there also an example with $X$ a projective curve of genus at least two? The fact that $\mathbf G_m$ might fail to have this property could be related to the fact that its Euler characteristic is zero. $\endgroup$
    – Theaux G.
    Sep 12, 2013 at 17:19
  • 3
    $\begingroup$ Yes, there is. Take any étale double cover $\pi\colon Y \rightarrow X$ of smooth projective curves over a field $k$. For example, we may take $X$ to have genus $2$ and $Y$ to have genus $3$. Then $\pi$ is an unramified covering map, and $Y$ has a unique structure of $\mathbf{Z}/2\mathbf{Z}$-torsor over $X$ (with $\mathbf{Z}/2\mathbf{Z}$ just interchanging the sheets of the covering). We can twist $\pi \colon Y \rightarrow X$ by any Galois cohomology class $c \in \operatorname{H}^1(k,\mathbf{Z}/2\mathbf{Z})$, of which there are infinitely many if e.g. $k$ is a number field. $\endgroup$
    – R.P.
    Sep 12, 2013 at 23:11

1 Answer 1

7
$\begingroup$

Just to add one word to ulrich's answer: for a geometrically irreducible $K$-variety $X$, there is a short exact sequence of etale fundamental groups. $$ 0 \to \pi_1^{\text{et}}(X\otimes_K \overline{K}) \to \pi_1^{\text{et}}(X) \to \text{Gal}(\overline{K}/K) \to 0. $$ An étale degree $d$ cover is equivalent to a homomorphism $\phi:\pi_1^{\text{et}}(X)\to \mathfrak{S}_d$. This homomorphism gives an irreducible cover of $X\otimes_K \overline{K}$ if the image of $\pi_1^{\text{et}}(X\otimes_K \overline{K})$ is a transtive subgroup of $\mathfrak{S}_d$. But for a given homomorphism $\phi_{\overline{K}}:\pi_1^{\text{et}}(X\otimes_K \overline{K}) \to \mathfrak{S}_d$ with transitive image, there may be many extensions of this to a homomorphism $\phi$. That does not in itself prove that there are infinitely many étale covers of, say $\mathbb{G}_{m,K}$. But it does help organize the infinitely many covers that ulrich listed.

$\endgroup$
0

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.