Sign up ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $\textbf{Ell}$ be the category of elliptic curves over various base schemes, and where a morphism between $E\rightarrow S$ and $E'\rightarrow S'$ is a cartesian diagram with those two maps as columns.

A moduli problem for elliptic curves is then just a contravariant functor $\textbf{Ell}\rightarrow\textbf{Sets}$.

For example, we usual level $N$ moduli problem is the functor sending $E/S$ to the set of isomorphisms $E[N]\stackrel{\sim}{\longrightarrow}(\mathbb{Z}/N\mathbb{Z})_S^2$.

There are a ton of these functors, mostly coming from various cohomology theories, but the only such functors I can think of that land in the category of finite sets all have to do with torsion points on the elliptic curve.

Does anyone have any examples of a contravariant functor $F:\textbf{Ell}\rightarrow\textbf{Sets}$ such that for $E/S$ with $S$ connected, $F(E/S)$ is finite, and doesn't have to do with torsion data?

Ideally, the functor will actually land in the category of groups, be generically of some order $M\ge 3$, and always be of order $\le M$.

share|cite|improve this question
Aside from the fact that the phrase "doesn't have to do with torsion data" is vague, if you consider a $\mathbf{Z}[1/N]$-schemes $S$ that is sufficiently disconnected then the set of level-$N$ structures on a fixed elliptic curve $E$ over $S$ can be arbitrarily large and in particular not finite (akin to global sections of a constant sheaf on a disconnected space). So do you mean just that for elliptic curves over algebraically closed fields the associated set should be finite? If so, then how about assigning to any $E \rightarrow S$ the automorphism group? – user28172 May 7 '13 at 19:52

1 Answer 1

up vote 11 down vote accepted

Sure -- try the set of homomorphisms from $\pi_1^{\mathrm{et}}(E - O)$ to a fixed finite group G. This is a "non-abelian level structure" of the sort considered by de Jong and Pikaart

Rachel Davis, a 2013 Wisconsin Ph.D. working with Nigel Boston, wrote her thesis about this kind of stuff in the case of elliptic curves.

share|cite|improve this answer
Hmm, so in that arxiv paper, they're using some symbols I'm not familiar with. In particular, in the first paragraph, what does $(\mathbb{Z}/\ltimes\mathbb{Z})^{\nvDash\partial}$, and $\text{Spec}(\mathbb{Z}[\nVdash/\#\mathbb{G}])$ mean? – William Chen May 8 '13 at 0:25
Also $\text{Spec}(\mathbb{Z}[\nVdash/\ltimes])$ in the statement of theorem 3.1.1 on the first page. – William Chen May 8 '13 at 0:27
Either that's a bizarre TeX error, or else I have a new entry for the question… – Marty May 8 '13 at 0:50
If you look at the source, you'll see that these are $(\mathbb{Z}/n\mathbb{Z})^{2g}$, Spec $\mathbb{Z}[1/\# G]$, and Spec $\mathbb{Z}[1/n]$, respectively. – S. Carnahan May 8 '13 at 1:27
Yes, in some sense it explains them all! This is a result of myself and McReynolds very much inspired by the old paper of Diaz, Donagi, and Harbater, "Every curve is a Hurwitz curve." – JSE May 9 '13 at 1:55

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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