4
$\begingroup$

As far as I understand it, by the work of Lafforgue (cf. Laumon, "Cohom. of Drinfeld ... II", Thm 12.4.1) there is a Galois representation associated to an irreducible cuspidal automorphic representation $\pi$ , if $\pi$ is Steinberg at some place $\infty$.

Do we expect Galois representations also if some of these conditions do not hold?

I might be totally off here, but R. Taylor constructed in his thesis (using results from Brylinki-Labesse) Galois rep's to Hilbert modular forms by congruence methods.

Has this be studied anywhere for function fields?

Any hint, where such things are discussed, would be greatly appreciated!

$\endgroup$
1
  • $\begingroup$ I'm definitely no expert, and I may be wrong, but you might expect L-functions associated to irreducible cuspidal automorphic representations to correspond exactly to primitive elements of the Selberg class. I don't know whether one can associate to any such primitive element a Galois representation though. $\endgroup$ Commented Oct 23, 2013 at 16:43

1 Answer 1

4
$\begingroup$

Let $X$ be a smooth projective geometrically irreducible curve over $\mathbb F_{q}$ a finite field and $F$ its global field. Let $\mathbf G/F$ be a split connected reductive group and $\widehat{\mathbf G}$ its Langlands dual. Let $\pi$ be an irreducible cuspidal automorphic representation of $\mathbf G(\mathbb A_{F})$. Then for all $\ell\nmid q$ there exists a $G_{F}$-representation $\sigma_{\pi,\ell}$ with values in $\widehat{\mathbf G}(\bar{\mathbb Q}_\ell)$ attached to $\pi$ in the usual sense. This result is essentially optimal, so that the answer to the question in the title seems to be "all of them".

This is a result of Laurent Lafforgue (Invent. Math. 147) when $\mathbf G=GL_{n}$ and of Vincent Lafforgue (preprint available from his webpage) in the other cases.

That said, these results are far beyond my own expertise, so I hope real experts will chime in. Regarding your second question, there exists unpublished work on congruences for Drinfeld modular varieties, but my impression is that not much work has been done in this direction.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .