1
$\begingroup$

Let $F$ and $E$ be number fields, $G_F$ be the absolute Galois group of $F$, and $S$ be a finite set of primes of $F$. For $\lambda$ a prime of $E$ we denote by $\ell$ its residual characteristic. We also denote the set of primes of $F$ over $\ell$ by $S_{\ell}$.

A compatible system of continuous semisimple $\lambda$-adic representations $\rho=\{\rho_{\lambda}:G_F \to \mathrm{GL}_n(\overline{E_{\lambda}})\}_{\lambda}$ satisfies (by definition) the following properties:

(1) If $v \not\in S \cup S_{\ell}$, then $\rho_{\lambda}$ is rational and unramified at $v$.

(2) If $v \not\in S \cup S_{\ell} \cup S_{\ell'}$, then the characteristic polynomials of the image of $\mathrm{Frob}_v$ under $\rho_{\lambda}$ and $\rho_{\lambda'}$ are the same.

In p.I.16 of Serre's book, 'Abelian $\ell$-adic representations and elliptic curves', it is defined the $L$-function attached to this compatible system by

$$ L(s,\rho):=\prod_{v \not\in S} \frac{1}{\det(1-\rho(\mathrm{Frob}_v)N(v)^{-s})}. $$

Here, the denominator can be taken independently for infinitely many $\lambda$ (precisely, $\lambda$ with $\ell \neq p_v$).

I would like to know whether there are some results related to the analytic properties of this $L$-function, such as the holomorphy or existence of poles(of analytic continuation, if exists), absolute convergence of the formal Dirichlet series on the right half plane, convexity bounds, and similar properties concerning $L(s,\rho \times \rho)$.

If we consider the Artin representation instead of a compatible system of $\lambda$-adic representations, then Artin conjecture states that the associated $L$-function is entire. However, the Artin representation is a complex representation of $G(M/F)$ of finite extension $M$ of $F$. I wonder if is there a similar conjecture or theorem about the $\lambda$-adic representations.

$\endgroup$
4
  • $\begingroup$ I found the reciprocity conjecture in the Langlands program predicts the meromorphic continuation, ensuring that the system is weakly compatible so that each of $\rho_{\lambda}$ is algebraic. The potential automorphy and the automorphy are proven only for a few cases, so the problem I asked seems to be widely open. Once if we have automorphy or potential automorphy, then we can think of automorphic $L$-function instead of $L$-function of Galois representation. If there are strange points in my comment, please feel free to point out. $\endgroup$
    – LWW
    Commented Jan 24 at 8:58
  • $\begingroup$ There are many results. Are you asking about analytic results in cases where we don't know (potential) automorphy? $\endgroup$
    – Kimball
    Commented Jan 24 at 13:49
  • $\begingroup$ @Kimball Yes, now the question can be understood in that direction (when I asked the question, I didn't know that this question is related to the potential automorphy). $\endgroup$
    – LWW
    Commented Jan 25 at 3:26
  • 1
    $\begingroup$ In this generality it is not even known that $L(s, \rho)$ converges for any $s \in \mathbf{C}$, so it might not even make sense as a function at all. $\endgroup$ Commented Jan 25 at 6:38

0

You must log in to answer this question.