2
$\begingroup$

Consider characters of $PGL(2)$. The conductor of a local component $\pi_p$ of an automorphic representation is defined as the smallest index such that there is nontrivial fixed vector by the congruence subgroup $$K_0(p^r) = \left\{ M \in GL(2, \mathcal{O}_{F_p}) \ : \ M \equiv \left( \begin{array}{cc} \star & \star \\ & \star \end{array} \right) \mod p^r \right\}$$

In the case of a character of $GL(2)$ of the form $\chi_0 \circ \det$, where $\chi_0$ is a character on the base field $F_p$, is there a natural relation between this notion and the conductor of $\chi_0$, defined as the smallest index $r$ such that $\chi_0$ is trivial on $1 + p^r$?

$\endgroup$
1
  • 4
    $\begingroup$ I guess the $\mod p$ in the definition of $K_0(p^r)$ is meant to be $\mod{p^r}$, and that the $\star$s are meant to stand for elements of the ring of integers? (Otherwise $K_0(p^r)$ doesn't seem to be a group.) Then it seems that $\det(K_0(p^r))$ is the unit group of the ring of integers for all $r \ge 0$, so that, if $\chi_0 \circ \det$ is trivial on any $K_0(p^r)$, then it is trivial on all of them. $\endgroup$
    – LSpice
    Commented Feb 3, 2018 at 15:56

1 Answer 1

4
$\begingroup$

The lower-right corner is supposed to be a $1$, not a star (unless the central character is trivial, which it isn't in your case unless $\chi_0$ is quadratic).

Regardless, as LSpice notes, these representations do not have a conductor in that sense, unless the conductor of $\chi_0$ is $1$, because they are not invariant under any of these subgroups.

This is not so surprising as the theorem only says that generic representations have a conductor.

Maybe a better thing to try is to tensor your representation with the Steinberg representation, obtaining an infinitesimally equivalent representation which is generic, whose conductor is equal to $\min(c(\chi_0)^2, p)$.

$\endgroup$

You must log in to answer this question.

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