1
$\begingroup$

Suppose $X / K$ is a variety over a finitely generated field over $\mathbb{Q}$. Fix an embedding $K \subset \mathbb{C}$ and let $\pi := \pi_1(X(\mathbb{C}), x)$ be the topological fundamental group. Spreading out to $A \subset K$ we can get specialization maps

$$ \pi \to \hat{\pi} = \pi_1^{\mathrm{et}}(X_{\mathbb{C}}, \bar{x}) \to \pi_1^t(X_{\bar{s}}, \bar{x}) \subset \pi_1^t(X_{s}, \bar{x}) $$

where $s \in \mathrm{Spec}(A)$ is a point of characteristic $p$. Suppose for $\ell \neq p$ I have an $\ell$-adic arithmetic representation:

$$ \rho : \pi_1^t(X_{\bar{s}}, \bar{x}) \to \mathrm{SL}_r(\overline{\mathbb{Q}}_\ell) $$

which for simplicity I have assumed is actually a representation of the full arithmetic fundamental group of $X_s$. Then for any $\ell' \neq \ell, p$ and a field isomorphism $\sigma : \overline{\mathbb{Q}}_{\ell} \to \overline{\mathbb{Q}}_{\ell'}$ there is a $\sigma$-companion: a continuous arithmetic $\ell'$-adic representation

$$ \rho^{\sigma} : \pi_1^t(X_s, \bar{s}) \to \mathrm{SL}_r(\overline{\mathbb{Q}}_{\ell'}) $$

compatible in the sense that the Frobenius characteristic polynomials are transported by $\sigma$.

My question is: if we compose with the specialization to get $\rho|_{\pi}$ and $\rho^{\sigma}|_{\pi}$ is there anything we can say about how these representations are related. Because $\pi$ is dense (in the geometric $\pi_1$), it is clear that $\rho|_{\pi}$ uniquely determines $\rho^{\sigma}|_{\pi}$ (up to conjugation since the extension to the arithmetic group is unique up to conjugation) but this is the extent of my knowledge.

A naive question is if the characteristic polynomials of $\rho|_{\pi}$ and $\rho^{\sigma}|_{\pi}$ are also transported by $\sigma$? This seems unlikely but I don't see a counterexample.

$\endgroup$
2
  • $\begingroup$ If $\rho$ comes from a motive then the companions will all come from the motive and hence be isomorphic representations. Conjecturally $\rho$ comes from a motive(up to some twisting if $\rho$ is not geometrically irreducible) but there's no proof. $\endgroup$
    – Will Sawin
    Commented Jul 15 at 18:07
  • $\begingroup$ @WillSawin, so suppose $\rho$ arises from some smooth proper morphism $f : Y \to X_{\bar{s}}$ then if I interpret what you are saying correctly, it is that in this case, my claim about the characteristic polynomials of $\rho|_{\pi}$ and $\rho^{\sigma}|_{\pi}$ is true. I don't see how to prove it, could you point me in the right direction? Thank you. Also could you elaborate on what sort of twist I need to do if $\rho$ is not irreducible? $\endgroup$
    – Ben C
    Commented Jul 15 at 19:58

0

You must log in to answer this question.