Let $F$ be a $p$-adic local field with residue field $\kappa$, and $q = |\kappa|$. Then the residue field $\overline{\kappa}$ of $F^{\textrm{ur}}$ is an algebraic closure of $\kappa$. The local Weil group $W_F$ is the group of $\sigma \in \operatorname{Gal}(\overline{F}/F)$ such that $\bar{\sigma}(x) = x^{q^n}$ for some $n \in \mathbf Z$ and all $x \in \overline{\kappa}$. Let $I = \operatorname{Gal}(\overline{F}/F^{\textrm{ur}})$ be the inertia group, which is the kernel of $W_F \rightarrow \operatorname{Gal}(\overline{\kappa}/\kappa)$, and $\Phi \in W_F$ be a geometric Frobenius, which induces the automorphism $x \mapsto x^{q^{-1}}$ on $\overline{\kappa}$. Then $W_F$ is topologically the semidirect product of $I$ and $\langle \Phi \rangle \cong \mathbf Z$.

Here is what John Tate says in his Corvallis article on the Weil group:

enter image description here

I don't understand what is going on here. First, is Tate claiming that there is an identification with the underlying space of $\operatorname{Spec} A_n$ and the elements of $\Phi^n I$? Or is he saying that $W_F$ occurs as the $\mathbb Q$-rational points of that $\mathbb Q$-scheme? Do we lose the topology on $W_F$ when we view it as a group scheme?


The construction will produce $W_F$ both as the spectrum and the set of $\mathbb Q$-valued points - i.e. the set of points is $W_F$, and all the residue fields are $\mathbb Q$.

The topology on $W_F$ is visible in the first perspective, as the topology of the scheme but not directly in the second perspective.

It is easies to verify this facts first for $I$ (i.e. that $I$ is the spectrum of the ring of locally constant functions he constructs, and that all the residue fields are $\mathbb Q$) and then observe that they hold for $W_F$ as it is the infinite disjoint union of countably many copies of $I$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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