Let $\Sigma$ be an $n$-dimensional representation of the global Weil group $W_F$ for a number field $F$, and $\Pi$ an automorphic representation of $\operatorname{GL}_n(\mathbb A_F)$. Suppose that at every place $v$ of $F$ where $\Sigma_v$ and $\Pi_v$ are both unramified, we have $\Pi_v \leftrightarrow \Sigma_v$ under the local Langlands correspondence.

In Guy Henniart's proof on the local Langlands correspondence, he claims that $\operatorname{Det} \Sigma$ then corresponds to the central character of $\Pi$ under the isomorphism $W_F^{\operatorname{ab}} \cong \mathbb A_F^{\ast}/F^{\ast}$ of global class field theory.

enter image description here

Why should this be the case? It is true that at almost all places $v$, $\operatorname{Det} \Sigma_v$ identifies with the central character of $\Pi_v$ under the isomorphism $F_v^{\ast} \cong W_{F_v}^{\operatorname{ab}}$. We have for $a = (a_v) \in \mathbb A_F^{\ast}$,

$$\varpi_{\Pi}(a) = \prod\limits_v \varpi_{\Pi_v}(a_v)$$

so it seems like something could go wrong with $\varpi_{\Pi}$ at the ramified places.


1 Answer 1


Now that I have thought about it a little more, I think this just comes down to the following fact: If $\chi = \otimes \chi_v$ is a character of $\mathbb A_F^{\ast}/F^{\ast}$, and $S$ is a finite set of places of $F$, then $\chi$ is completely determined by the characters $\chi_v : v \not\in S$. This is because if $\eta = \otimes_v \eta_v$ is another character of $\mathbb A_F^{\ast}/F^{\ast}$ with $\eta_v = \chi_v$ for all $v \not\in S$, then $\chi \eta^{-1}$ defines a character of $\prod\limits_{v \in S} k_v^{\ast}$ which is trivial on the diagonal embedding $\Delta$ of $k^{\ast}$. But $\Delta$ is dense in this product by weak approximation, so in fact $\chi = \eta$ on $\mathbb A_F^{\ast}$.

Now $\operatorname{Det} \circ \Sigma$ corresponds to a character of $\mathbb A_F^{\ast}/F^{\ast}$ which agrees at almost all places with the central character of $\Pi$ (also trivial on $F^{\ast}$), so they must be the same character.


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.