6
$\begingroup$

Let $\mathbb{F}_p((x))$ be the local field of formal laurent series over the finite field $\mathbb{F}_p$ considered as a locally compact field.

What are the additive continuous characters $\mathbb{F}_p((x)) \to \mathbb{C}^{\times}$?

What are the multiplicative characters $\mathbb{F}_p((x))^{\times} \to \mathbb{C}^{\times}$?

I know how characters of $\mathbb{Q}_p$ are all built out of taking the sum of the fractional part of the $p$-adic expansion which is well defined upto an integer and I understand why it's continuous. For $\mathbb{F}_p((t))$ however I have no idea where to start. The presence of the finite field confuses me when I try to define a homomorphism.

I also know that $\mathbb{F}_p((x))^{\times}=\mathbb{F}_p[[x]]^{\times}\times x^\mathbb{Z}$. So it will presumably be enough to understand additive characters of $\mathbb{F}_p[[x]]$. Unfortunately here i'm stuck with the same problem as before...

$\endgroup$
  • $\begingroup$ What are the additive and multiplicative characters of $\mathbb{F}_p[x,x^{-1}]$, which ones are continuous for $|u| = 2^{-ord(u)}$ ? $\endgroup$ – reuns Mar 14 '17 at 22:46
  • 4
    $\begingroup$ A nice question, but: What do you know? What have you found out so far? Do you know of a single example of each kind? When you post nothing but a question, those who may be in a position to answer it don’t know what expositional level to use. $\endgroup$ – Lubin Mar 15 '17 at 3:20
  • $\begingroup$ @Lubin Added some of my thoughts in the direction. $\endgroup$ – Saal Hardali Mar 15 '17 at 19:07
4
$\begingroup$

This is a subject that I fear that I have not paid as much attention to as I should have. If any expert would like to chime in and accuse me of vapidity (or error) in what I say below, I will step back.

Let’s call our field $\Bbb F_p((x))=k$ and its ring of integers $\Bbb F_p[[t]]=\mathfrak o$. First, for the additive characters, note that, when we look modulo a basic neighborhood of $0$, i.e. $t^m\mathfrak o$, we have $$ X_m=k/t^m\mathfrak o\cong\bigoplus_{j<m}\Bbb F_p\,, $$ an infinite direct sum. This, of course, is radically different from the situation for $\Bbb Q_p$. I guess that (since everything goes into the $p$-th roots of unity in $\Bbb C^\times)$ the set of characters of $X_m$ is $Y_m=\prod_{j<m}C_p$. where by $C_p$ I just mean a cyclic group of order $p$. Now, since $k^+$ is the inverse limit of all the $X_m$, I guess what you want is the direct limit of the $Y_m$’s, but I have no feel for what this is. That’s all I can say about $k^+$.

For the multiplicative characters, the story is far more interesting. You have correctly identified $k^\times=x^{\Bbb Z}\oplus\mathfrak o^\times$, but $\mathfrak o^\times$ is equal, in turn, to $\Bbb F_p^\times\oplus(1+x\mathfrak o)$. What’s inside the parentheses here is, explicitly, $1+x\Bbb F_p[[x]]$, the principal units of $\mathfrak o$. And the structure of this has nothing whatever to do with any additive structure. Indeed, in this case the principal units are the direct product (not sum) of copies of $\Bbb Z_p$, each rank-one factor being the span of (as one possible choice among many) $1+x^n$, where the $n$ here run through all the integers prime to $p$. Explicitly, then, $$ 1+x\mathfrak o=\prod_{n:(n,p)=1}(1+x^n)^{\Bbb Z_p}\,. $$ The reason for this is that you can raise any series in $1+x\mathfrak o$ to any $p$-adic integer, ’cause the powers $(1+x)^{p^m}$ go to $1$ as $m\to\infty$. Now, for any one factor here, the continuous characters of $(1+x^n)^{\Bbb Z_p}$ into $\Bbb C^\times$ are the (continuous) characters of $\Bbb Z_p$ into $\Bbb C^\times$, and these are just the group of torsion elements of $\Bbb C^\times$ of $p$-power order. You may recognize this group as $\Bbb Q_p/\Bbb Z_p$. So it seems to me that the continuous characters of $1+x\mathfrak o$ into $\Bbb C^\times$ are $$ \bigoplus_{n:(n,p)=1}\Bbb Q_p/\Bbb Z_p\,. $$ I just hope I haven’t missed any subtleties, and I hope this will prove to be of some help.

$\endgroup$

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.