1
$\begingroup$

Can one characterize the elements $\alpha$ of $L=\overline{\mathbb F_q(T)}\subset\Omega$, a completion of $L$ for the $\left(\frac1T\right)$-adic valuation such that for every conjugate $\beta$ of $\alpha$, there exists a continuous $\mathbb F_q\left(\left(\frac1T\right)\right)$-isomorphism of $\Omega$ with $\sigma(\alpha)=\beta$.

$\endgroup$
0

2 Answers 2

3
$\begingroup$

Let $\Omega'$ be an algebraic closure of $\mathbb F_q (( \frac{1}{T}))$. Then we can embed $\Omega'$ into $\Omega$ because every extension of $\mathbb F_q (( \frac{1}{T}))$ is defined over $\mathbb F_q(T)$, and $\Omega'$ is dense in $\Omega$ because it contains $L$, so therefore $\Omega$ is the completion of $\Omega'$. Thus every continuous automorphism of $\Omega'$ extends to a continuous automorphism of its completion $\Omega$.

Thus it's equivalent to ask that every Galois conjugate of $\alpha$ is conjugate, over $\mathbb F_q (( \frac{1}{T}))$, in $\Omega'$, to $\alpha$. Since $\Omega'$ is an algebraic closure of $\mathbb F_q (( \frac{1}{T}))$, it's equivalent to ask that the minimal polynomial of $\alpha$ over $\mathbb F_q(T)$ remains irreducible over $\mathbb F_q((1/T))$.

There are many criteria for when a polynomial over a local field is / isn't irreducible, for instance in terms of the Newton polygon, so I think this is the best way to express it.

$\endgroup$
0
$\begingroup$

Please define correctly your symbols, you meant $L$ is the algebraic closure of $\Bbb{F}_q(T)$ and $\Omega$ is the completion of $L$ for a discrete valuation above $T$.

Take $f\in \Bbb{F}_q[T][X]$ monic irreducible such that $f\bmod T\in \Bbb{F}_q[X]$ is separable and it splits completely in $\Bbb{F}_p$ (for example $q=p=3,f=X^2-1+T$) then $f$ splits completely in $\Bbb{F}_q[[T]]\subset \Omega$ and its roots are all $\Bbb{F}_q(T)$-conjugates, of course there is no $\sigma\in Aut(\Omega/\Bbb{F}_q((T)))$ acting non-trivially on those roots.

A continuous automorphism of $\Omega$ restricts to an automorphism of $O_{\Omega} =\{ a\in \Omega,v(a)\ge 0\}$ sending the maximal $\mathfrak{m} = \{ a\in \Omega,v(a)> 0\}$ to itself, thus it gives an automorphism of $O_\Omega/\mathfrak{m} \cong \overline{\Bbb{F}}_q$ and since it acts trivally on $\Bbb{F}_p$ it acts trivially on the reductions on the roots thus on their unique lift to $\Bbb{F}_q[[T]]$.

$\endgroup$

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.