1
$\begingroup$

What does $\mathrm{Gal}(\overline{\mathbb{Q}}_{p}/\mathbb{Q})$ mean? ($p$ is a prime number.)

If it is defined as $\mathrm{Aut}(\overline{\mathbb{Q}}_{p}/\mathbb{Q})$, then does it have any property like usual galois groups?

For example, is the following statement true? :

Let $\alpha \in \overline{\mathbb{Q}}_{p}$, and assume that for any $\sigma \in \mathrm{Aut} (\overline{ \mathbb{Q} } _{p} / \mathbb{Q})$, $\sigma (\alpha) = \alpha$. Then, $\alpha \in \mathbb{Q}$.

If this is true, then how can I prove it?

Please give me any advice.

Thanks!

$\endgroup$
5
  • $\begingroup$ Please read mathoverflow.net/faq#whatnot . $\overline{\mathbb{Q}_p} / \mathbb{Q}_p$ is an infinite Galois extension. $\endgroup$ Nov 13, 2011 at 12:21
  • $\begingroup$ (but over $\mathbb{Q}$ it's not Galois) $\endgroup$ Nov 13, 2011 at 12:22
  • 1
    $\begingroup$ @Hiro: One can use the axiom of choice to construct an isomorphism of abstract fields (i.e. ignoring the topology) $\overline{\mathbb{Q}_p} \cong \overline{\mathbb{Q}}$. $\endgroup$ Nov 13, 2011 at 12:32
  • 9
    $\begingroup$ @Daniel: No, one cannot. $\overline{\mathbb{Q}}$ is countable, $\overline{\mathbb{Q}_p}$ is not. $\endgroup$ Nov 13, 2011 at 12:48
  • $\begingroup$ @Robert: Yes of course I was being completely stupid, thanks. $\endgroup$ Nov 13, 2011 at 14:56

1 Answer 1

6
$\begingroup$

Yes, this is true. A way to prove this is to use the existence of transcendence basis for field extensions, and the fact that $\overline{\mathbf{Q}_p}$ is algebraically closed.

First, assume $\alpha$ is transcendental. Then there exists a transcendence basis $S$ of $\overline{\mathbf{Q}_p}/\mathbf{Q}$ containing $\alpha$. By permuting the elements of $S$, there exists an automorphism $\sigma$ of $\mathbf{Q}(S)$ such that $\sigma(\alpha) \neq \alpha$. Since $\overline{\mathbf{Q}_p}$ is an algebraic closure of $\mathbf{Q}(S)$, you can extend $\sigma$ to an automorphism of $\overline{\mathbf{Q}_p}$, which gives a contradiction.

Finally, if $\alpha \in \overline{\mathbf{Q}}$ but $\alpha \not\in \mathbf{Q}$, then you can take the Galois closure $L$ of $\mathbf{Q}(\alpha)$ and find an automorphism $\sigma$ of $L$ such that $\sigma(\alpha) \neq \alpha$. Then, repeating the above argument, you can extend $\sigma$ to $\overline{\mathbf{Q}_p}$, which gives a contradiction.

Note that the argument works replacing $\overline{\mathbf{Q}_p}$ by any algebraically closed field of characteristic $0$.

$\endgroup$
1
  • $\begingroup$ Oh, I did not think that it is such a general fact! And your proof is clear and interesting! Thank you so much! $\endgroup$
    – Hiro
    Nov 13, 2011 at 13:06

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