I've been tasked with the following:
Let $\zeta_8=\exp(2\pi i/8)$. Compute the group of field automorphisms of $\mathbb{Q}(\zeta_8)$.
This seems easy enough, I assume that the problem reduces to showing that automorphisms of $\mathbb{Q}(\zeta_8)$ simply permute the roots of the polynomial $x^7+x^6+\cdots +x+1$, and hence it'll probably be isomorphic to $S_7$ or something similar.
However I've hit a snag: from my knowledge, we only know that a field automorphism $\sigma$ of $\mathbb{Q}(\zeta_8)$ permutes the roots of the cyclotomic polynomial if $\sigma$ fixes $\mathbb{Q}$. I don't quite know how to show this though. We've shown that the only automorphism of $\mathbb{Q}$ is the trivial automorphism, so I assume that ties in somewhere, but I can't see where
Any help is truly appreciated. If you can't tell, this is a homework exercises. So just hints if possible. Thanks in advance.
Edit: is it because we know that $\sigma(\mathbb{Q})\cong\mathbb{Q}/\ker\sigma=\mathbb{Q}$ for each $\sigma\in\text{Aut}\left(\mathbb{Q}(\zeta_8)\right)$? In this case how do we know that $\mathbb{Q}$ doesn't map onto an isomorphic copy of $\mathbb{Q}$ distinct from $\mathbb{Q}$?
If $K$ is any field containing $\mathbb{Q}$ and $f:K\to K$ is any automorphism, then $f(q)=q$ for all $q\in\mathbb{Q}$. First, $f(0)=0$ and $f(1)=1$ by definition of a homomorphism. It follows that $f(2)=f(1)+f(1)=1+1=2$, $f(3)=f(2)+f(1)=2+1=3$, and so on, so $f(n)=n$ for all $n\in\mathbb{N}$. Since $f$ preserves additive inverses it follows that $f(n)=n$ for all $n\in\mathbb{Z}$, and since $f$ preserves multiplicative inverses $f(1/n)=1/n$ for all nonzero $n\in\mathbb{Z}$. Thus if $n/m\in\mathbb{Q}$, $f(n/m)=f(n)f(1/m)=n/m$.