1
$\begingroup$

Suppose that $K/F$ is a normal extension of fields of finite degree and consider $G' \leq \mbox{Gal}(K,F)$. If $\mathcal{F} = K_{G'}$ is the fixed field of $G'$, then show that for any $\tau \in \mbox{Gal}(K,F)$, $K_{\tau G' \tau^{-1}} = \tau(\mathcal{F})$.

Here's my attempt. Suppose $a \in \mathcal{F}$ and $g' \in G'$. Then $(\tau g' \tau^{-1})(\tau a) = \tau (g' a) = \tau a$, since $g'$ fixes $a \in \mathcal{F}$. Hence, $\tau G' \tau^{-1}$ fixes $\tau(\mathcal{F})$. So if I can show that the group fixing $\tau(\mathcal{F})$ has order equal to the degree of $K$ over $\tau({\mathcal{F}})$, which is the same as the degree of $K$ over $\mathcal{F}$, and I then show that the degree of $K$ over $\mathcal{F}$ is the order of $G'$, then the proof is complete. If $K/F$ is a Galois extension, then all of the above holds, but I'm not sure that it holds in the case that $K/F$ is a normal extension of fields of finite degree.

$\endgroup$
1
$\begingroup$

Some of your notation is unusual: $G'$ usually denotes the derived group of $G$ while the fixed field of $H$ is usually denoted $K^H$ rather than $K_H$.

I'll use standard notation: let $G=\textrm{Aut}(K/F)$, $H$ be a subgroup of $G$ and $K^H$ be its fixed field. Let $\tau\in G$. Then $a\in K^{\tau H\tau^{-1}}$ iff $\tau\sigma\tau^{-1}(a)=a$ for all $\sigma\in H$ iff $\sigma\tau^{-1}(a)=\tau^{-1}(a)$ for all $\sigma\in H$ iff $\tau^{-1}(a)\in K^H$ iff $a\in\tau(K^H)$. Therefore $K^{\tau H\tau^{-1}}=\tau(K^H)$.

None of this depends on $K/F$ being a Galois extension. Here $G=\textrm{Aut}(K/F)$ is the group of $F$-automorphisms of $K$. The extension is Galois iff $F=K^G$, but that's not necessary for the argument above.

$\endgroup$
2
  • $\begingroup$ Does your proof make use of the assumption that $K/F$ is a normal extension? $\endgroup$
    – user95224
    Feb 23 '18 at 5:24
  • 1
    $\begingroup$ @user95224 On reading my proof, I find no assumptions made whatsoever about the extension $K/F$. $\endgroup$ Feb 23 '18 at 5:38

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.