1
$\begingroup$

Suppose we have a tower of field extensions:

$\overline{F} \subset K \subset E \subset F$

Is it true in general that $|G(K/F)| = |G(K/E)| \cdot |G(E/F)|$?

I was able to verify some specific examples, like $\mathbb{Q}(\sqrt[3]{2}, \omega)$ for $x^3-2$ and another extension, but how could I show that this holds in general for all such towers of extensions?

$\endgroup$
  • $\begingroup$ Are you assuming that $K$ and $E$ are Galois extensions of $F$?. If so then this is equivalent to the standard identity $|K:F| = |K:E| \cdot |E:F|$. $\endgroup$ – Rolf Hoyer May 17 '16 at 19:37
  • $\begingroup$ All I know is that they are extensions. $\endgroup$ – Ninosław Brzostowiecki May 17 '16 at 19:38
  • $\begingroup$ Then does $G(K/F)$ just mean $Aut(K/F)$, the automorphisms of $K$ fixing $F$? Or is it the Galois group of the normal closure of $K$ over $ F$? $\endgroup$ – Rolf Hoyer May 17 '16 at 19:42
  • $\begingroup$ It would have to be the latter,the Galois group of the normal closer of K over F. $\endgroup$ – Ninosław Brzostowiecki May 17 '16 at 19:52
1
$\begingroup$

This is generally not the case. For instance, take $K = \Bbb Q(\sqrt[4]{2}), E = \Bbb Q(\sqrt{2}), F = \Bbb Q$. Then $K/E$ and $E/F$ are both Galois extensions with $|G(K/E)| = |G(E/F)| = 2$. However, the extensions $K/F$ is not Galois, with there being only two automorphisms of $K$ fixing $F$, since such an automorphism is determined by its action on $\sqrt[4]{2}$, which can only be sent to $\pm \sqrt[4]{2}$, the only roots of $x^4-2$ lying in $K$. Thus, $|Aut(K/F)| = 2$.

If you interpret $|G(K/F)|$ to be the order of the Galois group of the Galois closure of $K$ over $F$, then the order is 8 instead, as the Galois closure, $\Bbb Q(\sqrt[4]{2}, i)$ is a degree 8 extension.

$\endgroup$
  • $\begingroup$ So if one of our assumptions was that these are Galois extensions, the equality would hold? $\endgroup$ – Ninosław Brzostowiecki May 18 '16 at 20:21
  • $\begingroup$ @NinosławCiszewski If all three extensions $K/F, E/F, K/E$ are Galois, then the equality would hold, yes. $\endgroup$ – Rolf Hoyer May 18 '16 at 20:25

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.