3
$\begingroup$

I am learning about cyclotomic fields (hoping to understand the basics of Iwasawa theory) using Washington, Introduction to Cyclotomic Fields. I was reading this proposition, it seems to be using "quite advanced" results such as

If $K$ is a number field and $d_K$ is the discriminant of $K$, then

$$ p|d_K \text{ if and only if } p \text{ ramifies in } K$$

and $$ |d_K| > 1 \text{ if } K \neq \mathbb{Q}$$

which follows from Minkowski's bound according to Wikipedia.

enter image description here

According to Lang (the bottom image),

enter image description here

which somehow magically proves the same statement from

$$ \mathbb{Q}(\zeta_n) \mathbb{Q}(\zeta_m) = \mathbb{Q}(\zeta_{mn}) $$

Am I missing something, or is it just a typo? Thank you in advance.

$\endgroup$
4
$\begingroup$

Here is what I think Lang had in mind: earlier in the same chapter, Lang proves that if $F$ and $K$ are extensions of a field $k$ with $K$ Galois over $k$, then $KF$ is Galois over $F$ and $K$ is Galois over $K\cap F$, with $\mathrm{Gal}(KF/F)\simeq \mathrm{Gal}(K/K\cap F)$. In particular, $$ [KF:F]=[K:K\cap F]$$

In your problem, let $K=\mathbb{Q}(\zeta_m)$, $F=\mathbb{Q}(\zeta_n)$, and $k=\mathbb{Q}$. Then $KF=\mathbb{Q}(\zeta_{mn})$, hence the above equality becomes $$ [\mathbb{Q}(\zeta_{mn}):\mathbb{Q}(\zeta_n)]=[\mathbb{Q}(\zeta_m):\mathbb{Q}(\zeta_m)\cap\mathbb{Q}(\zeta_n)]$$ Now the left-hand side is equal to $$ \frac{[\mathbb{Q}(\zeta_{mn}):\mathbb{Q}]}{[\mathbb{Q}(\zeta_{n}):\mathbb{Q}]}=\frac{\phi(mn)}{\phi(n)}=\phi(m)$$ by the tower rule, hence the right-hand side must also be $\phi(m)=[\mathbb{Q}(\zeta_m):\mathbb{Q}]$. But this implies that $\mathbb{Q}(\zeta_m)\cap\mathbb{Q}(\zeta_n)=\mathbb{Q}$.

$\endgroup$
  • $\begingroup$ Thank you for the reference. (Theorem 1.12 in the same chapter) $\endgroup$ – libofmath Jul 13 '17 at 6:21

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.