3
$\begingroup$

I am working through a number theory proof and there is one part whose justification I do not fully understand. I believe the reasoning is merely something related to localization.

Suppose $R$ is a Dedekind domain with fractions $K$. Let $L/K$ be normal with $G$ Galois group. Denote $R'$ the integral closure of $R$ in $L$. Let $\beta$ be a nonzero prime of $R'$, $p = R \cap \beta$ its contraction. Let $N$ denote the norm of an ideal, defined to be $$N(\mathfrak{a}) = \sum_{a \in \mathfrak{a}} RN(a)$$ I wish to show $N(\beta)$ is a power of the prime $p$. Suppose not, that $N(\beta) = qm$ with $q \ne p$ prime and $m \subseteq R$. Let $C$ be the complement of $q$ in $R$ so that $R_q = R_C$. Since $q \ne p$, the prime ideal $\beta$ is not a factor of $qR'$. Thus $\beta_C = R'_C$ as the only prime ideals of $R_C '$ are the extensions of the primes appearing in the factorization of $qR'$.

I understand that if $S$ is a multiplicative set and we localize $S^{-1} A$ a ring with respect to $S$, then the primes in the localization correspond bijectively to the primes of $A$ such that $p \cap S = \emptyset$. In the case $S = R - q$, then this corresponds to the primes contained within $q$. The images then correspond to the primes appearing in the factorization of the image of $q$, namely $qR'$. I understand that since $q \ne p$, then $\beta$ is not a factor of $qR'$. I don't see how one puts this together to get the desired equality.

I am thinking that one shows the set of prime ideals on each side is the same and since $\beta \subset R'$, then this would show that their localizations are the same.

$\endgroup$
  • $\begingroup$ Your definition of norm is not clear: what is $N(a)$ for $a\in \mathfrak a$? $\endgroup$ – Ferra Sep 16 '16 at 11:01
0
$\begingroup$

You just need to note that no prime $q\ne p$ divides $N(\beta)$ because $\Bbb Z$ has unique factorization. What the proof is doing is just saying another prime dividing $N(\beta)$ would imply that $qR'\supseteq N(\beta)$, but then take norms again and you get $N(N(\beta)) = q^{[R':R]}$ is a power of a different prime, $q$. But this is a contradiction to the fact that $(p)\supseteq\beta$.

$\endgroup$

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.