0

$\divideontimes$ Refer to the Janusz, Algebraic number fields, Chap II- Example 2 or Nuekirch, Algebraic number theory, p.69.

Let $\mathcal{o}$ be a Dedekind domain wiht fraction field $K$. Let $\mathfrak{p} \neq 0$ be a nonzero prime ideal in $\mathcal{o}$. For any nonzero $x$ in $\mathcal{o}$ let $v_{\mathfrak{p}}(x)$ denote the power to which $\mathfrak{p}$ appears in the factorization of $(x)$. We may then wrie

$$ (x) = \prod \mathfrak{p}^{v_{\mathfrak{p}(x)}}, \ \mathfrak{p} \text{ runs through the primes. }$$

Then I don't know how can we define corresponding valuation $v_{\mathfrak{p}} : K^{*}\to \mathbb{Z}$ clearly. How can we lift the function $v_{\mathfrak{p}}$ on $\mathcal{o}$ to $K^{*}$?

And assoicated question : Let $L|K$ be a ( finite ) Galois extension of number fields. $\mathfrak{P}_1, \mathfrak{P}_j \subseteq \mathcal{O}_{L}$ be prime ideals of $L$ such that $\sigma \mathfrak{P}_1 = \mathfrak{P}_j$ for some $\sigma \in G(L|K)$. Then the corresponding valuations satisfy $v_{\mathfrak{P}_j} \circ \sigma = v_{\mathfrak{P}_1}$ ?

It is sufficient to show the relation only on $\mathcal{O}_L$? Assume this. Fix $x\in \mathcal{O}_L$. Then

$$ \mathfrak{P}_j^{v_{\mathfrak{P}_j}(\sigma(x))} \prod_{\mathfrak{P}_j \neq \mathfrak{P} } \mathfrak{P}^{v_{\mathfrak{P}}(\sigma(x))} = \prod_{\mathfrak{P}}\mathfrak{P}^{v_{\mathfrak{P}}(\sigma(x))} = ( \sigma(x) ) \stackrel{?}{=} \sigma((x)) = \sigma(\prod_{\mathfrak{P}}\mathfrak{P}^{v_{\mathfrak{P}}(x)}) = \sigma ( \mathfrak{P}_1^{v_{\mathfrak{P}_1}(x)} \prod_{\mathfrak{P}_1 \neq \mathfrak{P}}\mathfrak{P}^{v_{\mathfrak{P}}(x)}) \stackrel{?}{=} \sigma(\mathfrak{P}_1)^{v_{\mathfrak{P}_1}(x)} \prod_{\mathfrak{P}_1 \neq \mathfrak{P}}(\sigma(\mathfrak{P}))^{v_{\mathfrak{P}}(x)} = \mathfrak{P}_j^{v_{\mathfrak{P}_1}(x)}\prod_{\mathfrak{P}_1 \neq \mathfrak{P}}(\sigma(\mathfrak{P}))^{v_{\mathfrak{P}}(x)} . \tag{1} $$

Note that $\{ \mathfrak{P} \in \operatorname{Spec}(\mathcal{O}_L) : \mathfrak{P}_j \neq \mathfrak{P} \} = \{ \sigma(\mathfrak{P}) : \mathfrak{P} \in \operatorname{Spec}(\mathcal{O}_L), \mathfrak{P}_1 \neq \mathfrak{P} \}$ ( Check ). So for far left and far right in $(1)$, in $\prod_{\mathfrak{P}_j \neq \mathfrak{P} } \mathfrak{P}^{v_{\mathfrak{P}}(\sigma(x))}$ and $\prod_{\mathfrak{P}_1 \neq \mathfrak{P}}(\sigma(\mathfrak{P}))^{v_{\mathfrak{P}}(x)}$, there is no $\mathfrak{P}_j$-term. So by comparing exponentials, $v_{\mathfrak{P}_j}(\sigma(x)) = v_{\mathfrak{P}_1}(x)$. And it remains to prove our claim for arbitrary $x\in K^{*}$. For this, proof of the case $x\in \mathcal{O}_L$ is sufficient? Can anyone help?

Plantation
  • 3,710
  • 1
    For your first question, knowing that $K$ is the fraction field of $\mathcal{o}$, you can guess how one extends the valuation to $K^{\times}$ (so as to obtain a group homomorphim). Then you can check that your guess is well defined. This shows that the valuation on the Dedekind domain determines the valuation on the fraction field, and answers your second question. – François Gatine Aug 06 '24 at 07:44
  • Thank you. I guess that $\bar{v_{\mathfrak{p}}}(a/b) := v_{\mathfrak{p}}(a)-v_{\mathfrak{p}}(b)$. ( C.f. https://math.stackexchange.com/questions/674658/how-do-we-extend-the-valuation-on-kx-to-a-valuation-on-kx/674944#674944 ) Perhaps, do you know whether there is any other reference which deal with this kind of stuff? – Plantation Aug 06 '24 at 08:27
  • 1
    That's right (and you immediately see the independence on the representative). The main "official" reference I know of is Neukirch, although there might be plenty more. When I first studied Dedekind domains and valuation rings for a master thesis, I just looked at loads of online notes, and that did the trick. – François Gatine Aug 06 '24 at 08:42
  • Wow, thanks. I also looked up Neukirch's book, perhaps, do you know what pages of Neukrich's book the related content appears on? – Plantation Aug 06 '24 at 08:50

0 Answers0