MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $F$ be a finite extension of $Q_2$ (2-adic field) or $F_2((x))$ (function field). Let $E/F$ be a separable extension of degree $2$.

  1. What is the image of the norm map $N_{E/F}$?

  2. In particular - is it true that the index $[F^{\star} : N_{E/F}(E^{\star})]$ depends only on the ramification $e(E|F)$?

share|cite|improve this question
What does "some extension" mean? If you meant it to be a finite extension or an algebraic extension, please say so. Also, is the quadratic extension in the function field case supposed to be Galois or not? – KConrad Feb 14 '11 at 11:09
Thanks for you comment KConard. I have edited the question. – Pooja Feb 14 '11 at 12:55
Just a remark: while (as the answers below show) it is not true that $[F^\times : N_{E/F} E^\times]$ has anything to do with ramification, it is true for any finite abelian extension of local fields that $[\mathcal{O}_F^\times : N_{E/F} \mathcal{O}_E^\times] = e_{E/F}$. – David Loeffler Feb 14 '11 at 15:49
Thank you very much for your comment David. It is very useful for me. Would you suggest a reference for this result? I am also trying to understand the image of groups $1 + \pi^t \mathcal{O}_E$, $t \in \mathbb{Z}$ under the norm map in above situation. Do you have any comments for these groups? – Pooja Feb 15 '11 at 8:24
Serre's "Local Fields" is the bible here. Milne's notes on local class field theory (from his website) are also very good. As for the subgroups you mention: if t is positive and reasonably large, the p-adic logarithm map will give an isomorphism between $(1 + \pi^t \mathcal{O}_E, \times)$ and $(\pi^t \mathcal{O}_E, +)$, and the norm on the left-hand side corresponds to the trace on the right, which reduces it to a much easier question. For t positive but small there might be some messier behaviour. I don't know what the question means for $t < 0$. – David Loeffler Feb 15 '11 at 14:01
up vote 6 down vote accepted

As KConrad points out, you perhaps mean to say that $F$ is a finite extension of $\mathbf{Q}_2$ or of $\mathbf{F}_2((x))$, and that the quadratic extesnions $E|F$ is separable (and hence galoisian) in the second case.

With this interpretation of the question, $N_{E|F}(E^\times)$ is a closed subgroup of index $2$ in $F^\times$, and every closed subgroup of index $2$ in $F^\times$ is of this form. In particular, the ramification index $e_{E|F}$ does not determine the subgroup in question.

For more on this, see the relevant chapter in Serre's Corps locaux (=Local fields) or the book by Fesenko and Vostokov, among many other places.

Addendum. In the same vein as David Speyer's example, it might also be instructive to work out the case $F=\mathbf{F}_2((x))$,by using "Artin-Schreier theory" instead of "Kummer theory". Cyclic quadratic extensions of $F$ correspond to $\mathbf{F}_2$-lines $D\subset F/\wp(F)$, where $\wp$ is the endomorhphism $t\mapsto t^2-t$ of the additive group of $F$. The quotient $F/\wp(F)$ carries a natural filtration, coming from the filtration on the additive group $F$, so every line $D$ has a level in this filtration. The ramification index of the cyclic quadratic extension $E_D=F(\wp^{-1}(D))$ of $F$ corresponding to $D$ depends only on this level.

It would be interesting to work out the norm subgroup $N_{E_D|F}(E_D^\times)\subset F^\times$ for each line $D$.

By the way, the Second Extended Edition of Fesenko-Vostokov is available online.

Addendum 2. See this answer for a generalisation from $p=2$ to arbitrary primes $p$.

share|cite|improve this answer
Thank you very much for your reply. I can not read the fourth line of your reply, would you please edit it a little bit? Thanks. – Pooja Feb 14 '11 at 13:01
Is it better now ? – Chandan Singh Dalawat Feb 14 '11 at 13:04
Yes, surely it is. Thank you very much for your reply. – Pooja Feb 15 '11 at 7:57

It might be worth working the example of $\mathbb{Q}_2$. Recall that $u \in \mathbb{Q}_2$ is a square if and only if $u$ is of the form $4^k (1+8 \ell)$ for $\ell \in \mathbb{Z}_2$. So $\mathbb{Q}_2^{\times} / (\mathbb{Q}_2^{\times})^2$ has order $8$, with representative elements being $1$, $3$, $5$, $7$, $2$, $6$, $10$ and $14$. This is an important computation for two reasons:

(1) Quadratic extensions of a characteristic zero field are of the form $K(\sqrt{a})$ for some nonsquare $a$, and two different $a$'s give the same extension if there ratio is a square. So there are $7$ quadratic extensions of $\mathbb{Q}_2$. The unramified one is $\mathbb{Q}_2(\sqrt{5})$.

(2) If $L$ is any quadratic extension, then $(\mathbb{Q}_2^\times)^2 \subset N_{L/\mathbb{Q}_2} (L^\times)$ as, for $a \in \mathbb{Q}_2$, we have $N(a)=a^2$. So we can describe the norm group by giving its image in the $8$ element group $\mathbb{Q}^{\times}_2/(\mathbb{Q}_2^{\times})^2$.

So, for example, in $\mathbb{Q}_2(\sqrt{3})$, the norms are elements of the form $a^2 - 3 b^2$. A little checking shows that the image in $\mathbb{Q}^{\times}_2/(\mathbb{Q}_2^{\times})^2$ is represented by $\{ 1, 1-3 \cdot 2^2, 3^2 -3, 1 - 3 \} \equiv \{ 1, 5, 6, 10 \}$. You can enjoy writing down the $4$ element subgroup of $\{ 1,3,5,7,2,6,10,14 \}$ corresponding to each of the $7$ quadratic extensions.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.