I am reading Serre's Local Fields. In Section V.4, Serre considers a finite totally ramified extension of local fields $L/K$ with the residue field $\bar{L}=\bar{K}$ a perfect field. For $\bar{K}'$ a finite extension of $\bar{K}$, let $K'$ be the corresponding unramified extension of $K$, and let $L'=LK'$. The norm maps (for $n\geq 1$) $$ N_n: U^{\psi(n)}_L/U^{\psi(n)+1}_L \to U^n_K/U^{n+1}_K $$ $$ N'_n: U^{\psi(n)}_{L'}/U^{\psi(n)+1}_{L'} \to U^n_{K'}/U^{n+1}_{K'} $$ can be identified with $N_n:\bar{K}\to \bar{K}$ and $N'_n:\bar{K}'\to \bar{K}'$, given by polynomials that are determined in Section V.3. He remarks that the coefficients of the polynomials are the same for $N_n$ and $N'_n$ (which I believe), and then he writes:

In the language of algebraic geometry, that means we have for each $n\geq 1$ a homomorphism $v_n:G_a\to G_a$ rational over $\bar{K}$, and that $N_n'$ is the restriction of $v_n$ to the points of $G_a$ rational over $\bar{K}'$.

Can some one elaborate on it? I have some exposure to algebraic geometry (from Hartshorne's textbook) but I fail to recognize anything familiar here. I do not even know what the word rational means here.



$G_a$ is the additive group variety. There are various more and less complicated versions of how to describe it but let me give a very simple one you may believe.

For $F$ a field, the affine line $\mathbb{A}^1_F$ is also equipped with point-wise addition. If we denote this pair of $(\mathbb{A}^1_F, +)$ as $G_a/F$ then because this addition is polynomial one can formulate that $G_a/F$ not just equipped point-wise with an operation, but that this operation can be phrased in terms of the category of $F$-schemes where $G_a/F$ will be a group object.

Here Serre is saying that the Norm maps for $n \geq 1$ are group homomorphisms $(\overline{K},+) \rightarrow (\overline{K}, +)$ which are given by polynomials. Now $(\overline{K},+)$ is the $\overline{K}$ points of $G_a/\overline{K}$ and so one can expect that these polynomials can be lifted to $G_a/\overline{K}$ and made into a morphism of group objects.


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.