3
$\begingroup$

This is from Silverman's book "The arithmetic of elliptic curves" (AEC), p.36, lemma 5.8.1.

Lemma 5.8.1 states

Let $V$ be a $\overline{K}$-vector space, and assume that $G_{\overline{K}/K}$ acts continuously on $V$ in a manner compatible with its action on $\overline{K}$. Then, $V\cong \overline{K} \otimes_{K} V_K$.

AEC reads there is a fancy proof which uses Hilbert's theorem 90, which says that $H^1(G_{\overline{K}/K},\operatorname{GL}_n(K))=0$ (Exercise 2.13).

How can I apply Hilbert's theorem 90 to the proof of lemma 5.8.1?

$\endgroup$

1 Answer 1

11
$\begingroup$

Here's a sketch of the proof. I encourage you to fill in the details yourself. The definition of $V_K$ is $V_K=H^0(G_{\overline K/K},V)$. The key part of the proof is to show that $V$ has a $\overline{K}$ basis consisting of vectors in $V_K$. To find such a basis, start with an arbitrary basis $v_1,\ldots,v_n\in V$. Then each $\sigma\in G_{\overline K/K}$ gives a change of basis matrix $A_\sigma$ defined by writing $v_i^\sigma$ as a linear combination of $v_1,\ldots,v_n$. The map $$ G_{\overline K/K} \longrightarrow \operatorname{GL}_n(\overline K),\quad \sigma\longmapsto A_\sigma $$ is a $1$-cocycle. Using Hilbert Theorem 90, it follows that it is a coboundary, so there is a matrix $B\in\operatorname{GL}_n(\overline K)$ satisfying $A_\sigma=B^\sigma B^{-1}$. Now use $B$ (or maybe $B^{-1}$) to change the basis $v_1,\ldots,v_n$ to a new basis $w_1,\ldots,w_n$. If you've done it properly, the vectors $w_1,\ldots,w_n$ will be in $V_K$.

$\endgroup$
2
  • $\begingroup$ I didn't expect to receive an answer from the author. I love AEC,thank you for writing this great book. In the case $V$ is not finite dimmensional vector space over $\overline{K}$, how can I take $v_1,v_2,・・・,v_n$ from $V$ ? $\endgroup$ Jan 5, 2022 at 7:52
  • 1
    $\begingroup$ @Neronoggshafareivh Glad you like the book. If $V$ has a countable basis, probably you can do it by taking a sequence of finite dimensional subspaces $V_1\subset V_2\subset V_3\subset\cdots$ with $\dim(V_n)=n$ and generating the $K$-bases inductively. If $\dim(V)$ is uncountable, I never really thought about it, but maybe try some sort of Zorn's lemma argument? $\endgroup$ Jan 5, 2022 at 11:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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