Let $(K,u)$ be a complete valued field, $u$ be its discrete absolute value (corresponds to a discrete valuation on $K$), then:

($\ast)$ Let $E/K$ is a finite separable field extension, then the absolute value $w$ of $E$ which extends $u$ is unique.

In the book Algebraic Number Theory by Fröhlich, A., and Taylor, M.J., it first proves the above property, then gives an alternate proof as follows:

Let $\mathfrak o$ denotes the valuation ring in $K$ associates with $u$, $\mathfrak o_E$ the integral closure of $\mathfrak o$ in $E$,
then in order to prove ($\ast$), it suffices to show that $\mathfrak o_E$ has the unique prime ideal $\mathfrak B$ above $\mathfrak o$.

($\mathfrak B$ is above $\mathfrak o$ means $\mathfrak B \cap \mathfrak o=\mathfrak p$ is an unique prime ideal in $\mathfrak o$.)

My question is, how the uniqueness of $\mathfrak B$ implies the uniqueness of extension $w$, why it suffices?

I know there exists a absolute value $w'$ of $E$ extends $u$ which gives $\mathfrak {B} =\{ x; w'(x)<1 \}$, but this only works for such specific $w'$, for an arbitrary extension $w$, not even being discrete, I am not sure how it gives a prime ideal in $\mathfrak o_E$ to obtain the uniquemess.


You talk about Theorem 16 on Page 103 and its alternate proof in a special case starting at the bottom of Page 105.

Indeed there seems to be a small gap in the alternate proof as it only deals with discrete extensions $w$ of the given discrete absolute value $u$. Probably what the authors had in mind is this: let us assume that both $u$ and $w$ are discrete absolute values in Theorem 16, then we can give an alternate proof by algebraic methods.


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.