5
$\begingroup$

Suppose that $p$ is unramified in a number field $K$ of degree $n$. Apparently, Stickelberger proved that $\big( \frac{Disc(K)}{p}\big) = (-1)^{n - g}$, where $g$ is the number of prime ideal factors over $(p)$ in $K$.

Is there a convenient reference for this fact?

Thank you! -Frank

$\endgroup$
4
  • 1
    $\begingroup$ mscand.dk/article.php?id=1427 $\endgroup$ Commented Apr 22, 2012 at 23:10
  • 4
    $\begingroup$ Frank, this is proved in Hasse's "Number Theory". $\endgroup$
    – KConrad
    Commented Apr 22, 2012 at 23:22
  • 4
    $\begingroup$ Stickelberger's paper appeared in the proceedings of the 1st ICM in 1897. A related, but simpler, result goes back to Pellet in 1878: for an odd prime $p$ and squarefree polynomial $h(x)$ in ${\mathbf F}_p[x]$ of degree $n$ with $g$ monic irreducible factors, $(\frac{{\rm disc} h}{p}) = (-1)^{n-g}$. $\endgroup$
    – KConrad
    Commented Apr 22, 2012 at 23:27
  • $\begingroup$ Stickelberger's original formulation is a little gem : Die Diskriminante des Körpers $\Omega$ ist durch die Primzahl $p$ nicht teilber, wenn $p$ ein Produkt von lauter verschiedenen Primidealen in ist; zugleich ist sie, wenn ungerade, quadratischer Rest oder Nichtrest von $p$, je nachdem die Anzahl der in $p$ aufgehenden Primideale von geradem Grade eine gerade oder ungerade ist, oder je nachdem die Anzahl aller Primfaktoren von $p$ dem Grade des Körpers kongruent ist nach dem Modul Zwei oder nicht. $\endgroup$ Commented Dec 30, 2012 at 10:09

1 Answer 1

6
$\begingroup$

First reduce the question to the local case--This can be done since after base change to $\mathbb{Q}_{p}$, $K \otimes_{\mathbb{Q}} \mathbb{Q}_{p}=K_1 \times...\times K_{g}$, the trace form is the orthogonal sum of the local trace forms so the discriminant is the product of the local discriminants, and then use that the Legendre symbol is multiplicative. Now, since $p$ is unramified the extension $K_{i}/\mathbb{Q}_{p}$ is a cyclic extension of degree $f_{i}$, notation as usual, with Galois group generated by $\sigma$. We want to show that $\left ( \frac{disc(K_i)}{p}\right)= (-1)^{f_i -1}$. This is done in the exact same way one proves that a cubic extension is Galois iff its discriminant is a square; just notice that $(-1)^{f_i -1}$ is the the signature of $\sigma$ under the regular representation of $Gal(K_{i}/\mathbb{Q}_{p})$.

Remark: If you, like Conway, like the notation $p=-1$ for the infinite prime then the above proof works at that prime too. Just read Kronecker symbol instead of Legendre, and what one gets is the usual "sign of discriminant is $(-1 )^{s}$ "

$\endgroup$
1
  • $\begingroup$ This proof goes back to Hensel, Crelle, 1905. $\endgroup$ Commented Dec 30, 2012 at 10:09

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .