MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

Hermite's theorem, as it is typically called, is that there are only finitely many number fields of bounded (equivalently, fixed) discriminant.

The usual proof (see Neukirch's Algebraic Number Theory for example) proceeds as follows. First, one proves the Minkowski bound: $|Disc(K)| > (C + o(1))^{deg \ K}$ where $C > 1$. This reduces the problem to fields of fixed degree. For fields $K$ of degree $n$ with $Disc(K) < X$, one can then obtain bounds on the coefficients of the minimal polynomial of $K$, and in particular there are only finitely many possibilities.

However, Minkowski's work was more than thirty years after Hermite's. I looked at Hermite's original paper and although I confess to not having read it in detail, it seems to be essentially the proof I described above. In particular he only claims on the first page to prove the theorem for fields of a fixed degree, and there is nothing I found in the paper which looks like it applies to all degrees.

So did Hermite actually prove the result that bears his name, or has he been given credit for the jazzed-up version, which apparently could only have been proved thirty years later?

And regardless of whether he actually did, might he have reasonably done so other than coming up with Minkowski's bound on his own?

Thank you!

share|cite|improve this question
The Minkowski bound provides a lower bound for the discriminant, and this is what you need to reduce the problem to fields of fixed degree. I corrected the inequality in your post. – GH from MO Jul 26 '12 at 4:19
@GH: Oops, that was indeed a typo. Thank you! – Frank Thorne Jul 26 '12 at 13:31

Hermite and Minkowski used completely different ideas.

Hermite, by induction on $n$, proved that a positive definite quadratic form $q = X^TAX$ of determinant $ |A| = 1$ and $n$-dimensional column-vector $X$ takes a nonzero value $q \leq (4/3)^{(n-1)/2}$ for some integer vector.

Minkowsky, by a purely geometrical consideration, discovering that spheres of radius of half the distance of two nearest lattice-points could be replaced by any symmetric convex body, achieved his estimation which is better than Hermite's for large $n$. Hilbert, in his obituary, praised: Minkowski's proof of a deep number theoretic theorem without computational help, mainly on geometrical vizualization, is a pearl of Minkowski's inventiveness.

A masterly and/but very readable presentation of both ideas has been given by C.L. Siegel: Lectures on Quadratic Forms

share|cite|improve this answer

What Rhett Bulter says is also detailed in a chapter entitled 'From Hermite to Minkowski' in a book of Scharlau and Opolka (From Fermat to Minkowski, Lectures on the Theory of Numbers and Its Historical Development, Springer, 1984).

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.