(Quadratic) Gauss sums express square root of any integer as a sum of roots of unity (or of cosines of rational multiples of $2\pi$, if you will) with rational coefficients.

But Kronecker-Weber guarantees that any root of any integer can be expressed as a sum of that kind. What is the Is there corresponding formula for, say, $\sqrt[3]p$?

Upd. I'm sorry, but original question doesn't make much sense. The question I, perhaps, meant to ask is (as Matt E kindly points out) discussed in David Speyer's answer.

In particular, for $p=3k+1$ the cubic sum $\displaystyle\sum_{t\in\mathbb Z/p}\cos\left(\frac{2\pi t^3}p\right)$ is a root of the equation $x^3-3px-Ap=0$ where $4p=A^2+27B^2$ and $A\equiv 1\pmod 3$ (the discriminant of the cubic equation is $(27pB)^2$).


The Kronecker--Weber theorem does not guarantee what you claim. Indeed, $\mathbb Q(p^{1/3})$ is not Galois over $\mathbb Q$, and its Galois closure is equal to $\mathbb Q(\sqrt{-3},p^{1/3})$, which is an $S_3$ (and hence non-abelian) extension of $\mathbb Q$. It is an abelian extension of $\mathbb Q(\sqrt{-3})$, but not all extensions of this field are cyclotomic; rather, they are described by the theory of complex multiplication (Kronecker's Jugendtraum).

  • $\begingroup$ (Actually, first I wanted to ask about "Gauss-sum-type" formulas for roots of a cubic polynomial with perfect square discriminant -- but then thought that just cube roots should be more tractable.) $\endgroup$ – Grigory M Jan 26 '12 at 20:19
  • 2
    $\begingroup$ @Grigory: Dear Grigory, If the discriminant is square, then you do get an abelian extension of $\mathbb Q$, which is thus cyclotomic. David Scheyer has a post explaining a very concrete approach to Kronecker--Weber in this case: math.stackexchange.com/a/31600/221 Regards, $\endgroup$ – Matt E Jan 26 '12 at 20:22

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.