# Gauss sum

In algebraic number theory, a Gauss sum or Gaussian sum is a particular kind of finite sum of roots of unity, typically

${\displaystyle G(\chi ):=G(\chi ,\psi )=\sum \chi (r)\cdot \psi (r)}$

where the sum is over elements r of some finite commutative ring R, ψ is a group homomorphism of the additive group R+ into the unit circle, and χ is a group homomorphism of the unit group R× into the unit circle, extended to non-unit r, where it takes the value 0. Gauss sums are the analogues for finite fields of the Gamma function.[clarification needed]

Such sums are ubiquitous in number theory. They occur, for example, in the functional equations of Dirichlet L-functions, where for a Dirichlet character χ the equation relating L(s, χ) and L(1 − s, χ) (where χ is the complex conjugate of χ) involves a factor[clarification needed]

${\displaystyle {\frac {G(\chi )}{|G(\chi )|}}.}$

## History

The case originally considered by Carl Friedrich Gauss was the quadratic Gauss sum, for R the field of residues modulo a prime number p, and χ the Legendre symbol. In this case Gauss proved that G(χ) = p12 or ip12 for p congruent to 1 or 3 modulo 4 respectively (the quadratic Gauss sum can also be evaluated by Fourier analysis as well as by contour integration).

An alternate form for this Gauss sum is:

${\displaystyle \sum e^{\frac {2\pi ir^{2}}{p}}}$

Quadratic Gauss sums are closely connected with the theory of theta functions.

The general theory of Gauss sums was developed in the early 19th century, with the use of Jacobi sums and their prime decomposition in cyclotomic fields. Gauss sums over a residue ring of integers mod N are linear combinations of closely related sums called Gaussian periods.

The absolute value of Gauss sums is usually found as an application of Plancherel's theorem on finite groups. In the case where R is a field of p elements and χ is nontrivial, the absolute value is p12. The determination of the exact value of general Gauss sums, following the result of Gauss on the quadratic case, is a long-standing issue. For some cases see Kummer sum.

## Properties of Gauss sums of Dirichlet characters

The Gauss sum of a Dirichlet character modulo N is

${\displaystyle G(\chi )=\sum _{a=1}^{N}\chi (a)e^{\frac {2\pi ia}{N}}.}$

If χ is also primitive, then

${\displaystyle |G(\chi )|={\sqrt {N}},}$

in particular, it is nonzero. More generally, if N0 is the conductor of χ and χ0 is the primitive Dirichlet character modulo N0 that induces χ, then the Gauss sum of χ is related to that of χ0 by

${\displaystyle G(\chi )=\mu \left({\frac {N}{N_{0}}}\right)\chi _{0}\left({\frac {N}{N_{0}}}\right)G\left(\chi _{0}\right)}$

where μ is the Möbius function. Consequently, G(χ) is non-zero precisely when N/N0 is squarefree and relatively prime to N0.[1]

Other relations between G(χ) and Gauss sums of other characters include

${\displaystyle G({\overline {\chi }})=\chi (-1){\overline {G(\chi )}},}$

where χ is the complex conjugate Dirichlet character, and if χ is a Dirichlet character modulo N such that N and N are relatively prime, then

${\displaystyle G\left(\chi \chi ^{\prime }\right)=\chi \left(N^{\prime }\right)\chi ^{\prime }(N)G(\chi )G\left(\chi ^{\prime }\right).}$

The relation among G(χχ′), G(χ), and G(χ′) when χ and χ are of the same modulus (and χχ is primitive) is measured by the Jacobi sum J(χ, χ′). Specifically,

${\displaystyle G\left(\chi \chi ^{\prime }\right)={\frac {G(\chi )G\left(\chi ^{\prime }\right)}{J\left(\chi ,\chi ^{\prime }\right)}}.}$