Power residue symbol

In algebraic number theory the n-th power residue symbol (for an integer n > 2) is a generalization of the (quadratic) Legendre symbol to n-th powers. These symbols are used in the statement and proof of cubic, quartic, Eisenstein, and related higher[1] reciprocity laws.[2]

Background and notation

Let k be an algebraic number field with ring of integers O k {\displaystyle {\mathcal {O}}_{k}} that contains a primitive n-th root of unity ζ n . {\displaystyle \zeta _{n}.}

Let p O k {\displaystyle {\mathfrak {p}}\subset {\mathcal {O}}_{k}} be a prime ideal and assume that n and p {\displaystyle {\mathfrak {p}}} are coprime (i.e. n p {\displaystyle n\not \in {\mathfrak {p}}} .)

The norm of p {\displaystyle {\mathfrak {p}}} is defined as the cardinality of the residue class ring (note that since p {\displaystyle {\mathfrak {p}}} is prime the residue class ring is a finite field):

N p := | O k / p | . {\displaystyle \mathrm {N} {\mathfrak {p}}:=|{\mathcal {O}}_{k}/{\mathfrak {p}}|.}

An analogue of Fermat's theorem holds in O k . {\displaystyle {\mathcal {O}}_{k}.} If α O k p , {\displaystyle \alpha \in {\mathcal {O}}_{k}-{\mathfrak {p}},} then

α N p 1 1 mod p . {\displaystyle \alpha ^{\mathrm {N} {\mathfrak {p}}-1}\equiv 1{\bmod {\mathfrak {p}}}.}

And finally, suppose N p 1 mod n . {\displaystyle \mathrm {N} {\mathfrak {p}}\equiv 1{\bmod {n}}.} These facts imply that

α N p 1 n ζ n s mod p {\displaystyle \alpha ^{\frac {\mathrm {N} {\mathfrak {p}}-1}{n}}\equiv \zeta _{n}^{s}{\bmod {\mathfrak {p}}}}

is well-defined and congruent to a unique n {\displaystyle n} -th root of unity ζ n s . {\displaystyle \zeta _{n}^{s}.}

Definition

This root of unity is called the n-th power residue symbol for O k , {\displaystyle {\mathcal {O}}_{k},} and is denoted by

( α p ) n = ζ n s α N p 1 n mod p . {\displaystyle \left({\frac {\alpha }{\mathfrak {p}}}\right)_{n}=\zeta _{n}^{s}\equiv \alpha ^{\frac {\mathrm {N} {\mathfrak {p}}-1}{n}}{\bmod {\mathfrak {p}}}.}

Properties

The n-th power symbol has properties completely analogous to those of the classical (quadratic) Jacobi symbol ( ζ {\displaystyle \zeta } is a fixed primitive n {\displaystyle n} -th root of unity):

( α p ) n = { 0 α p 1 α p  and  η O k : α η n mod p ζ α p  and there is no such  η {\displaystyle \left({\frac {\alpha }{\mathfrak {p}}}\right)_{n}={\begin{cases}0&\alpha \in {\mathfrak {p}}\\1&\alpha \not \in {\mathfrak {p}}{\text{ and }}\exists \eta \in {\mathcal {O}}_{k}:\alpha \equiv \eta ^{n}{\bmod {\mathfrak {p}}}\\\zeta &\alpha \not \in {\mathfrak {p}}{\text{ and there is no such }}\eta \end{cases}}}

In all cases (zero and nonzero)

( α p ) n α N p 1 n mod p . {\displaystyle \left({\frac {\alpha }{\mathfrak {p}}}\right)_{n}\equiv \alpha ^{\frac {\mathrm {N} {\mathfrak {p}}-1}{n}}{\bmod {\mathfrak {p}}}.}
( α p ) n ( β p ) n = ( α β p ) n {\displaystyle \left({\frac {\alpha }{\mathfrak {p}}}\right)_{n}\left({\frac {\beta }{\mathfrak {p}}}\right)_{n}=\left({\frac {\alpha \beta }{\mathfrak {p}}}\right)_{n}}
α β mod p ( α p ) n = ( β p ) n {\displaystyle \alpha \equiv \beta {\bmod {\mathfrak {p}}}\quad \Rightarrow \quad \left({\frac {\alpha }{\mathfrak {p}}}\right)_{n}=\left({\frac {\beta }{\mathfrak {p}}}\right)_{n}}

All power residue symbols mod n are Dirichlet characters mod n, and the m-th power residue symbol only contains the m-th roots of unity, the m-th power residue symbol mod n exists if and only if m divides λ ( n ) {\displaystyle \lambda (n)} (the Carmichael lambda function of n).

Relation to the Hilbert symbol

The n-th power residue symbol is related to the Hilbert symbol ( , ) p {\displaystyle (\cdot ,\cdot )_{\mathfrak {p}}} for the prime p {\displaystyle {\mathfrak {p}}} by

( α p ) n = ( π , α ) p {\displaystyle \left({\frac {\alpha }{\mathfrak {p}}}\right)_{n}=(\pi ,\alpha )_{\mathfrak {p}}}

in the case p {\displaystyle {\mathfrak {p}}} coprime to n, where π {\displaystyle \pi } is any uniformising element for the local field K p {\displaystyle K_{\mathfrak {p}}} .[3]

Generalizations

The n {\displaystyle n} -th power symbol may be extended to take non-prime ideals or non-zero elements as its "denominator", in the same way that the Jacobi symbol extends the Legendre symbol.

Any ideal a O k {\displaystyle {\mathfrak {a}}\subset {\mathcal {O}}_{k}} is the product of prime ideals, and in one way only:

a = p 1 p g . {\displaystyle {\mathfrak {a}}={\mathfrak {p}}_{1}\cdots {\mathfrak {p}}_{g}.}

The n {\displaystyle n} -th power symbol is extended multiplicatively:

( α a ) n = ( α p 1 ) n ( α p g ) n . {\displaystyle \left({\frac {\alpha }{\mathfrak {a}}}\right)_{n}=\left({\frac {\alpha }{{\mathfrak {p}}_{1}}}\right)_{n}\cdots \left({\frac {\alpha }{{\mathfrak {p}}_{g}}}\right)_{n}.}

For 0 β O k {\displaystyle 0\neq \beta \in {\mathcal {O}}_{k}} then we define

( α β ) n := ( α ( β ) ) n , {\displaystyle \left({\frac {\alpha }{\beta }}\right)_{n}:=\left({\frac {\alpha }{(\beta )}}\right)_{n},}

where ( β ) {\displaystyle (\beta )} is the principal ideal generated by β . {\displaystyle \beta .}

Analogous to the quadratic Jacobi symbol, this symbol is multiplicative in the top and bottom parameters.

  • If α β mod a {\displaystyle \alpha \equiv \beta {\bmod {\mathfrak {a}}}} then ( α a ) n = ( β a ) n . {\displaystyle \left({\tfrac {\alpha }{\mathfrak {a}}}\right)_{n}=\left({\tfrac {\beta }{\mathfrak {a}}}\right)_{n}.}
  • ( α a ) n ( β a ) n = ( α β a ) n . {\displaystyle \left({\tfrac {\alpha }{\mathfrak {a}}}\right)_{n}\left({\tfrac {\beta }{\mathfrak {a}}}\right)_{n}=\left({\tfrac {\alpha \beta }{\mathfrak {a}}}\right)_{n}.}
  • ( α a ) n ( α b ) n = ( α a b ) n . {\displaystyle \left({\tfrac {\alpha }{\mathfrak {a}}}\right)_{n}\left({\tfrac {\alpha }{\mathfrak {b}}}\right)_{n}=\left({\tfrac {\alpha }{\mathfrak {ab}}}\right)_{n}.}

Since the symbol is always an n {\displaystyle n} -th root of unity, because of its multiplicativity it is equal to 1 whenever one parameter is an n {\displaystyle n} -th power; the converse is not true.

  • If α η n mod a {\displaystyle \alpha \equiv \eta ^{n}{\bmod {\mathfrak {a}}}} then ( α a ) n = 1. {\displaystyle \left({\tfrac {\alpha }{\mathfrak {a}}}\right)_{n}=1.}
  • If ( α a ) n 1 {\displaystyle \left({\tfrac {\alpha }{\mathfrak {a}}}\right)_{n}\neq 1} then α {\displaystyle \alpha } is not an n {\displaystyle n} -th power modulo a . {\displaystyle {\mathfrak {a}}.}
  • If ( α a ) n = 1 {\displaystyle \left({\tfrac {\alpha }{\mathfrak {a}}}\right)_{n}=1} then α {\displaystyle \alpha } may or may not be an n {\displaystyle n} -th power modulo a . {\displaystyle {\mathfrak {a}}.}

Power reciprocity law

The power reciprocity law, the analogue of the law of quadratic reciprocity, may be formulated in terms of the Hilbert symbols as[4]

( α β ) n ( β α ) n 1 = p | n ( α , β ) p , {\displaystyle \left({\frac {\alpha }{\beta }}\right)_{n}\left({\frac {\beta }{\alpha }}\right)_{n}^{-1}=\prod _{{\mathfrak {p}}|n\infty }(\alpha ,\beta )_{\mathfrak {p}},}

whenever α {\displaystyle \alpha } and β {\displaystyle \beta } are coprime.

See also

Notes

  1. ^ Quadratic reciprocity deals with squares; higher refers to cubes, fourth, and higher powers.
  2. ^ All the facts in this article are in Lemmermeyer Ch. 4.1 and Ireland & Rosen Ch. 14.2
  3. ^ Neukirch (1999) p. 336
  4. ^ Neukirch (1999) p. 415

References

  • Gras, Georges (2003), Class field theory. From theory to practice, Springer Monographs in Mathematics, Berlin: Springer-Verlag, pp. 204–207, ISBN 3-540-44133-6, Zbl 1019.11032
  • Ireland, Kenneth; Rosen, Michael (1990), A Classical Introduction to Modern Number Theory (Second edition), New York: Springer Science+Business Media, ISBN 0-387-97329-X
  • Lemmermeyer, Franz (2000), Reciprocity Laws: from Euler to Eisenstein, Springer Monographs in Mathematics, Berlin: Springer Science+Business Media, doi:10.1007/978-3-662-12893-0, ISBN 3-540-66957-4, MR 1761696, Zbl 0949.11002
  • Neukirch, Jürgen (1999), Algebraic number theory, Grundlehren der Mathematischen Wissenschaften, vol. 322, Translated from the German by Norbert Schappacher, Berlin: Springer-Verlag, ISBN 3-540-65399-6, Zbl 0956.11021