Blog Archive

Friday, April 5, 2024

More on Legendre symbol

From Legendre symbol: An exact sequence point of view we know that the Legendre symbol is the unique group homomorphism Up→{−1,1}.

In this blog, we would like to introduce some useful fact about it.

Proposition.

(−1p)=(−1)p−12={1,p≡1mod4−1,p≡3mod4

Firstly let me explain why we only need to consider p≡1mod4 or p≡3mod4.

Since p≥3, p has to be odd number. If p≡0mod4 or p≡2mod4, then obviously p is even number.

Proof.

If p≡1mod4, p=4k+1

(−1p)=(−1)p−12=(−1)4k+1−12=(−1)2k=1

If p≡3mod4

(−1p)=(−1)p−12=(−1)2k+1=−1◻

Gauss Lemma.

We already know that the Legendre symbol is nothing but UP→{−1,1}.

We could embed i:UP↪Sp−1. If i(a) is an odd permutation, then (−1p)=−1, if i(a) is even, then (−1p)=1​.

As we know, sign:Sp−1→{−1,1} is defined by

∏1≤i<i≤p−1(aj−ai)=sign(a)∏1≤i<i≤p−1(j−i)

Traditionaly we should consider the unit group Up, let a∈Up acts on it self, we get:​

(123...p−1a2a3a...a(p−1))

But we already know that

(ap)≡ap−12modp

Hence let us consider the a∈Up acts on S={1,2,3,...,p−12} and see what happens.

Let us rewrite the set {1,2,3,...,p−1} to {−p−12,...,−1,1,...,p−12}

Hence

{a,2a,3a,...,ap−12}≡{±s1,±s2,±s3,...,±sp−12}

Where sk∈{1,2,3,...,p−12} . Since the acttion is faithful, or image that Fp as a one dimensional vector space,

then v1≠±v2⟹av1≠±av2, since a is invertible.

Then we claim that:

{s1,...,sp−12}={1,...,p−12}

Let μ=|{x∈S|ax≥p2}|. The Gauss Lemma tell us that

(ap)=(−1)μ

Proof.

Since

{a,2a,3a,...,ap−12}≡{±s1,±s2,±s3,...,±sp−12}

Take the product both sides, we get

ap−12(p−12)!≡(−1)μs1...sp−12≡(−1)μ(p−12)!modp

Hence

ap−12≡(−1)μmodp

Proposition.

(2p)=(−1)p2−18={1,p≡±1mod8−1,p≡±3mod8

According to Gauss Lemma, we need to find the number that p−12<2k≤2p−12.

i.e

p−14<k≤p−12⟺⌊p−14⌋<k≤p−12

When p=8k+1,8k+3,8k+5,8k+7:

p−12=4k,4k+1,4k+2,4k+3, ⌊p−14⌋=2k,2k,2k+1,2k+1

Hence

μ=p−14−⌊p−14⌋=2k,2k+1,2k+1,2k+2

By Gauss Lemma

(2p)=(−1)p2−18={1,p≡±1mod8−1,p≡±3mod8

For the proof of quadratic reciprocity law, I recommend this paper:

[1804.00199] Yet another proof of the quadratic reciprocity law (arxiv.org)

No comments:

Post a Comment

Popular Posts