Exercise 15.4.8

Consider the finite ring [ i ] 2 ( 1 + i ) [ i ] , and let β [ i ] be odd.

(a)
Prove that ( [ i ] 2 ( 1 + i ) [ i ] ) = { ± [ 1 ] , ± [ i ] } , and then explain why this implies that β i 𝜀 ( mod 2 ( 1 + i ) ) for some 𝜀 { 0 , 1 , 2 , 3 } .
(b)
Prove that φ ( β ϖ 2 ) = i 𝜀 .

Answers

Proof. (a) By Exercise 1, | [ i ] 2 ( 1 + i ) [ i ] | = N ( 2 ( 1 + i ) ) = 8 . We want to find a complete system of representatives of the Gaussian integers modulo 2 ( 1 + i ) .

Note first that 2 i 2 ( mod 2 ( 1 + i ) ) , and 4 = ( 1 i ) [ 2 ( 1 + i ) ] 0 ( mod 2 ( 1 + i ) ) .

Let z = a + ib be any Gaussian integer. The Euclidean division in gives b = 2 q + r where r { 0 , 1 } . Then

z = a + ( 2 q + r ) i a 2 q + ri c + ri ( mod 2 ( 1 + i ) ) ,

where c , r { 0 , 1 } . If we write c = 4 q + s , 0 s < 4 , then

z s + ri ( mod 2 ( 1 + i ) ) , s { 0 , 1 , 2 , 3 } , r { 0 , 1 } .

Two distinct Gaussian integers among the set { s + ri s { 0 , 1 , 2 , 3 } , r { 0 , 1 } } are not congruent modulo 2 ( 1 + i ) , otherwise | [ i ] 2 ( 1 + i ) [ i ] | < 8 . Thus

[ i ] 2 ( 1 + i ) [ i ] = { [ 0 ] , [ 1 ] , [ 2 ] , [ 3 ] , [ i ] , [ 1 + i ] , [ 2 + i ] , [ 3 + i ] } = { [ 0 ] , [ 1 ] , [ 2 ] , [ 1 ] , [ i ] , [ 1 + i ] , [ i ] , [ 1 + i ] } ,

since 3 1 and 2 + i = i + ( 2 + 2 i ) i ( mod 2 ( 1 + i ) ) .

A coset [ α ] , where α [ i ] , is invertible in [ i ] 2 ( 1 + i ) [ i ] if and only if α and 2 ( 1 + i ) are relatively prime. Since 2 ( 1 + i ) = i ( 1 + i ) 3 , where i is a unit, and 1 + i is prime. Thus α and 2 ( 1 + i ) are relatively prime if and only if α and 1 + i are relatively prime. By Section A about Gaussian integers, this is equivalent to α is odd. Therefore

( [ i ] 2 ( 1 + i ) [ i ] ) = { ± [ 1 ] , ± [ i ] } .

Since β is odd, β and 1 + i are relatively prime. By the preceding reasoning, β ( [ i ] 2 ( 1 + i ) [ i ] ) , therefore

β { ± [ 1 ] , ± [ i ] } .

Thus β = i 𝜀 , 𝜀 { 0 , 1 , 2 , 3 } . (b) By (15.44), when β is odd,

φ ( βz ) = i 𝜀 φ ( z ) P β ( φ 4 ( z ) ) Q β ( φ 4 ( z ) ) .

Then

φ ( β ϖ 2 ) = i 𝜀 φ ( ϖ 2 ) P β ( φ 4 ( ϖ 2 ) ) Q β ( φ 4 ( ϖ 2 ) ) .

We know from Section (15.1) that φ ( ϖ 2 ) = 1 . Moreover Theorem 15.4.4 shows that Q β ( u ) = u d P β ( 1 u ) . Taking u = 1 , we obtain P ( 1 ) = Q ( 1 ) , therefore

φ ( β ϖ 2 ) = i 𝜀 .

User profile picture
2022-07-19 00:00
Comments