Exercise 15.4.2

Let α [ i ] be nonzero. The goal of this exercise is to prove part (a) of Lemma 15.4.2, which asserts that | [ i ] αℤ [ i ] | = N ( α ) . The idea is to forget multiplication and think of [ i ] and [ i ] αℤ [ i ] as groups under addition. let m be the greatest common divisor of the real and imaginary parts of α , so that α = m ( a + bi ) , where gcd ( a , b ) = 1 . Then pick c , d such that ad bc = 1 .

(a)
Show that the map [ i ] defined by μ + νi μ ( d , b ) + ν ( c , a ) = ( μd νc , μb + νa )

is a group isomorphism under addition.

(b)
Show that the map of part (a) takes α and to ( m , 0 ) and ( m ( ac + bd ) , m ( a 2 + b 2 ) ) , respectively. Then use this to show that the map takes αℤ [ i ] [ i ] to the subgroup mℤ m ( a 2 + b 2 ) .

(c)
Use part (b) to conclude that | [ i ] αℤ [ i ] | = N ( α ) .

Answers

Proof. (a) Consider

ψ { [ i ] μ + νi μ ( d , b ) + ν ( c , a ) = ( μd νc , μb + νa ) .

We verify that φ is a group homomorphism: if z = μ + νi , z = μ + ν i [ i ] , then

ψ ( z + z ) = ψ ( ( μ + μ ) + i ( ν + ν ) = ( μ + μ ) ( d , b ) + ( ν + ν ) ( c , a ) = [ μ ( d , b ) + ν ( c , a ) ] + [ μ ( d , b ) + ν ( c , a ) ] = ψ ( z ) + ψ ( z ) .

Let ( u , v ) be any element of . For all μ + [ i ] , since ad bc = 1 ,

( u , v ) = ψ ( μ + ) { μd νc = u , μb + νa = v , { μ ( ad bc ) = au + cv , ν ( ad bc ) = bu + dv , { μ = au + cv , ν = bu + dv .

Conversely, if μ = au + cv , ν = bu + dv , then

{ μd νc = ( au + cv ) d ( bu + dv ) c = u ( ad bc ) = u , μb + νa = ( au + cv ) b + ( bu + dv ) a = v ( ad bc ) = v .

We have proved, for all z [ i ] , for all ( u , v ) × , that

( u , v ) = ψ ( z ) z = ( au + cv ) + i ( bu + dv )

This shows that ψ is bijective, and for all ( u , v ) ,

ψ 1 ( u , v ) = ( au + cv ) + i ( bu + dv ) = ( a + ib ) u + ( c + id ) v .

To conclude, ψ is a group isomorphism. (b) We compute the images of α and by the homomorphism ψ :

ψ ( α ) = ψ ( ma + mbi ) = ma ( d , b ) + mb ( c , a ) = ( m ( ad bc ) , 0 ) = ( m , 0 ) , ψ ( ) = ψ ( mb , ma ) = mb ( d , b ) + ma ( c , a ) = ( m ( ac + bd ) , m ( a 2 + b 2 ) ) .

( α , ) is a -basis of αℤ [ i ] , i.e. every element of αℤ [ i ] writes uniquely as a linear combination of α , with integer coefficients. Moreover ψ ( α ) = ( m , 0 ) mℤ × m ( a 2 + b 2 ) , and ψ ( ) = ( m ( ac + bd ) , m ( a 2 + b 2 ) ) mℤ × ( a 2 + b 2 ) = mℤ ( a 2 + b 2 ) , therefore

ψ ( αℤ [ i ] ) mℤ ( a 2 + b 2 ) .

Conversely, let ( u , v ) be any element of mℤ × ( a 2 + b 2 ) . There are some λ , μ such that

u = λm , v = μ ( a 2 + b 2 ) .

Using the formula which gives ψ 1 ( u , v ) in part (a), we obtain

ψ 1 ( u , v ) = ( a + ib ) λm + ( c + id ) μ ( a 2 + b 2 ) = ( a + ib ) [ λm + μ ( c + id ) ( a ib ) ] ,

thus ψ 1 ( u , v ) αℤ [ i ] , and ( u , v ) ψ ( αℤ [ i ] ) . This proves mℤ ( a 2 + b 2 ) ψ ( αℤ [ i ] ) , thus

ψ ( αℤ [ i ] ) = mℤ ( a 2 + b 2 ) .

(c) If A , B are Abelian groups, and I , J are subgroups of A , B respectively, the surjective homomorphism

A × B A I × B J ( a , b ) A I × B J

has kernel I × J , so that

( A × B ) ( I × J ) A I × B J .

This general property gives here

( ) ( mℤ ( a 2 + b 2 ) ) ( mℤ ) × ( ( a 2 + b 2 ) .

Since the isomorphism ψ maps [ i ] on , and αℤ [ i ] on mℤ ( a 2 + b 2 ) , this implies

[ i ] αℤ [ i ] ( ) ( mℤ ( a 2 + b 2 ) ) mℤ × ( a 2 + b 2 ) .

Therefore

| [ i ] αℤ [ i ] | = | mℤ | | ( a 2 + b 2 ) | = m ( a 2 + b 2 ) = N ( α ) .

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