Problem 5.2.14 (Dual group)

For any group G define the dual group of G (denoted Ĝ ) to be the set of all homomorphisms from G into the multiplicative group of roots of unity in ℂ . Define a group operation in Ĝ by pointwise multiplication of functions.

(a)
Show that this operation on Ĝ makes Ĝ into an abelian group.
(b)
If G is a finite abelian group, prove that Ĝ ≃ G . [Write G as ⟨ x 1 ⟩ × ⋯ × ⟨ x r ⟩ and if n i = | x i | define χ i to be the homomorphism which sends x i to e 2 𝜋𝑖 ∕ n i and sends x j to 1 , for all j ≠ i . Prove χ i has order n i in Ĝ and Ĝ = ⟨ χ 1 ⟩ × ⋯ × ⟨ χ r ⟩ .

Answers

Proof. Let G be a finite group Ĝ = Hom ( G , ℂ ∗ ) be the set of all homomorphisms from G into the multiplicative group ℂ ∗ . Let χ ∈ Ĝ . Since every element x ∈ G has a finite order r , then x r = 1 implies χ ( x ) r = 1 , so | χ ( x ) | = 1 and χ ( r ) ∈ 𝕌 , where 𝕌 = { z ∈ ℂ ∣ | z | = 1 } is a multiplicative subgroup of ℂ ∗ , so Ĝ ≃ Hom ( G , 𝕌 ) .

(a)
We know that the set ( ℂ ∗ ) G of the applications from G in ℂ ∗ is a group for pointwise multiplication. We verify that Ĝ is a subgroup of this group
  • Ĝ ⊆ ( ℂ ∗ ) G . Moreover If χ , λ ∈ Ĝ , then for all g , h ∈ G ,

    ( 𝜒𝜆 ) ( 𝑔h ) = χ ( 𝑔h ) λ ( 𝑔h ) = χ ( g ) χ ( h ) λ ( g ) λ ( h ) = χ ( g ) λ ( g ) χ ( h ) λ ( h ) = ( 𝜒𝜆 ) ( g ) ( 𝜒𝜆 ) ( h ) ,

    so 𝜒𝜆 ∈ Ĝ .

  • Let χ be any element of Ĝ . Then the inverse of χ in the group ( ℂ ∗ ) G is χ − 1 , defined by χ − 1 ( g ) = 1 ∕ χ ( g ) ( = χ ( g ) ¯ ) for all g ∈ G . Then for all g , h ∈ G ,

    χ − 1 ( 𝑔h ) = 1 χ ( 𝑔h ) = 1 χ ( g ) χ ( h ) = 1 χ ( g ) 1 χ ( h ) = χ − 1 ( g ) χ − 1 ( h ) ,

    so χ − 1 ∈ Ĝ .

This shows that Ĝ is a subgroup of ( ℂ ∗ ) G .

(b)
Let G be an abelian group. We can write G ≃ ⟨ x 1 ⟩ × ⋯ × ⟨ x r ⟩ , where  | x i | = n i .

By Exercise 13, G has presentation

G ≃ ⟨ x 1 , … , x r ∣ x 1 n 1 = ⋯ = x r n r = 1 , x i x j = x j x i ( 1 ≤ i ≤ n , 1 ≤ j ≤ n ) ⟩ . (1)

Fix some i ∈ [ [ 1 , r ] ] . Since ( e 2 𝜋𝑖 ∕ n i ) n i = 1 , and 1 n j = 1 for j ≠ i , there exists by Exercise 13 a unique homomorphism χ i ∈ Ĝ such that

χ i ( x i ) = e 2 𝜋𝑖 ∕ n i , χ i ( x j ) = 1  if  j ≠ i .

Then χ i n i ( x i ) = ( e 2 𝜋𝑖 ∕ n i ) n i = e 2 𝜋𝑖 = 1 , and χ i n i ( x j ) = 1 if j ≠ 1 . Since G = ⟨ x 1 ⟩ × ⋯ × ⟨ x r ⟩ , we obtain

χ i n i = 1 .

Conversely, if χ i k = 1 for some integer k , then χ i k ( g ) = 1 for all g ∈ G . In particular, [ χ i ( x i ) ] k = 1 , thus e 2 𝑘𝜋𝑖 ∕ n i = 1 , so e 2 𝑘𝜋𝑖 ∕ n i = 1 , which gives 2 𝑘𝜋 ∕ n i ≡ 0 ( 𝑚𝑜𝑑 2 π ) and n i ∣ k .

So n i is the least positive integer among the integers k such that χ i k = 1 . This shows

| χ i | = n i .

Since χ i n i = 1 (where Ĝ is abelian), there exists by (1) a unique homomorphism φ : G → Ĝ such that φ ( x i ) = χ i for all i ∈ [ [ 1 , r ] ] .

Furthermore, let χ be any element of Ĝ .

Since x i n i = 1 , then χ ( x i ) n i = 1 , thus there is some integer k i such that

χ ( x i ) = e k i 2 𝑖𝜋 ∕ n i ( 1 ≤ i ≤ r ) .

If g is any element of G , then g = x 1 a 1 ⋯ x r a r , thus

χ ( g ) = χ ( x 1 ) a 1 ⋯ χ ( x r ) a r = e a 1 k 1 2 𝑖𝜋 ∕ n i ⋯ e a r k r 2 𝑖𝜋 ∕ n i = χ 1 ( g ) k 1 ⋯ χ r ( g ) a r = ( χ 1 k 1 ⋯ χ r k r ) ( g ) ,

so χ = χ 1 k 1 ⋯ χ r k r . This shows

Ĝ = ⟨ χ 1 , … , χ r ⟩ ,

and consequently φ is surjective.

Now we prove that φ is injective. Let g = x 1 a 1 ⋯ x r a r ∈ ker ⁡ ( φ ) . Then

1 = φ ( g ) = φ ( x 1 ) a 1 ⋯ φ ( x r ) a r = χ 1 a 1 ⋯ χ r a r .

Then 1 = ( χ 1 a 1 ⋯ χ r a r ) ( h ) for all ∈ G . In particular, for h = x i , we obtain for all i

1 = ( χ 1 a 1 ⋯ χ r a r ) ( x i ) = χ i ( x i ) a i = e 2 𝜋𝑖 a i ∕ n i .

Therefore n i ∣ a i , thus x i a i = 1 for all i , so g = 1 . This shows that φ is injective. Hence φ is an isomorphism, and

Ĝ ≃ G .

□
User profile picture
2026-09-09 11:08
Comments