CHARACTERS OF FINITE ABELIAN GROUPS (SHORT VERSION)

KEITH CONRAD

1. Introduction
The theme we will study is an analogue on finite abelian groups of Fourier analysis on
R. A Fourier series on the real line is the following type of series in sines and cosines:
X X
f (x) = an cos(nx) + bn sin(nx).
n≥0 n≥1

This is 2π-periodic. Since einx = cos(nx) + i sin(nx) and e−inx = cos(nx) − i sin(nx), a
Fourier series can also be written in terms of complex exponentials:
X
f (x) = cn einx ,
n∈Z

where the summation runs over all integers (c0 = a0 , cn = 21 (an − bn i) for n > 0, and
cn = 12 (a|n| + b|n| i) for n < 0). The convenient algebraic property of einx , which is not
shared by sines and cosines, is that it is a group homomorphism from R to the unit circle
S 1 = {z ∈ C : |z| = 1}:
0 0
ein(x+x ) = einx einx .
We now replace the real line R with a finite abelian group. Here is the analogue of the
functions einx .
Definition 1.1. A character of a finite abelian group G is a homomorphism χ : G → S 1 .
We will usually write abstract groups multiplicatively, so χ(g1 g2 ) = χ(g1 )χ(g2 ) and
χ(1) = 1.
Example 1.2. The trivial character of G is the homomorphism 1G defined by 1G (g) = 1
for all g ∈ G.
Example 1.3. Let G be cyclic of order 4 with generator γ. Since γ 4 = 1, a character χ of
G has χ(γ)4 = 1, so χ takes only four possible values at γ, namely 1, −1, i, or −i. Once
χ(γ) is known, the value of χ elsewhere is determined by multiplicativity: χ(γ j ) = χ(γ)j .
So we get four characters, whose values can be placed in a table. See Table 1.

1 γ γ2 γ3
1G 1 1 1 1
χ1 1 −1 1 −1
χ2 1 i −1 −i
χ3 1 −i −1 i
Table 1.

1
2 KEITH CONRAD

When G has size n and g ∈ G, for any character χ of G we have χ(g)n = χ(g n ) = χ(1) = 1,
so the values of χ lie among the nth roots of unity in S 1 . More precisely, the order of χ(g)
divides the order of g (which divides |G|).
Characters on finite abelian groups were first studied in number theory, since number
theory is a source of many interesting finite abelian groups. For instance, Dirichlet used
characters of the group (Z/(m))× to prove that when (a, m) = 1 there are infinitely many
primes p ≡ a mod m. The quadratic reciprocity law of elementary number theory is con-
cerned with a deep property of a particular character, the Legendre symbol. Fourier series
on finite abelian groups have applications in engineering: signal processing (the fast Fourier
transform [1, Chap. 9]) and error-correcting codes [1, Chap. 11].
To provide a context against which our development of characters on finite abelian groups
can be compared, Section 2 discusses classical Fourier analysis on the real line. In Section
3 we will run through some properties of characters of finite abelian groups and introduce
their dual groups. Section 4 uses characters of a finite abelian group to develop a finite
analogue of Fourier series.
Our notation is completely standard, but we make two remarks about it. For a complex-
valued function f (x), the complex-conjugate function is usually denoted f (x) instead of
f (x) to stress that conjugation creates a new function. (We sometimes use the overline
notation also to mean the reduction g into a quotient group.) For n ≥ 1, we write µn for
the group of nth roots of unity in the unit circle S 1 . It is a cyclic group of size n.
Exercises.

1. Make a character table for Z/(2) × Z/(2), with columns labeled by elements of the
group and rows labeled by characters, as in Table 1.
2. Let G be a finite nonabelian simple group. (Examples include An for n ≥ 5.) Show
the only group homomorphism χ : G → S 1 is the trivial map.

2. Classical Fourier analysis

This section on Fourier analysis on R serves as motivation for our later treatment of
finite abelian groups, where there will be no delicate convergence issues (just finite sums!),
so we take a soft approach and sidestep the analytic technicalities that a serious treatment
of Fourier analysis on R would demand.
Fourier analysis for periodic functions on R is based on the functions einx for n ∈ Z. Any
“reasonably nice” function f : R → C that has period 2π can be expanded into a Fourier
series X
f (x) = cn einx ,
n∈Z
where the sum runs over Z and the nth Fourier coefficient cn can be recovered as an integral:
Z 2π
1
(2.1) cn = f (x)e−inx dx.
2π 0
This formula for cn can be explained by replacing f (x) in (2.1) by its Fourier series and in-
tegrating termwise (for “reasonably nice” functions this termwise integration is analytically
justifiable), using the formula
Z 2π (
1 1, if m = n,
eimx e−inx dx =
2π 0 0, if m 6= n.
 CHARACTERS OF FINITE ABELIAN GROUPS (SHORT VERSION) 3

Rather than working with functions f : R → C having period 2π, formulas look cleaner
using functions f : R → C having period 1. The basic exponentials become e2πinx and the
Fourier series and coefficients for f are
X Z 1
(2.2) f (x) = cn e 2πinx
, cn = f (x)e−2πinx dx.
n∈Z 0

Note cn in (2.2) is not the same as cn in (2.1).

In addition to Fourier series there are Fourier integrals. The Fourier transform of a
function f that decays rapidly at ±∞ is the function fb: R → C defined by the integral
formula Z
fb(y) = f (x)e−2πixy dx.
R
The analogue of the expansion (2.2) of a periodic function into a Fourier series is the Fourier
inversion formula, which expresses f in terms of its Fourier transform fb:
Z
f (x) = fb(y)e2πixy dy.
R
2
Example 2.1. √ A Gaussian is a function of the form ae−bx , where b > 0. For example, the
2
Gaussian (1/ 2π)e−(1/2)x is important in probability theory. The Fourier transform of a
Gaussian is another Gaussian:
Z r
−bx2 −2πixy π −π2 y2 /b
(2.3) ae e dx = ae .
R b
This formula shows that a highly peaked Gaussian (large b) has a Fourier transform that is
a spread out Gaussian (small π 2 /b) and vice versa. More generally, there is a sense in which
a function and its Fourier transform can’t both be highly localized; this is a mathematical
incarnation of Heisenberg’s uncertainty principle from physics.
There are several conventions for where 2π appears in the Fourier transform. Table 2
collects three different 2π-conventions. The first column of Table 2 is a definition and the
second column is a theorem (Fourier inversion).

fb(y) f (x)
−2πixy dx 2πixy dy
R R
R f (x)e R f (y)e
b
−ixy dx 1 ixy dy
R R
R f (x)e 2π R f (y)e
b
√1 f (x)e−ixy dx √1 fb(y)eixy dy
R R
2π R 2π R
Table 2.

A link between Fourier series and Fourier integrals is the Poisson summation formula:
for a “nice” function f : R → C that decays rapidly enough at ±∞,
X X
(2.4) f (n) = fb(n),
n∈Z n∈Z
4 KEITH CONRAD

2
where fb(y) = R f (x)e−2πixy dx. For example, when f (x) = e−bx (with b > 0), the Poisson
R
summation formula says r
X π
−bn2 2 2
X
e = e−π n /b ,
b
n∈Z n∈Z
To prove the Poisson summation formula, we use Fourier series. Periodize f (x) as
X
F (x) = f (x + n).
n∈Z

Since F (x + 1) = F (x), write F as a Fourier series: F (x) = n∈Z cn e2πinx . Then

P
Z 1
cn = F (x)e−2πinx dx
0
Z 1 X !
= f (x + m) e−2πinx dx
0 m∈Z
XZ 1
= f (x + m)e−2πinx dx
m∈Z 0
XZ m+1
= f (x)e−2πinx dx
m∈Z m
Z
= f (x)e−2πinx dx
R
= fb(n).
Therefore the expansion of F (x) into a Fourier series is equivalent to
X X
(2.5) f (x + n) = fb(n)e2πinx ,
n∈Z n∈Z

which becomes the Poisson summation formula (2.4) by setting x = 0.

Exercises.

1. Without dwelling on analytic subtleties, check from Fourier inversion that fb(x) =
b
f (−x) (if the Fourier transform is defined suitably).
2. For a function f : R → C and c ∈ R, let g(x) = f (x + c). Define the Fourier
h(y) = R h(x)e−2πixy dx. If f has a Fourier transform,
R
transform of a function h by b
show g has Fourier transform gb(y) = e2πicy fb(y).
3. Assuming the Fourier inversion formula holds for a definition of the Fourier trans-
form as in Table 2, check that for all α and β in R× that if we set
Z
(Ff )(y) = α f (x)e−iβxy dx
R
for all x then Z
β
f (x) = (Ff )(y)eiβxy dy.
2πα R
(If β = 2πα2 then these two equations are symmetric in the roles of f and Ff except
for a sign in the exponential term.)
 CHARACTERS OF FINITE ABELIAN GROUPS (SHORT VERSION) 5

3. Finite Abelian Group Characters

We leave the real line and turn to the setting of finite abelian groups G. Our interest shifts
from the functions einx to characters: homomorphisms from G → S 1 . The construction of
characters of these groups begins with the case of cyclic groups.

Theorem 3.1. Let G be a finite cyclic group of size n with a chosen generator γ. There
are exactly n characters of G, each determined by sending γ to the different nth roots of
unity in C.

Proof. We mimic Example 1.3, where G is cyclic of size 4. Since γ generates G, a character
is determined by its value on γ and that value must be an nth root of unity (not necessarily
of exact order n, e.g., 1G (γ) = 1), so there are at most n characters. We now write down
n characters.
Let ζ be any nth root of unity in C. Set χ(γ j ) = ζ j for j ∈ Z. This formula is well-
defined (if γ j = γ k for two different integer exponents j and k, we have j ≡ k mod n so
ζ j = ζ k ), and χ is a homomorphism. Of course χ depends on ζ. As ζ changes, we get
different characters (their values at γ are changing), so in total we have n characters. 

To handle characters of non-cyclic groups, the following lemma is critical.

Lemma 3.2. Let G be a finite abelian group and H ⊂ G be a subgroup. Any character of
H can be extended to a character of G in [G : H] ways.

Proof. We will induct on the index [G : H] and we may suppose H 6= G. Pick a ∈ G with
a 6∈ H, so
H ⊂ hH, ai ⊂ G.
Let χ : H → S 1 be a character of H. We will extend χ to a character χ e of hH, ai and count
the number of possible χ e. Then we will use induction to lift characters further from hH, ai
all the way up to G.
What is a viable choice for χ e(a)? Since a 6∈ H, χe(a) is not initially defined. But some
k
power a is in H for k ≥ 1 (e.g., k = [G : H]), and therefore χe(ak ) is defined: χ
e(ak ) = χ(ak ).
k
Pick k ≥ 1 minimal with a ∈ H. That is, k is the order of a in G/H, so k = [hH, ai : H].
If χ
e is going to be a character then χ e(a) must be an k-th root of χ(ak ). That is our clue:
define χe(a) ∈ S 1 to be a solution to z k = χ(ak ):

(3.1) e(a)k = χ(ak ).

χ

Every number in S 1 has k different k-th roots in S 1 , so there are k potential choices for
χ
e(a). We will show they all work.
Once we have chosen χ e on hH, ai by
e(a) to satisfy (3.1), define χ

e(hai ) := χ(h)e
χ χ(a)i .

This formula does cover all possible elements of hH, ai, but is χ e well-defined? Perhaps H
and hai overlap nontrivially, so the expression of an element of hH, ai in the form hai is not
unique. We have to show this doesn’t lead to an inconsistency in the value of χ e. Suppose
0 0
hai = h0 ai . Then ai−i ∈ H, so i0 ≡ i mod k since k is denoting the order of a in G/H.
6 KEITH CONRAD

0
Write i0 = i + kq, so h = h0 ai −i = h0 akq . The terms h, h0 , and ak are in H, so
0
χ(h0 )e
χ(a)i = χ(h0 )e
χ(a)i χ
e(a)kq
= χ(h0 )e
χ(a)i χ(ak )q since χ
e(a)k = χ(ak )
= χ(h0 akq )e
χ(a)i
χ(a)i .
= χ(h)e
Therefore χe : hH, ai → S 1 is a well-defined function and it is easily checked to be a homo-
morphism. It restricts to χ on H. The number of choices of χ e extending χ is the number
of choices for χe(a), which is k = [hH, ai : H]. Since [G : hH, ai] < [G : H], by induction on
the index there are [G : hH, ai] extensions of each χ e to a character of G, so the number of
extensions of a character on H to a character on G is [G : hH, ai][hH, ai : H] = [G : H]. 
Theorem 3.3. If g 6= 1 in a finite abelian group G then χ(g) 6= 1 for some character χ of
G. The number of characters of G is |G|.
Proof. The cyclic group hgi is nontrivial, say of size n, so n > 1. The group µn of n-th
roots of unity in S 1 is also cyclic of order n, so there is an isomorphism hgi ∼
= µn . This
isomorphism can be viewed as a character of the group hgi. By Lemma 3.2 it extends to a
character of G and does not send g to 1.
To show G has |G| characters, apply Lemma 3.2 with H the trivial subgroup. 
We have used two important features of S 1 as the target group for characters: for any
k ≥ 1 the kth power map on S 1 is k-to-1 (proof of Lemma 3.2) and for each k ≥ 1 there is
a cyclic subgroup of order k in S 1 (proof of Theorem 3.3).
Corollary 3.4. If G is a finite abelian group and g1 6= g2 in G then there is a character of
G that takes different values at g1 and g2 .
Proof. Apply Theorem 3.3 to g = g1 g2−1 . 
Corollary 3.4 shows the characters of G “separate” the elements of G: different elements
of the group admit a character taking different values on them.
Corollary 3.5. If G is a finite abelian group and H ⊂ G is a subgroup and g ∈ G with
g 6∈ H then there is a character of G that is trivial on H and not equal to 1 at g.
Proof. We work in the group G/H, where g 6= 1. By Theorem 3.3 there is a character of
G/H that is not 1 at g. Composing this character with the reduction map G → G/H yields
a character of G that is trivial on H and not equal to 1 at g. 
It is easy to find functions on G that separate elements without using characters. For
g ∈ G, define δg : G → {0, 1} by
(
1, if x = g,
(3.2) δg (x) =
0, if x 6= g.
These functions separate elements of the group, but characters do this too and have better
algebraic properties: they are group homomorphisms.
Our definition of a character makes sense on nonabelian groups, but there will not be
enough such characters for Theorem 3.3 to hold if G is finite and nonabelian: any homomor-
phism χ : G → S 1 must equal 1 on the commutator subgroup [G, G], which is a nontrivial
subgroup, so such homomorphisms can’t distinguish elements in [G, G] from each other. If
 CHARACTERS OF FINITE ABELIAN GROUPS (SHORT VERSION) 7

g 6∈ [G, G] then in the finite abelian group G/[G, G] the coset of g is nontrivial so there
is a character G/[G, G] → S 1 that’s nontrivial on g. Composing this character with the
reduction map G → G/[G, G] produces a homomorphism G → S 1 that is nontrivial on g.
Definition 3.6. For a character χ on a finite abelian group G, the conjugate character is
the function χ : G → S 1 given by χ(g) := χ(g).
Since any complex number z with |z| = 1 has z = 1/z, χ(g) = χ(g)−1 = χ(g −1 ).
Definition 3.7. The dual group of a finite abelian group G is the set of homomorphisms
G → S 1 with the group law of pointwise multiplication of functions: (χψ)(g) = χ(g)ψ(g).
The dual group of G is denoted G.
b

The trivial character of G is the identity in G

b and the inverse of a character is its conjugate
character. Note Gb is abelian since multiplication in C× is commutative.
Theorem 3.3 says in part that
(3.3) |G| = |G|.
b

In fact, the groups G and G

b are isomorphic. First let’s check this on cyclic groups.

Theorem 3.8. If G is cyclic then G ∼

=Gb as groups.

Proof. We will show G

b is cyclic. Then since G and G
b have the same size they are isomorphic.
Let n = |G| and γ be a generator of G. Set χ : G → S 1 by χ(γ j ) = e2πij/n for all j. For
any other character ψ ∈ G,b we have ψ(γ) = e2πik/n for some integer k, so ψ(γ) = χ(γ)k .
Then
ψ(γ j ) = ψ(γ)j = χ(γ)jk = χ(γ j )k ,
which shows ψ = χk . Therefore χ generates G.b 

Lemma 3.9. If A and B are finite abelian groups, there is an isomorphism A

\ ×B ∼
= A×
b B.
b

Proof. Let χ be a character on A × B. Identify the subgroups A × {1} and {1} × B of

A × B with A and B in the obvious way. Let χA and χB be the restrictions of χ to A and
B respectively, i.e., χA (a) = χ(a, 1) and χB (b) = χ(1, b). Then χA and χB are characters
of A and B and χ(a, b) = χ((a, 1)(1, b)) = χ(a, 1)χ(1, b) = χA (a)χB (b). So we get a map
(3.4) A
\ ×B →A
b×B
b

by sending χ to (χA , χB ). It is left to the reader to check (3.4) is a group homomorphism.

Its kernel is trivial since if χA and χB are trivial characters then χ(a, b) = χA (a)χB (b) = 1,
so χ is trivial. Both sides of (3.4) have the same size by (3.3), so (3.4) is an isomorphism. 

Theorem 3.10. If G is a finite abelian group then G is isomorphic to G.

b

Proof. The case when G is cyclic was Theorem 3.8. Lemma 3.9 extends easily to several
factors in a direct product:
(3.5) (H1 × · · · × Hr ) b ∼
=Hb1 × · · · × H
br .

When Hi is cyclic, H bi ∼
= Hi , so (3.5) tells us that that character group of H1 × · · · × Hr is
isomorphic to itself. Every finite abelian group is isomorphic to a direct product of cyclic
groups, so the character group of any finite abelian group is isomorphic to itself. 
8 KEITH CONRAD

Although G and G b are isomorphic groups, there is not any kind of natural isomorphism
between them, even when G is cyclic. For instance, to prove G ∼ =G b when G is cyclic we
had to choose a generator. If we change the generator, then the isomorphism changes.1
The double-dual group G b is the dual group of G.
b Since G and G b are isomorphic, G and
b

G
b are isomorphic. However, while there isn’t a natural isomorphism from G to G, b there is
b

a natural isomorphism from G to G. b The point is that there is a natural way to map G to
b
its double-dual group: associate to each g ∈ G the function “evaluate at g,” which is the
function Gb → S 1 given by χ 7→ χ(g). Here g is fixed and χ varies. This is a character of G,
b
since (χ1 χ2 )(g) = χ1 (g)χ2 (g) by definition.

Theorem 3.11. Let G be a finite abelian group. The homomorphism G → G

b associating
b
to g ∈ G the function “evaluate at g” is an isomorphism.
Proof. Since a finite abelian group and its dual group have the same size, a group and its
double-dual group have the same size, so it suffices to show this homomorphism is injective.
If g ∈ G is in the kernel then every element of Gb is 1 at g, so g = 1 by Theorem 3.3. 
Theorem 3.11 is called Pontryagin duality. This label actually applies to a more general
result about characters of locally compact abelian groups. Finite abelian groups are a
special case, where difficult analytic techniques can be replaced by counting arguments.
The isomorphism between G and its double-dual group given by Pontryagin duality lets us
think about any finite abelian group G as a dual group (namely the dual group of G).b
The isomorphism in Pontryagin duality is natural: it does not depend on any ad hoc
choices (unlike the isomorphism between a finite abelian group and its dual group).
Exercises.

1. Let’s find the characters of the additive group (Z/(m))r , an r-fold direct product.
(a) For k ∈ Z/(m), let χk : Z/(m) → S 1 by
χk (j) = e2πijk/m ,
so χk (1) = e2πik/m . Show χ0 , χ1 , . . . , χm−1 are all the characters of Z/(m) and
χk χl = χk+l .
(b) Let r ≥ 1. For r-tuples a, b in (Z/(m))r , let
a · b = a1 b1 + · · · + ar br ∈ Z/(m)
be the usual dot product. For k ∈ (Z/(m))r , let χk (j) = e2πi(j·k)/m . Show the
functions χk are all the characters of (Z/(m))r and χk χl = χk+l .
2. Show the following are equivalent properties of a character χ: χ(g) = ±1 for all g,
χ(g) = χ(g) for all g, and χ2 = 1G .
3. Describe the error in the following bogus proof of Lemma 3.2. Let m = [G : H] and
pick a set of coset representatives g1 , . . . , gm for G/H. Given a character χ on H,
define χ
e on G by first picking the m (= [G : H]) values χ e(gi ) for 1 ≤ i ≤ m and
then writing any g ∈ G in the (unique) form gi h and defining χ e(g) = χe(gi )χ(h).
This defines χb on G, and since we had to make m choices there are m characters.
1If G is trivial or of order 2, then it has a unique generator, so in that case we could say the isomorphism
G∼
=Gb is canonical.
 CHARACTERS OF FINITE ABELIAN GROUPS (SHORT VERSION) 9

4. For finite nonabelian G, show the characters of G (that is, homomorphisms G → S 1 )

separate elements modulo [G, G]: χ(g1 ) = χ(g2 ) for all χ if and only if g1 = g2 in
G/[G, G].
5. This exercise will give an interpretation of characters as eigenvectors. For a finite
abelian group G and g ∈ G, let Tg : L(G) → L(G) by (Tg f )(x) = f (gx).
(a) Show the Tg ’s are commuting linear transformations and any character of G
is an eigenvector of each Tg .
(b) If f is a simultaneous eigenvector of all the Tg ’s, show f (1) 6= 0 (if f (1) = 0
conclude f is identically zero, but the zero vector is not an eigenvector) and then
after rescaling f so f (1) = 1 deduce that f is a character of G. Thus the characters
of G are the simultaneous eigenvectors of the Tg ’s, suitably normalized.
(c) Show the Tg ’s are each diagonalizable. Deduce from this and parts (a) and
(b) that Gb is a basis of L(G), so |G|
b = dim L(G) = |G|. (This gives a different proof
that G and G b have the same size.)
6. For a subgroup H of a finite abelian group G, let
H ⊥ = {χ ∈ G
b : χ = 1 on H}.

These are the characters of G that are trivial on H. For example, G⊥ = {1G } and
{1}⊥ = G.b Note H ⊥ ⊂ G b and H ⊥ depends on H and G.
⊥
Show H is a subgroup of G, [ and G/(H
b it is isomorphic to G/H, ⊥ ) ∼ H.
b = b In
⊥
particular, |H | = [G : H].
7. Let G be finite abelian and H ⊂ G be a subgroup.
(a) Viewing H ⊥⊥ = (H ⊥ )⊥ in G using Pontryagin duality, show H ⊥⊥ = H.
(Hint: The inclusion in one direction is easy. Count sizes for the other inclusion.)
(b) Show for each m dividing |G| that
|{H ⊂ G : |H| = m} = |{H ⊂ G : [G : H] = m}|
by associating H to H ⊥ and using a (fixed) isomorphism of G with G. b
(c) For a finite abelian group G, part b says the number of subgroups of G with
index 2 is equal to the number of elements of G with order 2. Use this idea to count
the number of subgroups of (Z/(m))× with index 2. (The answer depends on the
number of odd prime factors of m and the highest power of 2 dividing m.)
(d) Show, for a prime p, that the number of subspaces of (Z/(p))n with dimension
d equals the number of subspaces with dimension n − d.
8. For a finite abelian group G, let G[n] = {g ∈ G : g n = 1} and Gn = {g n : g ∈ G}.
Both are subgroups of G. Prove G[n]⊥ = (G) b n and (Gn )⊥ = G[n]
b in G.
b

4. Finite Fourier series

Let G be a finite abelian group. Set
L(G) = {f : G → C},
the C-valued functions on G. This is a C-vector space of functions. Every f ∈ L(G) can
be expressed as a linear combination of the delta-functions δg from (3.2):
X
(4.1) f= f (g)δg .
g∈G
10 KEITH CONRAD

Indeed, evaluate both sides at each x ∈ G and we get the P same value. The functions δg
span L(G) by (4.1) and they are linearly independent: if g ag δg = 0 then evaluating the
sum at x ∈ G shows ax = 0. Thus the functions δg are a basis of L(G), so dim L(G) = |G|.
The next theorem is the first step leading to an expression for each δg as a linear combi-
nation of characters of G, which will lead to a Fourier series expansion of f . It is the first
time we add character values.
Theorem 4.1. Let G be a finite abelian group. Then
( (
X |G|, if χ = 1G , X |G|, if g = 1,
χ(g) = χ(g) =
g∈G
0, if χ 6= 1G , 0, if g 6= 1.
χ∈G b
P
Proof. Let S = g∈G χ(g). If χ is trivial on G then S = |G|. If χ is not trivial on G, say
P P
χ(g0 ) 6= 1. Then χ(g0 )S = g∈G χ(gg0 ) = g∈G χ(g) = S, so S = 0.
The second formula in the theorem can be viewed as an instance of the first formula via
Pontryagin duality: the second sum is a sum of the character “evaluate at g” over the group
G,
b and this character on G b is nontrivial when g 6= 1 by Pontryagin duality. 
Theorem 4.1 says the sum of a nontrivial character over a group vanishes and the sum
of all characters of a group evaluated at a nontrivial element vanishes, so the sum of the
elements in each row and column of a character table of G is zero except the row for the
trivial character and the column for the identity element. Check this in Table 1.
Corollary 4.2. For characters χ1 and χ2 in G b and g1 and g2 in G,
( (
X |G|, if χ1 = χ2 , X |G|, if g1 = g2 ,
χ1 (g)χ2 (g) = χ(g1 )χ(g2 ) =
g∈G
0, if χ 1 6
= χ2 , 0, if g1 6= g2 .
b χ∈G

Proof. In the first equation of Theorem 4.1 let χ = χ1 χ2 . In the second equation of Theorem
4.1 let g = g1 g2−1 . (Alternatively, after proving the first equation for all G we observe that
the second equation is a special case of the first by Pontryagin duality.) 
The equations in Corollary 4.2 are called the orthogonality relations. They say that
the character table of G has orthogonal rows and orthogonal columns when we define
orthogonality of two n-tuples of complex numbers as vanishing of their Hermitian inner
product in Cn : h(z1 , . . . , zn ), (w1 , . . . , wn )i := nk=1 zk wk .
P
By the second equation in Corollary 4.2 we can express the delta-functions in terms of
characters: X 1 X
χ(g)χ(x) = |G|δg (x) =⇒ δg (x) = χ(g)χ(x).
|G|
χ∈G
b χ∈G
b

Substituting this formula for δg into (4.1) gives

 
X 1 X
f (x) = f (g)  χ(g)χ(x)
|G|
g∈G χ∈G
b
XX 1
= f (g)χ(g)χ(x)
|G|
χ∈G
b g∈G
X
(4.2) = cχ χ(x),
χ∈G
b
 CHARACTERS OF FINITE ABELIAN GROUPS (SHORT VERSION) 11

where
1 X
(4.3) cχ = f (g)χ(g).
|G|
g∈G

The expansion (4.2) is the Fourier series for f .

Equation (4.3) is similar to the formula for the coefficient cn of einx in (2.1): an integral
over [0, 2π] divided by 2π is replaced by a sum over G divided by #G and f (x)e−inx
is replaced by f (g)χ(g). The number e−inx is the conjugate of einx , which is also the
relation between χ(g) and χ(g). Equation (4.2) shows G b is a spanning set for L(G). Since
|G| = |G| = dim L(G), G is a basis for L(G).
b b

Definition 4.3. Let G be a finite abelian group. If f ∈ L(G) then its Fourier transform is
the function fb ∈ L(G)
b given by
X
fb(χ) = f (g)χ(g).
g∈G

By (4.2) and (4.3),

1 X b
(4.4) f (x) = f (χ)χ(x).
|G|
χ∈G
b

Equation (4.4) is called the Fourier inversion formula since it tells us how to recover f from
its Fourier transform.
Remark 4.4. Classically the Fourier transform of a function R → C is another function
R → C. The finite Fourier transform, however, is defined on the dual group instead of on
the original group. We can also interpret the classical Fourier transform to be a function of
characters. For y ∈ R let χy (x) = eixy . Then χy : R → S 1 is a character and fb(y) could be
R
viewed as fb(χy ) = R f (x)χy (x) dx, so fb is a function of characters rather than of numbers.

Example 4.5. Let f = δg . Then fb(χ) = χ(g) = χ(g −1 ).

Since L(G) is spanned by both the characters of G and the delta-functions, any linear
identity in L(G) can be verified by checking it on characters or on delta-functions.
Let’s look at Fourier transforms for functions on a cyclic group. By writing a cyclic
group in the form Z/(m), we can make an isomorphism with the dual group explicit: every
character of Z/(m) has the form χk : j 7→ e2πijk/m for a unique k ∈ Z/(m) (Exercise 3.1).
The Fourier transform of a function f : Z/(m) → C can be regarded as a function not on
\ but on Z/(m):
Z/(m),
X X
(4.5) fb(k) := f (j)χk (j) = f (j)e−2πijk/m .
j∈Z/(m) j∈Z/(m)

This is similar to the classical viewpoint of the Fourier transform of a function on R as

another function of R.
Example 4.6. Let f : Z/(8) → C have the periodic values 5, 3, 1, and 1. Both f and its
Fourier transform are in Table 3. This f has frequency 2 (its period repeats twice) and the
Fourier transform vanishes except at 0, 2, 4, and 6, which are multiples of the frequency.
12 KEITH CONRAD

n 0 1 2 3 4 5 6 7
f (n) 5 3 1 1 5 3 1 1
f (n) 20 0 8 + 4i 0 4 0 8 − 4i 0
b
Table 3.

Example 4.7. Consider a function f : Z/(45) → C with the four successive repeating
values 1, 8, 19, 17 starting with f (0) = 1. It is not a periodic function on Z/(45) since 4
does not divide 45, but the sequence 1, 8, 19, 17 repeats nearly 11 times. (The value of
f (44) is 1.) A calculation of |fb(n)|, the absolute value of the Fourier transform of f , reveals
sharp peaks at n = 0, 11, 22, 23, and 34. A plot of |fb(n)| is below. The red peaks are cut
off because the lowest red bar would be around three times as tall as the highest black bar.
Peaks in |fb(n)| occur approximately at multiples of the approximate frequency!

0 11 22 23 34

As Example 4.6 suggests, the Fourier transform of a periodic function on Z/(m) knows
the frequency of the original function by the positions where the Fourier transform has
nonzero values (Exercise 4.2). For nearly periodic functions on Z/(m), the approximate
frequency is reflected in where the Fourier transform takes on its largest values. This idea
is used in Shor’s quantum algorithm for integer factorization [2], [3, Chap. 17].
Exercises.

1. Let f : Z/(8) → C take the four values a, b, c, and d twice in this order. Compute
fb(n) explicitly and determine some values for a, b, c, and d such that fb(n) is nonzero
for n = 0, 2, and 6, but fb(4) = 0.
2. Let H be a subgroup of a finite abelian group G.
(a) Suppose f : G → C is constant on H-cosets (it is H-periodic). For χ ∈ G b with
χ 6∈ H ⊥ , show fb(χ) = 0. Thus the Fourier transform of an H-periodic function on
G is supported on H ⊥ .
 CHARACTERS OF FINITE ABELIAN GROUPS (SHORT VERSION) 13

(b) If f : Z/(m) → C has period d where d | m, show fb: Z/(m) → C is supported

on the multiples of m/d. (See Example 4.6.)
3. Let G be a finite abelian group and H be a subgroup. For any function f : G → C,
Poisson summation on G says
1 X 1 X b
f (h) = f (χ),
|H| |G| ⊥
h∈H χ∈H

where H⊥ is as in Exercise 3.6. Prove this formula in two ways:

P a) Copy the classical proof sketched in Section 2 (start with the function F (x) =
h∈H f (xh), which is H-periodic so it defines a function on G/H) to obtain
1 X 1 X b
(4.6) f (xh) = f (χ)χ(x)
|H| |G| ⊥
h∈H χ∈H
for any x ∈ G and then set x = 1.
b) By linearity in f of both sides of the desired identity, verify Poisson summation
directly on the delta-functions of G. (Corollary 3.5 and Example 4.5 will be useful.)

References
[1] A. Terras, “Fourier Analysis on Finite Groups and Applications,” Cambridge Univ. Press, Cambridge,
1999.
[2] P. Shor, Polynomial-Time Algorithms for Prime Factorization and Discrete Logarithms on a Quantum
Computer, http://arxiv.org/abs/quant-ph/9508027v2.
[3] W. Trappe and L. Washington, “Introduction to Cryptography with Coding Theory,” Prentice-Hall,
Upper Saddle River, NJ 2002.