1 Representation theory of finite groups

1.1 Definitions and first examples

Let k be a field (we will almost always take k=), and let G be a group.

Definition 1.1.

A representation of G over k is a pair (ρ,V) where the

  • V is a vector space over k, and

  • ρ:GGL(V) is a group homomorphism.

The dimension of the representation is the dimension of V. We will very often say that V is a representation of G, or that ρ is a representation of G, without mentioning the other part of the definition.

There is another way to think of this. Suppose that (ρ,V) is a representation of G. Then we define an action of G on V by gv=ρ(g)v. This is an action because ρ is a homomorphism, and it is linear, meaning that for every g the map taking v to gv is a linear map. Conversely, given a linear action of G on V, we can define ρ by ρ(g)v=gv. In other words:

A representation of G is a linear action on a vector space.

We will often use ρ(g)v and gv interchangeably.

If a basis is given for V, then an (invertible) linear map VV is just the same thing as an (invertible) n×n matrix, where n=dimV. So, once you choose a basis, a representation is just the same as a homomorphism GGLn(k). In particular:

A one-dimensional representation of G is the same as a homomorphism Gk×.

Example 1.2.

Let G=Sn. Recall that there is a homomorphism

ϵ:Sn{±1}

taking a permutation to its sign. As ±1×, this gives a one-dimensional representation of Sn called the sign representation.

Example 1.3.

If V is any vector space, then we can always take ρ:GGL(V) to be the homomorphism sending every element to the identity. We call this the trivial representation on V.

Example 1.4.

Suppose that G=(,+). Then, if ρ is a representation of G, it is completely determined by V and the invertible linear map ρ(1):VV (which can be anything). This is because we then have

ρ(n)=ρ(1++1)=ρ(1)n.

Thus a representation of is just a vector space V together with an invertible linear map from V to itself.

We can push this a bit further. Suppose that G is a cyclic group of order n with generator g, so

G={e,g,g2,,gn-1}

and gn=e. Then a representation of G is again determined by V and ρ(g), which can be any linear map T:VV such that Tn=I.

One source of more interesting examples is geometry.

Example 1.5.

Let G=Dn be the dihedral group of order 2n, the group of symmetries (rotations and reflections) of a regular n-gon — see Figure 1. Since each rotation/reflection is an invertible linear map from 22, we get a representation ρ of G on 2. Letting r be rotation by 2π/n and s be reflection in the vertical axis, Dn has the presentation

r,s:rn=s2=e,sr=r-1s.

As an explicit homomorphism ρ:DnGL2(), we have (with θ=2π/n)

ρ(r) =(cos(θ)-sin(θ)sin(θ)cos(θ))
ρ(s) =(-1001).
The action of the dihedral group on a polygon.
Figure 1: The action of the dihedral group on a polygon.
Example 1.6.

Let G=S4. You might remember that this is isomorphic to the group of symmetries (rotations/reflections) of the regular tetrahedron in 3. We therefore get a representation

ρ:S4GL3().

It would be a slightly unpleasant exercise to work the matrices out explicitly.

Note that S4 is also isomorphic to the group of rotations of the cube, giving another (different!) three-dimensional representation.

Another source of representations comes from actions of groups on (usually finite) sets.

Example 1.7.

Let G=Sn, and let V=kn. Define a representation of Sn on V via

σ(x1e1++xnen)=x1eσ(1)++xneσ(n)

where e1,,en is the standard basis. This is called the permutation representation (over k).

There is a warning here! If you write elements of V as (x1,,xn), as well you might, then it is not the case that g(x1,,xn)=(xg(1),,xg(n)). This actually would define a right action, not a left action. The correct formula is

g(x1,,xn)=(xg-1(1),,xg-1(n)).
Definition 1.8.

If G is a group acting on a set X, we consider a k-vector space V with basis {ex:xX}. It has a representation of G given by gex=egx, called the permutation representation associated to X.

Remark 1.9.

There is another point of view on this, the functional point of view. We didn’t cover this in lectures; as things stand this is optional. For simplicity, let X be finite, and define

kX={f:Xk}.

We give this an action of G by the formula

(gf)(x)=f(g-1(x))

for gG, fkX, xX.

Question 1.10.

Why don’t we define (gf)(x)=f(gx)?

Let δxkX be the function sending x to 1 and everything else to 0. Then the δx for xX are a basis for kX, and for gG you can check that

gδx=δgx.

In other words, the δx behave exactly like the ex in the permutation representation; when we have a bit more language, we can say that kX is isomorphic to the permutation representation.

The next example combines geometry and combinatorics.

Example 1.11.

Let G be the group of rotations of the cube and let X be the set of faces. We can think of an element of the permutation representation as being a way of writing a complex number on each face.

This term is all about finite groups, but next term we will consider representations of Lie groups such as the group SO(3) of rotations of 3. These have many applications in physics.

Example 1.12.

In a spherically symmetric situation we might be interested in (smooth) solutions f:3 to Laplace’s equation

f=0

with radial behaviour f(rx)=rlf(x) for scalars r, and some integer l0. Here is the Laplacian

2x2+2y2+2z2.

These form a representation of SO(3) of dimension 2l+1, and in fact the irreducible finite dimensional representations of SO(3) are exactly those obtained in this way for integers l0.

The solutions are called spherical harmonics, and representation theory can be used to find particularly nice bases of these spaces!

Here is another example:

Example 1.13.

Let G=GL2() and let V=X2,XY,Y2[X,Y]. Then V is a representation of G via

gF(X,Y)=F(aX+cY,bX+dY).