1.3. Lecture 3

One of the main goals of representation theory is to classify all the irreducible representations of a group GG (or more correctly, the isomorphism classes of irreducible representations). This entails producing a list of representations of GG and showing that

  1. (i)
    ​

    any irreducible representation of GG is isomorphic to some entry of the list, and

  2. (ii)
    ​

    there are no redundancies in the list, i.e all entries are non-isomorphic.

1.3.1. Example: dihedral groups

We list the elements of the dihedral group DnD_{n} as

{rk,s⁢rk|k=0,…,n−1}.\{r^{k},sr^{k}\,|\,k=0,\ldots,n-1\}.

We aim to show that Table 1.1 gives the complete list of irreducible representations of DnD_{n}, for nn odd. We leave the case of nn even as an exercise (there are two more 1-dimensional representations in this case).

Table 1.1. Representations of DnD_{n}.
Dimension ρ⁢(r)\rho(r) ρ⁢(s)\rho(s)
(Id,ℂ)(\mathrm{Id},{\mathbb{C}}) 11 11 11
(ϵ,ℂ)(\epsilon,{\mathbb{C}}) 11 11 −1-1
ρk1≤k<n/2\underset{1\leq k<n/2}{\rho_{k}} 22 (e2⁢π⁢i⁢kn00e−2⁢π⁢i⁢kn)\left(\begin{smallmatrix}e^{\frac{2\pi ik}{n}}&0\\ 0&e^{\frac{-2\pi ik}{n}}\end{smallmatrix}\right) (0110)\left(\begin{smallmatrix}0&1\\ 1&0\end{smallmatrix}\right)

We start by noting the following:

Proposition 1.3.1.

Let (π,V)(\pi,V) be an irreducible representation of a finite group GG. Then dim⁡π≤|G|\operatorname{dim}\pi\leq|G|; in particular, VV is finite-dimensional.

Proof.

Let 𝐯∈V{\bf v}\in V be a non-zero vector. Consider the subspace U=span⁢{π⁢(g)⁢𝐯|g∈G}⊆VU=\mathrm{span}\{\pi(g){\bf v}\,|\,g\in G\}\subseteq V. Since UU is spanned by the vectors π⁢(g)⁢𝐯\pi(g){\bf v} as gg runs over all elements of GG, we have 1≤dim⁡U≤|G|1\leq\operatorname{dim}U\leq|G|. We claim that (π,U)(\pi,U) is a subrepresentation of (π,V)(\pi,V). Given 𝐮∈U{\bf u}\in U, we may write it as a linear combination of the spanning elements, i.e.

𝐮=∑h∈Gzh⁢π⁢(h)⁢𝐯,{\bf u}=\sum_{h\in G}z_{h}\pi(h){\bf v},

where each zh∈ℂz_{h}\in{\mathbb{C}}. Given any g∈Gg\in G, we then have

π⁢(g)⁢𝐮=π⁢(g)⁢(∑h∈Gzh⁢π⁢(h)⁢𝐯)=∑h∈Gzh⁢π⁢(g)⁢π⁢(h)⁢𝐯=∑h∈Gzh⁢π⁢(g⁢h)⁢𝐯=∑h∈Gzg−1⁢h⁢π⁢(h)⁢𝐯.\pi(g){\bf u}=\pi(g)\left(\sum_{h\in G}z_{h}\pi(h){\bf v}\right)=\sum_{h\in G}% z_{h}\pi(g)\pi(h){\bf v}=\sum_{h\in G}z_{h}\pi(gh){\bf v}=\sum_{h\in G}z_{g^{-% 1}h}\pi(h){\bf v}.

The vector ∑h∈Gzg−1⁢h⁢π⁢(h)⁢𝐯\sum_{h\in G}z_{g^{-1}h}\pi(h){\bf v} is a linear combination of vectors in UU, and so is also in UU. The subspace UU is therefore GG-invariant, and hence a subrepresentation. Since (π,V)(\pi,V) is irreducible and U≠0U\neq 0, V=UV=U. ∎

Theorem 1.3.2.

Let nn be odd. Then Table 1.1 is a complete list of non-isomorphic irreducible representations of DnD_{n}.

Proof.

Firstly, we observe that the matrices in the table satisfy the group relations for DnD_{n}, and so do in fact define representations of DnD_{n}.

Now let (π,V)(\pi,V) to be an irreducible (complex) representation of DnD_{n}. Since VV is finite-dimensional, by Proposition 1.3.1, π⁢(r)\pi(r) has an eigenvector 𝐯∈V{\bf v}\in V with eigenvalue λ\lambda. Note that 𝐯=π⁢(e)⁢𝐯=π⁢(rn)⁢𝐯=π⁢(r)n⁢𝐯=λn⁢𝐯{\bf v}=\pi(e){\bf v}=\pi(r^{n}){\bf v}=\pi(r)^{n}{\bf v}=\lambda^{n}{\bf v}; λ\lambda is therefore an nn-th root of unity.

Consider the vector 𝐰=π⁢(s)⁢𝐯{\bf w}=\pi(s){\bf v}. The key calculation is:

π⁢(r)⁢𝐰=π⁢(r)⁢π⁢(s)⁢𝐯=π⁢(r⁢s)⁢𝐯=π⁢(s⁢r−1)⁢𝐯=π⁢(s)⁢π⁢(r)−1⁢𝐯=π⁢(s)⁢(λ−1⁢𝐯)=λ−1⁢𝐰.\pi(r){\bf w}=\pi(r)\pi(s){\bf v}\\ =\pi(rs){\bf v}=\pi(sr^{-1}){\bf v}=\pi(s)\pi(r)^{-1}{\bf v}=\pi(s)(\lambda^{-% 1}{\bf v})=\lambda^{-1}{\bf w}.

We also have π⁢(s)⁢𝐰=π⁢(s)2⁢𝐯=𝐯\pi(s){\bf w}=\pi(s)^{2}{\bf v}={\bf v} and so span⁢{𝐯,𝐰}\mathrm{span}\{{\bf v},{\bf w}\} is a subrepresentation of VV. As VV is irreducible, we see that V=span⁢{𝐯,𝐰}V=\mathrm{span}\{{\bf v},{\bf w}\}.

  • Case 1:
    ​

    Suppose that λ≠λ−1\lambda\neq\lambda^{-1}. Then 𝐯{\bf v} and 𝐰{\bf w} are eigenvectors of π⁢(r)\pi(r) with distinct eigenvalues, and so are linearly independent. Thus dim⁡V=2\operatorname{dim}V=2. In the basis 𝐯,𝐰{\bf v},{\bf w}, the representation is

    π⁢(r)\displaystyle\pi(r) =(λ00λ−1)\displaystyle=\left(\begin{matrix}\lambda&0\\ 0&\lambda^{-1}\end{matrix}\right)
    π⁢(s)\displaystyle\pi(s) =(0110).\displaystyle=\left(\begin{matrix}0&1\\ 1&0\end{matrix}\right).

    If λ=e2⁢π⁢i⁢k/n\lambda=e^{2\pi ik/n} for some 1≤k<n/21\leq k<n/2, then we get the representation ρk\rho_{k}. Otherwise, λ=e−2⁢π⁢i⁢k/n\lambda=e^{-2\pi ik/n} for some 1≤k<n/21\leq k<n/2 and we instead take the basis 𝐰,𝐯{\bf w},{\bf v} to get ρk\rho_{k} again.

  • Case 2:
    ​

    Suppose that λ=λ−1\lambda=\lambda^{-1}. Then λ=1\lambda=1 as nn is odd. Since

    π⁢(r)⁢(𝐯+𝐰)=π⁢(s)⁢(𝐯+𝐰)=𝐯+𝐰,\pi(r)({\bf v}+{\bf w})=\pi(s)({\bf v}+{\bf w})={\bf v}+{\bf w},

    we see that 𝐯+𝐰{\bf v}+{\bf w} spans a subrepresentation of VV. If 𝐯+𝐰≠0{\bf v}+{\bf w}\neq 0, then V=ℂ⁢(𝐯+𝐰)V={\mathbb{C}}({\bf v}+{\bf w}) is the trivial representation. Otherwise, π⁢(s)⁢𝐯=𝐰=−𝐯\pi(s){\bf v}={\bf w}=-{\bf v}, and we get the representation ϵ\epsilon.

    It remains to show that the representations in the table are non-isomorphic; this is left as Exercise 1.3.2 below.

∎

1.3.2. Exercises

..

Problem 25.

  1. (a)
    ​

    Show that if (π,V)(\pi,V) and (ρ,W)(\rho,W) are two isomorphic finite-dimensional representations of a group GG, then π⁢(g)\pi(g) and ρ⁢(g)\rho(g) have the same eigenvalues for all g∈Gg\in G.

  2. (b)
    ​

    Complete the proof of Theorem 1.3.2.

Problem 26. Find all the irreducible representations of DnD_{n} for nn even.

Problem 27. Show that the irreducible representations of S3S_{3} consist of (a) The 1-dimensional trivial representation (b) the sign representation (sgn,ℂ)(\mathrm{sgn},{\mathbb{C}}), where sgn⁢(σ)∈{±1}\mathrm{sgn}(\sigma)\in\{\pm 1\} is the sign of the permutation σ\sigma, and (c) the representation (π,W0)(\pi,W_{0}), where π\pi is the usual permutation representation of S3S_{3} on ℂ3{\mathbb{C}}^{3}, and W0={(z1,z2,z3)∈ℂ3|z1+z2+z3=0}W_{0}=\{(z_{1},z_{2},z_{3})\in{\mathbb{C}}^{3}\,|\,z_{1}+z_{2}+z_{3}=0\}. Hint: use the fact that we have a complete classification of the irreducible representations of DnD_{n}.

Problem 28. Classify the irreducible representations of Q8Q_{8}. Hint: adapt the strategy we used for DnD_{n}.