1.2. Lecture 2

1.2.1. Subrepresentations and irreducible representations

Definition 1.2.1.

A subrepresentation of a representation (ρ,V)(\rho,V) of GG is a pair (ρ|W,W)(\rho|_{W},W) consisting of a subspace WVW\subseteq V such that ρ(g)𝐰W\rho(g){\bf w}\in W for all 𝐰W{\bf w}\in W together with the group homomorphism

ρ|W:GGL(W);ρ|W(g)𝐰:=ρ(g)𝐰,\rho|_{W}:G\longrightarrow\operatorname{GL}(W)\;;\;\rho|_{W}(g){\bf w}:=\rho(g% ){\bf w},

for all wWw\in W.

Definition 1.2.2.

A representation (ρ,V)(\rho,V) of GG is irreducible if it is non-zero and has no subrepresentations except for {0}\{0\} and VV itself.

Example 1.2.3.

Since 1-dimensional vector spaces have no non-trivial subspaces, every 1-dimensional representation is irreducible.

Example 1.2.4.

Consider the permutation representation (π,3)(\pi,{\mathbb{C}}^{3}) of S3S_{3}. This is the representation which acts on via π(σ)𝐞i=𝐞σ(i)\pi(\sigma){\bf e}_{i}={\bf e}_{\sigma(i)}, where 𝐞1,𝐞2,𝐞3{\bf e}_{1},{\bf e}_{2},{\bf e}_{3} are the standard basis vectors. It is not irreducible!

Let 𝐯1=(1,1,1)t{\bf v}_{1}=(1,1,1)\,^{\mathrm{t}}\!. Then π(σ)𝐯1=𝐯1\pi(\sigma){\bf v}_{1}={\bf v}_{1} for every σS3\sigma\in S_{3}, so (π,𝐯1)(\pi,{\mathbb{C}}{\bf v}_{1}) is a subrepresentation of (π,3)(\pi,{\mathbb{C}}^{3}). Being 1-dimensional, it is in fact an irreducible subrepresentation.

Denote by W0W_{0} the subspace of 3{\mathbb{C}}^{3} consisting of all vectors (z1,z2,z3)t(z_{1},z_{2},z_{3})\,^{\mathrm{t}}\! such that z1+z2+z3=0z_{1}+z_{2}+z_{3}=0. Then since S3S_{3} permutes the coordinates of a vector, it doesn’t change their sum; so π(σ)W0W0\pi(\sigma)W_{0}\subseteq W_{0}, and W0W_{0} is a subrepresentation (of dimension 2).

We claim that W0W_{0} is irreducible. Indeed, suppose that UW0U\subseteq W_{0} is a non-zero subrepresentation; we have to show that U=W0U=W_{0}. Let (x,y,z)tU(x,y,z)\,^{\mathrm{t}}\!\in U be non-zero. As x=y=zx=y=z can’t happen, we can apply an element of S3S_{3} to permute the coordinates so that xyx\neq y. Then applying π(12)\pi(12), we have (y,x,z)tU(y,x,z)\,^{\mathrm{t}}\!\in U. Taking the difference, (xy,yx,0)tU(x-y,y-x,0)\,^{\mathrm{t}}\!\in U; scaling, (1,1,0)tU(1,-1,0)\,^{\mathrm{t}}\!\in U. Applying π(23)\pi(23), we have (1,0,1)tU(1,0,-1)\,^{\mathrm{t}}\!\in U. But these vectors span W0W_{0} (e.g. because they are linearly independent and dimW0=2\operatorname{dim}W_{0}=2), so U=W0U=W_{0} as required.

Example 1.2.5.

If (ρ,V)(\rho,V) is a finite-dimensional representation of G=G={\mathbb{Z}}, with T=ρ(1)T=\rho(1), then TT has an eigenvector 𝐯{\bf v} which spans a 1-dimensional subrepresentation of VV. Thus VV is not irreducible unless dimV=1\operatorname{dim}V=1. The irreducible subrepresentations of VV are exactly the lines spanned by TT-eigenvectors.

1.2.2. Homomorphisms and isomorphisms

Definition 1.2.6.

Suppose that (ρ,V)(\rho,V) and (σ,W)(\sigma,W) are representations of GG. Then a GG-homomorphism (or homomorphism of representations of GG, or map of representations of GG, or, if we are being lazy, just a homomorphism) VWV\rightarrow W is a linear map ϕ:VW\phi:V\rightarrow W such that

ϕ(ρ(g)𝐯)=σ(g)ϕ(𝐯)\phi(\rho(g){\bf v})=\sigma(g)\phi({\bf v})

for all 𝐯V{\bf v}\in V, gGg\in G.

In other words, ϕ\phi “commutes” with the action of GG. We write HomG(V,W)\operatorname{Hom}_{G}(V,W) for the (vector space) of GG-homomorphisms from VV to WW.

There is another word that is sometimes used for GG-homomorphism: intertwiner, or GG-intertwiner.

Definition 1.2.7.

A GG-isomorphism (or just an isomorphism) is a bijective GG-homomorphism.

If (π,V)(\pi,V) and (ρ,W)(\rho,W) are two representations of a group GG, and there exists a GG-isomorphism VWV\rightarrow W then we write (π,V)(ρ,W)(\pi,V)\cong(\rho,W), and say that they are isomorphic. The collection of all representations isomorphic to a representation (π,V)(\pi,V) is called the isomorphism class of (π,V)(\pi,V).

Lemma 1.2.8.

Suppose that VV and WW are representations of GG. If THomG(V,W)T\in\operatorname{Hom}_{G}(V,W) is an isomorphism, then T1HomG(W,V)T^{-1}\in\operatorname{Hom}_{G}(W,V).

Proof.

Exercise. ∎

Isomorphic representations are “different pictures of the same object”; if (π,V)(\pi,V) and (ρ,W)(\rho,W) are isomorphic, with T:VWT:V\rightarrow W begin a GG-isomorphism, then we have

ρ(g)=Tπ(g)T1\rho(g)=T\pi(g)T^{-1}

for all gGg\in G, i.e. π\pi and TT tells us everything about ρ\rho.

Lemma 1.2.9.

Let (π,V)(\pi,V) and (ρ,W)(\rho,W) be two representations of a group GG. If ϕHomG(V,W)\phi\in\operatorname{Hom}_{G}(V,W), then (π,ker(ϕ))(\pi,\ker(\phi)) and (ρ,im(ϕ))(\rho,{\mathrm{im}}(\phi)) are subrepresentations of VV and WW, respectively.

Proof.

We know from linear algebra that ker(ϕ)\ker(\phi) and im(ϕ){\mathrm{im}}(\phi) are subspaces, so we just have to show that they are preserved by the GG-actions.

For the kernel: suppose that 𝐯ker(ϕ){\bf v}\in\ker(\phi). Then ϕ(𝐯)=𝟎\phi({\bf v})={\bf 0}, and so for any gGg\in G, we have

ϕ(π(g)𝐯)=ρ(g)ϕ(𝐯)=ρ(g)𝟎=𝟎,\phi(\pi(g){\bf v})=\rho(g)\phi({\bf v})=\rho(g){\bf 0}={\bf 0},

so π(g)𝐯ker(ϕ)\pi(g){\bf v}\in\ker(\phi). Thus ker(ϕ)\ker(\phi) is GG-invariant, as required.

For the image: suppose that 𝐰im(ϕ){\bf w}\in{\mathrm{im}}(\phi). Then 𝐰=ϕ(𝐯){\bf w}=\phi({\bf v}) for some 𝐯V{\bf v}\in V. Then for any gGg\in G,

ρ(g)𝐰=ρ(g)ϕ(𝐯)=ϕ(π(g)𝐯)im(ϕ),\rho(g){\bf w}=\rho(g)\phi({\bf v})=\phi(\pi(g){\bf v})\in{\mathrm{im}}(\phi),

so the image of ϕ\phi is also GG-invariant. ∎

Example 1.2.10.

Let G=D3S3G=D_{3}\cong S_{3} where the isomorphism takes r(123)r\mapsto(123) and s(23)s\mapsto(23). Let (π,3)(\pi,{\mathbb{C}}^{3}) be the permutation representation of GG, so in the basis 𝐞1,𝐞2,𝐞3{\bf e}_{1},{\bf e}_{2},{\bf e}_{3} we have

π(r)=(001100010),π(s)=(100001010).\pi(r)=\begin{pmatrix}0&0&1\\ 1&0&0\\ 0&1&0\end{pmatrix},\qquad\pi(s)=\begin{pmatrix}1&0&0\\ 0&0&1\\ 0&1&0\end{pmatrix}.

Let (ρ,2)(\rho,{\mathbb{C}}^{2}) denote the defining representation of GG; thus

ρ(r)=(1/23/23/21/2),ρ(s)=(1001).\rho(r)=\left(\begin{matrix}-1/2&-\sqrt{3}/2\\ \sqrt{3}/2&-1/2\end{matrix}\right),\qquad\rho(s)=\left(\begin{matrix}1&0\\ 0&-1\end{matrix}\right).

Let T:32T:{\mathbb{C}}^{3}\rightarrow{\mathbb{C}}^{2} be the linear map defined through

T(𝐞1)=(10),T(𝐞2)=(1/23/2),T(𝐞2)=(1/23/2),T({\bf e}_{1})=\left(\begin{matrix}1\\ 0\end{matrix}\right),\qquad T({\bf e}_{2})=\left(\begin{matrix}-1/2\\ \sqrt{3}/2\end{matrix}\right),\qquad T({\bf e}_{2})=\left(\begin{matrix}-1/2\\ -\sqrt{3}/2\end{matrix}\right),

(i.e. TT maps the basis vectors 𝐞i{\bf e}_{i} to the vertices of the equilateral triangle around the origin with a vertex at (1,0)t(1,0)\,^{\mathrm{t}}\!). We want to show that TT is a GG-homomorphism.

To do this we first consider the matrix of TT is

(11/21/203/23/2).\begin{pmatrix}1&-1/2&-1/2\\ 0&\sqrt{3}/2&-\sqrt{3}/2\end{pmatrix}.

We need to show that Tπ(g)=ρT(g)T\pi(g)=\rho T(g) for all gD3g\in D_{3}. As D3D_{3} is generated by just rr and ss we only need to check two cases. We need to check that

(11/21/203/23/2)(001100010)=(1/23/23/21/2)(11/21/203/23/2),\begin{pmatrix}1&-1/2&-1/2\\ 0&-\sqrt{3}/2&\sqrt{3}/2\end{pmatrix}\begin{pmatrix}0&0&1\\ 1&0&0\\ 0&1&0\end{pmatrix}=\left(\begin{matrix}-1/2&\sqrt{3}/2\\ -\sqrt{3}/2&-1/2\end{matrix}\right)\begin{pmatrix}1&-1/2&-1/2\\ 0&\sqrt{3}/2&-\sqrt{3}/2\end{pmatrix},

which is true, and a similar equation coming from ss.

The kernel of the homomorphism TT is the subspace (𝐞1+𝐞2+𝐞3){\mathbb{C}}({\bf e}_{1}+{\bf e}_{2}+{\bf e}_{3}) and in fact TT defines an isomorphism from the subrepresentation

W0={(a,b,c)t3:a+b+c=0}3W_{0}=\{(a,b,c)\,^{\mathrm{t}}\!\in{\mathbb{C}}^{3}:a+b+c=0\}\subseteq{\mathbb% {C}}^{3}

to 2{\mathbb{C}}^{2}.

1.2.3. Exercises

.

Problem 20. Let (π,V)(\pi,V) be a representation of a group GG. Show that if (π,W1)(\pi,W_{1}), (π,W2)(\pi,W_{2}) are two irreducible subrepresentations of (π,V)(\pi,V), then either W1W2=0W_{1}\cap W_{2}=0 or W1=W2W_{1}=W_{2}.

Problem 21. Prove Lemma 1.2.8.

Problem 22. Let (π,V)(\pi,V) be an nn-dimensional representation of a group GG. Show that there exists a representation (π~,n)(\widetilde{\pi},{\mathbb{C}}^{n}) that is isomorphic to (π,V)(\pi,V).

Problem 23. Generalise Example 1.2.4 to the permutation representation (π,n)(\pi,{\mathbb{C}}^{n}) of SnS_{n}.

Problem 24. Let

C0(X)={f:X|f(x)=0forallbutfinitelymanyxX}.C_{0}(X)=\{f:X\rightarrow{\mathbb{C}}\,|\,f(x)=0\mathrm{\;for\;all\;but\;% finitely\;many}\;x\in X\}.

The quasi-regular representation of GG on C0(X)C_{0}(X) is denoted (λ,C0(X))(\lambda,C_{0}(X)) and is defined via the formula

(λ(g)f)(x):=f(g1x)\big{(}\lambda(g)f\big{)}(x):=f(g^{-1}\cdot x)

for all gGg\in G, xXx\in X and fC0(X)f\in C_{0}(X). This is called the functional point of view. Let GG be a group acting on a set XX. Show that (π,(X))(\pi,{\mathbb{C}}(X)), the representation arising from this group action, is isomorphic to (λ,C0(X))(\lambda,C_{0}(X)).