3.14. Lecture 14

3.14.1. Real forms

Here we will use our understanding of the representation theory of SL2()\operatorname{SL}_{2}(\mathbb{R}) and SU(2)\mathrm{SU}(2) to tell us about the representation theory of 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{C}}, and this will then be used to prove complete reducibility.

Definition 3.14.1.

A real form of a complex Lie algebra 𝔤\mathfrak{g} is a real Lie algebra 𝔥𝔤\mathfrak{h}\subseteq\mathfrak{g} such that every element ZZ of 𝔤\mathfrak{g} can be written uniquely as X+iYX+iY for X,Y𝔥X,Y\in\mathfrak{h}.

For dimension reasons, necessary and sufficient conditions are that 𝔥i𝔥=0\mathfrak{h}\cap i\mathfrak{h}=0 and dim𝔥=dim𝔤\dim_{\mathbb{R}}\mathfrak{h}=\dim_{\mathbb{C}}\mathfrak{g}.

Example 3.14.2.

The Lie algebra 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{R}} is a real form of 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{C}}.

Example 3.14.3.

The Lie algebra 𝔰𝔲n\mathfrak{su}_{n} is a real form of 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{C}}. Indeed, dim𝔰𝔲n=n21=dim𝔰𝔩n,\dim_{\mathbb{R}}\mathfrak{su}_{n}=n^{2}-1=\dim_{\mathbb{C}}\mathfrak{sl}_{n,% \mathbb{C}} and if X,iX𝔰𝔲nX,iX\in\mathfrak{su}_{n} then

iX=(iX)=iX=iXiX=-(iX)^{*}=iX^{*}=-iX

so X=0X=0.

Explicitly, we may write A𝔰𝔩n,A\in\mathfrak{sl}_{n,\mathbb{C}} as X+iYX+iY with

X=12(AA)X=\frac{1}{2}(A-A^{*})

and

Y=i2(A+A)Y=\frac{-i}{2}(A+A^{*})

in 𝔰𝔲n\mathfrak{su}_{n}.

Example 3.14.4.

Recall that

𝔰𝔬n={A𝔰𝔩n,|A+AT=0}.\mathfrak{so}_{n}=\{A\in\mathfrak{sl}_{n,\mathbb{R}}\,|\,A+A^{T}=0\}.

Let

𝔰𝔬n,={A𝔰𝔩n,|A+AT=0}.\mathfrak{so}_{n,\mathbb{C}}=\{A\in\mathfrak{sl}_{n,\mathbb{C}}\,|\,A+A^{T}=0\}.

Since the defining equation A+AT=0A+A^{T}=0 can be checked separately on the real and imaginary parts of AA,

𝔰𝔬n,={B+iC|B,C𝔰𝔬n}\mathfrak{so}_{n,\mathbb{C}}=\{B+iC\,|\,B,C\in\mathfrak{so}_{n}\}

and so 𝔰𝔬n\mathfrak{so}_{n} is a real form of 𝔰𝔬n,\mathfrak{so}_{n,\mathbb{C}}.

Proposition 3.14.5.

Let 𝔥\mathfrak{h} be a real form of 𝔤\mathfrak{g}. There is a one-to-one correspondence between representations of 𝔥\mathfrak{h} and complex-linear representations of 𝔤\mathfrak{g} under which irreducible representations correspond to irreducible representations.

Proof.

Given a \mathbb{C}-linear representation of 𝔤\mathfrak{g}, it is a representation of 𝔥\mathfrak{h} by restriction. Conversely, if (ρ,V)(\rho,V) is a representation of 𝔥\mathfrak{h}, then it extends to a unique \mathbb{C}-linear representation of 𝔤\mathfrak{g} given by the formula (forced by \mathbb{C}-linearity)

ρ(X+iY)𝐯=ρ(X)𝐯+iρ(Y)𝐯.\rho(X+iY){\bf v}=\rho(X){\bf v}+i\rho(Y){\bf v}.

It is an exercise to check this preserves the Lie bracket. The second statement is left as an exercise (Problem 3.14.3). ∎

As a corollary we immediately obtain

Theorem 3.14.6.

The representation theories of 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{R}} and 𝔰𝔲n\mathfrak{su}_{n} are ‘the same’ as the complex-linear representation theory of 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{C}}. All finite-dimensional irreducible representations of 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{R}}, SL2()\operatorname{SL}_{2}(\mathbb{R}), 𝔰𝔲2\mathfrak{su}_{2}, or SU(2)\mathrm{SU}(2) are of the form Symn(2)\operatorname{Sym}^{n}(\mathbb{C}^{2}).

Proof.

The claims about Lie algebras follow from Proposition 3.14.5. By Theorem 3.13.10, every (finite-dimensional) irreducible representation of 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{C}}, and thus 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{R}}, is of the form Symn(2)\operatorname{Sym}^{n}(\mathbb{C}^{2}), and these clearly exponentiate to representations of SL2()\operatorname{SL}_{2}(\mathbb{R}), despite this not being a simply connected group! Similarly for SU(2)\mathrm{SU}(2) (which is simply connected). Since SL2()\operatorname{SL}_{2}(\mathbb{R}) and SU(2)\mathrm{SU}(2) are connected, every representation of them is determined by its derivative, so we have a complete list of the irreducible representations. ∎

3.14.2. Decomposing representations

Theorem 3.14.7.

Let VV be any finite-dimensional complex-linear representation of 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{C}}. Then VV is completely reducible, that is, splits into a direct sum of irreducible representations.

Proof.

Let VV be a finite-dimensional complex-linear representation of 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{C}} and let WVW\subseteq V be a subrepresentation. Then WW is an 𝔰𝔲n\mathfrak{su}_{n}-subrepresentation, by Theorem 3.14.6. As SU(n)\mathrm{SU}(n) is simply-connected, VV and WW exponentiate to representations of SU(n)\mathrm{SU}(n). Since SU(n)\mathrm{SU}(n) is compact, by Maschke’s theorem (Theorem 2.11.4(ii)) there is a complementary SU(n)\mathrm{SU}(n)-subrepresentation WW^{\prime} with

V=WW.V=W\oplus W^{\prime}.

Then WW^{\prime} is a 𝔰𝔲n\mathfrak{su}_{n}-subrepresentation, and so a \mathbb{C}-linear 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{C}}-subrepresentation, so that

V=WWV=W\oplus W^{\prime}

as representations of 𝔰𝔩n,\mathfrak{sl}_{n,\mathbb{C}}. Complete reducibility follows. ∎

The argument in this theorem is called Weyl’s unitary trick. For a similar application of this idea, see the proof of Proposition 3.12.2.

Corollary 3.14.8.

Two finite-dimensional complex-linear representations of 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{C}} are isomorphic if and only if they have the same multiset of weights.

Proof.

Let (π,V)(\pi,V) and (ρ,W)(\rho,W) be finite-dimensional complex-linear representations of 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{C}}. We can decompose VV into irreducibles by looking at the weights. Firstly, look at the maximal weight kk of VV. Then there must be a weight vector 𝐯{\bf v} of weight kk, which is necessarily a highest weight vector, and so VV must contain a copy of Symk(2)\operatorname{Sym}^{k}(\mathbb{C}^{2}) — namely, the subspace 𝐯,Y𝐯,,Yk𝐯\left\langle{\bf v},Y{\bf v},\ldots,Y^{k}{\bf v}\right\rangle. By Theorem 3.14.7 we have

VSymk(2)V.V\cong\operatorname{Sym}^{k}(\mathbb{C}^{2})\oplus V^{\prime}.

The weights of VV^{\prime} are then obtained by removing the weights of Symk(2)\operatorname{Sym}^{k}(\mathbb{C}^{2}) from the weights of VV, and we repeat the process, noting that it terminates as VV is finite dimensional. If we then do the same for WW we see that they decompose into the same sum of copies of Symk(2)\operatorname{Sym}^{k}(\mathbb{C}^{2}) if and only if they have the same multiset of weights. ∎

3.14.3. Exercises

.

Problem 35. Show that the real Lie algebras 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{R}} and 𝔰𝔲2\mathfrak{su}_{2} are not isomorphic. Hint: consider the adjoint action of an arbitrary element of 𝔰𝔲2\mathfrak{su}_{2}.

Problem 36. Finish the proof of Proposition 3.14.5.

Problem 37. We have that Symn(2)\operatorname{Sym}^{n}(\mathbb{C}^{2}) is the irreducible representation of SU(2)\mathrm{SU}(2) of dimension n+1n+1. Let χn\chi_{n} be its character. Every conjugacy class of SU(2)\mathrm{SU}(2) contains an element of the form

exp(itH)=(eit00eit).\exp(itH)=\begin{pmatrix}e^{it}&0\\ 0&e^{-it}\end{pmatrix}.

Show that

χn(exp(itH))=sin((n+1)t)sin(t).\chi_{n}(\exp(itH))=\frac{\sin((n+1)t)}{\sin(t)}.

Problem 38. Let (ρ,V)(\rho,V) be an irreducible representation of 𝔤𝔩2,\mathfrak{gl}_{2,\mathbb{C}}, and let Z=(1001)Z=\begin{pmatrix}1&0\\ 0&1\end{pmatrix}.

  1. (a)

    Show that ρ(Z)\rho(Z) is as a scalar.

  2. (b)

    Show that the restriction of VV to 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{C}} is irreducible.

  3. (c)

    Show that for every λ\lambda\in\mathbb{C} and integer n0n\geq 0, there is a unique irreducible representation of 𝔤𝔩2,\mathfrak{gl}_{2,\mathbb{C}} of dimension n+1n+1 with ρ(Z)=λId\rho(Z)=\lambda\operatorname{Id}.

  4. (d)

    Which of these are derivatives of representations of GL2()\operatorname{GL}_{2}(\mathbb{C})? Hence classify the finite dimensional holomorphic irreducible representations of GL2()\operatorname{GL}_{2}(\mathbb{C}).

Problem 39. Let VV be a finite-dimensional representation of 𝔰𝔩2,\mathfrak{sl}_{2,\mathbb{C}}.

  1. (a)

    What are the weights of the dual representation VV^{*}?

  2. (b)

    Deduce that VVV\cong V^{*}.