REPRESENTATIONS OF LIE GROUPS AND LIE ALGEBRAS

JIAQI JIANG

Abstract. This paper studies the relationship between representations of a

Lie group and representations of its Lie algebra. We will make the correspondence in two steps: First we shall prove that a given representation of a Lie
group will provide us with a corresponding representation of its Lie algebra.
Second, we shall go backwards and see whether a given representation of a Lie
algebra will have a corresponding representation of its Lie group.

Contents
1. Introduction to Lie groups and Lie algebras
2. From Representations of Lie groups to Lie algebras
3. From Representations of Lie algebras to Lie groups
Acknowledgements
References

1
8
11
14
14

1. Introduction to Lie groups and Lie algebras

In this section, we shall introduce the notion of Lie group and its Lie algebra.
Since a Lie group is a smooth manifold, we shall also introduce some basic theory
of smooth manifolds here.
Definition 1.1. A Lie group is a smooth manifold G that also has a group structure, with the property that the multiplication map m : G G G and the
inversion map i : G G, given respectively by
m(g, h) = gh, i(g) = g 1 ,
are both smooth maps.
One of the most important examples of Lie groups is the group GL(V ). Suppose
V is some real or complex vector space. Then the group GL(V ) denotes the set
of all invertible linear transformations from V to itself. The group multiplication
is just composition. If V is finite-dimensional, then any basis for V will induce
an isomorphism of GL(V ) with GL(n, R) or GL(n, C), with n = dim V . (Here,
GL(n, R) is the general linear group consisting of all n n matrices with real
entries. Analogously, GL(n, C) is the complex general linear group).1 Therefore,
once we pick a basis for V , we get a chart on GL(V ) which sends an element
of GL(V ) to its matrix in the chosen basis (which can be thought as an element
Date: October 1st 2013.
1Both GL(n,R) and GL(n,C) are equipped with a standard smooth structure. For details,
please refer to chapter 1 of Lees Introduction to Smooth Manifolds.
1

JIAQI JIANG

in R2n . Obviously, this single chart covers the entirety of GL(V ), thus forming an
atlas. This makes GL(V ) a Lie group. If we choose another basis for V , then the
transition map between the two charts is given by a map of the form A BAB 1
with B the transition matrix between the two bases, it follows that the transition
maps are smooth. Thus, this smooth manifold structure on GL(V ) is independent
of the choice of basis. This Lie group GL(V ) will play an important role in the
later sections of this paper.
First we review the concept of tangent vectors on a manifold. There are in fact
several different ways to define tangent vectors to a smooth manifold. Here, we will
define a tangent vector at a point p in a smooth manifold M as a derivation at p.
Definition 1.2. Let M be a smooth manifold and p be a point of M . A derivation
at p is a linear map X : C (M ) R that satisfies
X(f g) = f (p)Xg + g(p)Xf
for any f, g C (M ).
The intuition behind such definition is that we can regard a derivation as a
directional derivative encountered in multivariable calculus theory. The set of all
derivations of C (M ) at p is called the tangent space to M at p, and is denoted
by Tp M . An element of Tp M is called a tangent vector at p. For any X, Y Tp M ,
we can define X + Y Tp M and cX Tp M for all c R by
(X + Y )f = Xf + Y f, (cX)f = c(Xf )

for all f C (M ). Under these operations, Tp M is clearly a vector space. Moreover, it is important to notice that the construction of the tangent space at a point
is purely a local construction.
Proposition 1.3. Suppose M is a smooth manifold, p M , and X Tp M . If
f, g C (M ) and both functons agree on some neighborhood of p, then Xf = Xg.
The proof of this proposition requires the use of smooth bump functions. The
existence of such functions is one of the properties of a smooth manifold.2
Proof. Let h = f g. Then by linearity of derivations, it is sufficient to show that
Xh = 0 whenever h vanishes in a neighborhood of p. Let C (M ) be a smooth
bump function that is equal to 1 on the support of h and is supported in M \ {p}.
Since 1 where h is not zero, then the product h is identically equal to h.
Since h(p) = (p) = 0, then Xh = X(h) = 0, which follows from the definition of
a derivation.

Having defined tangent vectors on a manifold, we then need to explore how
tangent vectors behave under smooth maps. Suppose M and N are two smooth
manifolds and F : M N is a smooth map. For each p M , there is a natural
map from Tp M to TF (p) N , which is called the push-forward of F :
Definition 1.4. If M and N are smooth manifolds and F : M N is a smooth
map, for each p M we define a map F : Tp M TF (p) N , called the push-forward
of F , by
(F X) (f ) = X (f F ) .

for all f C (N ).
2The proof can be found in chapter 2 of Lees Introduction to Smooth Manifolds.

REPRESENTATIONS OF LIE GROUPS AND LIE ALGEBRAS

It is straightforward to show that the map F : Tp M TF (p) N is a linear map

between the two vector spaces Tp M and TF (p) M . Moreover, it is not hard to verify
the following lemma by simply following the definition of push-forward.
Lemma 1.5. Let F : M N and G : N K be smooth maps, and let p M .
(a)F : Tp M TF (p) N is linear.
(b)(G F ) = G F
(c)(IdM ) = IdTp M
(d)If F is a diffeomorphism, then F : Tp M TF (p) N is an isomorphism
With the concept of push-forward, we can now construct a basis for Tp M for
any p M by the following theorem:3
Theorem 1.6. Let M be a smooth n-manifold. For any p M , Tp M is an ndimensional vector space. If (U, = (xi )) is any smooth chart containing p, the
coordinate vectors (/x1 |p , . . . , /xn |p ) form a basis for Tp M , where





1


.
xi 
xi p
(p)
With the concept of push-forward, we can now extend the familiar notion of
tangent vectors to a smooth curve in Rn to smooth curves in manifolds.
Definition 1.7. If M is a smooth manifold, we define a smooth curve in M to be
a smooth map : J M , where J R is an interval. (In most situations, we
shall let J be an open interval.)
Since we can regard J as an open submanifold of R with the standard smooth
structure, then it is natural to define the tangent vector to a smooth curve in M
as the push-forward of d/dt|to Tt0 R.
Definition 1.8. If is a smooth curve in a smooth manifold M , then the tangent
vector to at t0 J is the vector
 !
d 
0
(t0 ) =
T(t0 ) M,
dt t0
where d/dt|t0 is the standard coordinate basis for Tt0 R.
Vector fields on Rn are familiar objects to us. Now, we wish to define a vector
field on an abstract smooth manifold M . In order to make a precise definition, we
shall first introduce the concept of the tangent bundle of M .
Definition 1.9. Given a smooth manifold M , the tangent bundle of M , denoted
by T M , is the disjoint union of the tangent spaces at all points of M :
a
TM =
Tp M,
pM

together with the smooth structure defined in Theorem 1.10. An element of T M

will be written as an ordered pair (p, X).
3We have omitted the proof here. Readers can find the proof in chapter 3 of Lees Introduction
to Smooth Manifolds.

JIAQI JIANG

The tangent bundle is equipped naturally with a projection map : T M M

by (p, X) = p. From our definition, the tangent bundle seems to be merely a
collection of vector spaces on M . However, with the following theorem, we see that
T M has a smooth manifold structure on it such that the projection map defined
before is smooth with respect to this structure.
Theorem 1.10. For any smooth n-manifold M, the tangent bundle TM has a
topology and smooth structure that make it into a 2n-dimensional smooth manifold.
With this structure, : T M M is a smooth map.
However, we shall not provide the proof here.4 The smooth charts in the smooth

structure defined by the theorem are ( 1 (U ), ), where (U, ) are smooth charts

for M and the map : 1 (U ) R2n is defined by 5

 !


vi
= x1 (p) , . . . , xn (p) , v 1 , . . . , v n .
xi p
Now, we are ready to define a vector field on a smooth manifold M .
Definition 1.11. If M is a smooth manifold, a vector field on M is a continuous
map Y : M T M , usually denoted as p 7 Yp , satisfying
(1.12)

Y = IdM

A smooth vector field on M is then a smooth map Y : M T M which satisfies

(1.12). A rough vector field on M is a map (the map does not need to be continuous)
Y : M T M satisfying (1.10). We will use the notation X(M ) to denote the set of
all smooth vector fields on M . It becomes a vector space under pointwise addition
and scalar multiplication by:
(aY + bZ)p = aYp + bZp .
In additon, if f C (M ) and Y X(M ), then f Y : M T M defined by
(f Y )p = f (p)Yp
is another smooth vector field. We can also regard Y X(M ) as a map C (M )
C (M ) by f 7 Y f , where Y f is defined by
Y f (p) = Yp f.
From the product rule for tangent vectors, it follows that for f, g C (M ) and
Y X(M ),
(1.13)

Y (f g) = f Y g + gY f.

In general, we call a map from C (M ) to C (M ) which is linear over R and

satisfies equation (1.13) as a derivation. Actually, the derivations of C (M ) can
be identified with smooth vector fields on M . We shall merely quote the statement
of the following proposition here.6
Proposition 1.14. Let M be a smooth manifold. A map Y : C (M ) C (M )
is a derivation if and only if it is of the form Yf = Y f for some Y X(M ).
4For proof, see chapter 4 of Lees Introduction to Smooth Manifolds.
5We have used Einstein summation convention in the following equation.
6For proof, see chapter 4 of Lees Introduction to Smooth Manifolds.

REPRESENTATIONS OF LIE GROUPS AND LIE ALGEBRAS

Since we have defined the push-forward of a tangent vector, it is natural to

extend the idea to the vector fields.
Definition 1.15. Suppose F : M N is a smooth map between the smooth
manifolds M and N . Let Y X(M ) and Z X(N ). We say that the vector fields
Y and Z are F-related if F Yp = ZF (p) for each p M .
Given two smooth vector fields V and W , we can get a third smooth vector field,
called Lie bracket of V and W , defined by
[V, W ] f = V W f W V f.
We shall show that [V, W ] defined as above is indeed a smooth vector field.
Lemma 1.16. The Lie bracket [V, W ] of V, W X(M ) is also a smooth vector
field.
Proof. By Proposition 1.14, it is sufficient to show that [V, W ] is a derivation of
C (M ). For any f, g C (M ), we have
[V, W ] (f g) = V (W (f g)) W (V (f g))
= V (f W g + gW f ) W (f V g + gV f )
= f (V W g) + V f W g + g(V W f ) + V gW f
f (W V g) W f V g g(W V f ) W gV f
= f ([V, W ]g) + g([V, W ]f )

From the definition, it is not hard to see that Lie bracket has following properties.
Lemma 1.17. The Lie bracket satisfies following properties for all V, W, Y
X(M ):
(a)Bi-linearity: For a, b R,
[aV + bW, Y ] = a[V, Y ] + b[W, Y ],
[Y, aV + bW ] = a[Y, V ] + b[Y, W ].
(b)Anti-symmetry:
[V, W ] = [W, V ].
(c)Jacobi Identity:
[V, [W, Y ]] + [W, [Y, V ]] + [Y, [V, W ]] = 0.
Proof. Bi-linearity and anti-symmetry follow simply from our definition of the Lie
bracket. For the proof of Jacobi identity:
([V, [W, Y ]] + [W, [Y, V ]] + [Y, [V, W ]])f
= V [W, Y ]f [W, Y ]V f + W [Y, V ]f
[Y, V ]W f + Y [V, W ]f [V, W ]Y f
= V WY f V Y Wf WY V f + Y WV f + WY V f WV Y f
Y V Wf + V Y Wf + Y V Wf Y WV f V WY f + WV Y f
=0


JIAQI JIANG

Another important property of Lie bracket is expressed in the next theorem.

Theorem 1.18. Let F : M N be a smooth map, and let V1 , V2 X(M ) and
W1 , W2 X(N ). If Vi is F-related to Wi for i = 1, 2, then [V1 , V2 ] is F-related to
[W1 , W2 ].
Proof. Fix p M . Then for any f C (N ), we have
F [V1 , V2 ]p (f ) = [V1 , V2 ]p (f F )
= (V1 )p ((V2 )(f F )) (V2 )p ((V1 )(f F ))
= (V1 )p ((W2 f ) F ) (V2 )p ((W1 f ) F )
= (W1 )F (p) (W2 f ) (W2 )F (p) (W1 f )
= [W1 , W2 ]F (p) (f )

The most important application of Lie brackets actually occurs in the context of
Lie groups. Suppose G is a Lie group. According to the definition, for any g G,
we have the map Lg : G G, which is called the left translation, defined by
Lg (h) = gh
for all h G, is a smooth map. In fact, it is a diffeomorphism of G because the
smooth map Lg1 is clearly the inverse for Lg .
A vector field X on G is said to be left-invariant if it is Lg -related to itself for
every g G. This means
(Lg ) Xh = Xgh , for all g, h G.
Since Lg is also a diffeomorphism, then it follows that (Lg ) X = X for every g G.
Suppose X, Y are both smooth left-invariant vector fields on G, then it follows
that
(Lg ) (aX + bY ) = a(Lg ) X + b(Lg ) Y = aX + bY,
for all a, b R, so the set of all smooth left-invariant vector fields on G is a linear
sub-space of X(G). More importantly, this sub-space is closed under Lie brackets.
Theorem 1.19. Let G be a Lie group. If X and Y are both smooth left-invariant
vector fields on G, then [X, Y ] is also a smooth left-invariant vector field.
Proof. Fix some g G. By assumption, we have (Lg ) X = X and (Lg ) Y = Y .
Then by Theorem 1.18, we have [X, Y ] is Lg -related to [(Lg ) X, (Lg ) Y ] = [X, Y ],
so we have (Lg ) [X, Y ] = [X, Y ]. Therefore, [X, Y ] is a smooth left-invariant vector
field on G.

Now, we are ready to define the Lie algebra of a given Lie group. First, we will
give the definition for a general Lie algebra.
Definition 1.20. A real Lie algebra is a real vector space g endowed with a bilinear
map called the bracket [, ] : g g g, which is also anti-symmetric and satisfies
the Jacobi identity for all X, Y, Z g:
[X, [Y, Z]] + [Y, [Z, X]] + [Z, [X, Y ]] = 0

REPRESENTATIONS OF LIE GROUPS AND LIE ALGEBRAS

From Theorem 1.19, it follows that if G is a Lie group, the set of all smooth
left-invariant vector fields on G is a Lie algebra under the Lie bracket. We call this
Lie algebra the Lie algebra of G, and denote it by g.
Having defined the Lie algebra of a Lie group, we shall first explore some basic
properties of g.
Theorem 1.21. Let G be a Lie group. The evaluation map : g Te G, defined
by (X) = Xe , is a vector space isomorphism. Therefore, dimg = dimTe G = dimG.
Proof. Clearly by our definition of , it is a linear map. Therefore, to prove the
theorem, we only need to find a linear inversion for . For each V Te G, we shall

define a vector field V on G by

(1.22)

V g = (Lg ) V.

Clearly, if X is a left-invariant vector field on G such that Xe = V , then X has to

be given by the equation 1.22.

Now, we shall check that V is smooth. To verify this, it is sufficient to show that

V f is smooth for any f C (U ) for any open subset U of G.7

Let : (, ) G be such that (0) = e and 0 (0) = V .8 Then for any g U ,
we have

(V f )(g) = V g f = ((Lg ) V )f = V (f Lg )
= 0 (0)(f Lg )

d 
(f Lg )(t).
=
dt 

(1.23)

t=0

Let : (, ) G R be the map defined by (t, g) = f Lg (t). Since the

multiplication on G, f and are all smooth maps, then it follows that is also

smooth. By the equation (1.23), we can see that (V f )(g) = /t|(0,g) . Since is

smooth, /t|(0,g) depends smoothly on g. So V f is smooth.

It is easy to verify that V is left-invariant. For all h, g G, by definition of V ,

we have

(Lh ) V g = (Lh ) (Lg ) V = (Lhg ) V = V hg ,

where the second equality follows from Lemma 1.5.

Now, let : V 7 V be defined as above. The linearity of follows from its

definition. For any V Te G, we have

( (V )) = (V ) = (V )e = (Le ) V = V.
So = IdTe G . On the other hand, for any vector field X g, we have

( ((X)))g = ( (Xe ))g = X|g = (Lg ) Xe = Xg .

7For the details, check Lemma 4.6 which is proved in chapter 4 of Lees Introduction to Smooth

Manifolds.
8The existence of such a smooth curve is not hard to prove. Let (U, ) be a smooth coordinate

chart centered at e, and let V = V i /xi e in terms of the coordinate basis. Define the map
: (, ) U by (t) = (tV 1 , . . . , tV n ). Then this is a smooth curve with (0) = e and
0 (0) = V .

JIAQI JIANG

Thus, = Idg .

From the proof of Theorem 1.21, we see that if V is a left-invariant rough vector

field on the Lie group G, we then have V = Ve . This shows that every left-invariant
vector field on a Lie group is a smooth vector field.
Before finishing this section, we shall prove an important property of connected
Lie groups, as stated in the following theorem.
Theorem 1.24. Suppose G is a connected Lie group. Let U G be any open
neighborhood of the identity. Then every element of G can be written as a finite
product of elements in U , that is, U generates G.
Proof. Let U be an open neighborhood of e. For every n N, we let
U n = {u1 u2 un : u1 , u2 , . . . , un U } .
Fix some g U n . Then there exist some u U and v U n1 such that g = vu =
Lv u. Since Lv is a diffeomorphism, thus a homeomorphism, it then follows that
g Lv (U ) U n where Lv (U ) is open. Therefore, U n is an open set. Let
[
W =
U n.
nN

Then W is also an open set. Now, we want to show that W is also closed.
Let g W , the closure of W . Since the inversion map i : G G is smooth
and clearly i i = Id, then the inversion map i is a diffeomorphism of G, thus a
homeomorphism. So U 1 = i(U ) is also a open neighborhood of e. It follows that
gU 1 = Lg (U 1 ) is an open neighborhood of g. Since g W , then gU 1 W 6= .
Let h gU 1 W . Then, there exists some u U such that h = gu1 . So
g = hu. Since h W , then h U k for some k N. Then it follows that
g U k+1 W . So W = W , which implies that W is closed.
Therefore, we have shown that W is both closed and open. Since G is connected
and clearly W 6= , then we have G = W = U n . This implies that U generates
nN

G.


2. From Representations of Lie groups to Lie algebras

Having introduced the concept of a Lie group and its Lie algebra, we shall now
explore the connection between the two. In both mathematics and physics, the
study of the representations of a Lie group plays a major role in the understanding
of continuous symmetry. One of the basic tools in studying the representations
of Lie groups is the use of the corresponding representations of Lie algebras. In
this section, we shall focus on how to get a corresponding representation of the Lie
algebra, given a representation of its Lie group. First, we will give the definition of
a representation of a Lie group and that of a representation of a Lie algebra.
Definition 2.1. A representation of a Lie group G on a finite-dimensional vector
space V is a smooth group homomorphism
: G GL(V ).
Namely, for any h, g G and v V , we have
(hg)(v) = ((h) (g))(v).

REPRESENTATIONS OF LIE GROUPS AND LIE ALGEBRAS

Definition 2.2. A representation of a Lie algebra g on a finite-dimensional vector

space V is a Lie algebra homomorphism
: g gl(V ) = End(V ).
Namely, for any X, Y g and v V , we have
([X, Y ])(v) = ((X) (Y ))(v) ((Y ) (X))(v).
The important aspect of the Lie algebra of a Lie group arises from the fact that
each Lie group homomorphism induces a Lie algebra homomorphism, as stated in
the next theorem.
Theorem 2.3. Suppose G and H are Lie groups and let g, h be their Lie algebras.
If F : G H is a Lie group homomorphism, then for every X g, there is a
unique vector field Y h that is F -related to X. Let this vector field be denoted by
F X, then the map F : g h defined in this way is a Lie algebra homomorphism.
Proof. If there exists a vector field Y h that is F -related to X, then we must
have Ye = F Xe . Then according to the remark we have made after Theorem 1.21,
Y is uniquely determined by

^
Y = Ye = F
Xe .
Now, we need to show that the vector field Y determined this way is indeed F related to X. Since F is a Lie group homomorphism, we have
(F Lg )(g 0 ) = F (gg 0 ) = F (g)F (g 0 )
= (LF (g) F )(g 0 ).
So by Lemma 1.5, we get
F (Lg ) = (F Lg ) = (LF (g) F ) = (LF (g) ) F .
It then follows that
F Xg = F (Lg ) Xe = (LF (g) ) F Xe = (LF (g) ) Ye = YF (g) .
This proves that X and Y are F -related.
Thus, we define a map F : g h by X 7 F X = Y , where X g and Y is the
unique vector field in h that is F -related to X as we have shown above. Now what
remains to show is that F is a Lie algebra homomorphism. This is a immediate
consequence of Theorem 1.18, which states precisely that
F [X, Y ] = [F X, F Y ].

The map F : g h defined in the previous theorem will be called the induced
Lie algebra homomorphism of F . In some literature, F is called the differential of
map F .
Now consider a representation : G GL(V ) of a Lie group G on a finitedimensional (real or complex) vector space V . By definition, the map is actually
a Lie group homomorphism between the Lie group G and the Lie group GL(V ).
Thus, by applying Theorem 2.3, we get a induced Lie algebra homomorphism
: g Lie(GL(V )),
where Lie(GL(V )) denotes the Lie algebra of the Lie group GL(V ). In fact, the map
is indeed a representation of the Lie algebra g on the finite-dimensional vector

10

JIAQI JIANG

space V . This is a consequence of the important fact that Lie(GL(V )) is isomorphic

to gl(V ) in the Lie algebra sense, which is stated in the following theorem.
Theorem 2.4. The composition of the natural maps
Lie(GL(V )) TId GL(V ) gl(V )
gives a Lie algebra isomorphism between Lie(GL(V)) and gl(V).
Proof. Without loss of generality, we will assume that V is a finite-dimensional
complex vector space. Fix a basis for V . Then according to the remark made in
section one, GL(V ) can be then identified with GL(n, C). Analogously, gl(V ) can
be identified with gl(n, C). This provides us with globally defined coordinates on
GL(V ) by considering the corresponding matrix entries Xji of X GL(V ) resprect
to the basis we have previously chosen. Then we can get a natural isomorphism
TId GL(V ) gl(V ) which takes the form




i
Aj
 Aij ,
i
Xj 
Id

(Aij )

where
denotes the element A gl(V ) with respect to the basis we have chosen.
It then follows that any A = (Aij ) gl(V ) determines a left-invariant vector field

A Lie(GL(V )) as defined by equation 1.22, which in this case becomes

 !


i 
,
AX = (LX ) Aj

Xji 
Id

for any X GL(V ). Since LX is the linear map Y 7 XY for Y GL(V ), then it
follows that the coordinate representation of the push-forward (LX ) is given by
 !





i j
i
,
= Xj Ak
(2.5)
(LX ) Aj

Xji 
Xki X
Id

(Xki )

is the matrix representation of X under the chosen basis. Note that

where
Einstein summation convention is used in equation 2.5. It follows that for any

A gl(V ), the left-invariant vector field A Lie(GL(V )) determined by A is given

by



A = Xji Ajk
.
Xki X
Clearly, the map

A gl(V ) 7 A Lie(GL(V ))
and the composition map
X Lie(GL(V )) 7 XId TId GL(V ) 7 (Xji ) gl(V )
are linear inversions of each other. Thus, we have a natural isomorphism between
Lie(GL(V )) and gl(V ) in the sense of vector spaces. What remains to be shown
is that this natural isomorphism is indeed a Lie algebra isomorphism. For any

REPRESENTATIONS OF LIE GROUPS AND LIE ALGEBRAS

11

A, B gl(V ), we have



p q
[A, B]X = Xji Ajk
,
X
B
q r
Xrp X
Xki





  i j 
p q
p q
X
B
= Xji Ajk

X
Xj Ak
B
q r
q r
Xrp
Xrp X
Xki X
Xki




p q r

X
B
A
= Xji Ajk Brk
q
r
k

Xri X
Xkp X


 
 ,
(2.6)
= Xji Ajk Brk Xji Bkj Akr
X i 
r X

Xqp /Xki

where we have used the fact that

is equal to 1 if p = i and q = k, and 0
otherwise, as well as the fact that the mixed partial derivatives of a smooth function
can be taken in any order. Evaluating equation 2.6 at the identity map Id GL(V )
results




 .
[A, B]Id = Aik Brk Bki Akr
Xri Id
Since the matrix representation of [A, B] gl(V ) is just (Aik Brk Bki Akr ), thus
according to the natural isomorphism we get

^
[A, B] = [A,
B].
This proves that the natural isomorphism between Lie(GL(V ) and gl(V ) is a Lie
algebra isomorphism. Thus we can simply identify Lie(GL(V )) with gl(V ).

Combining Theorem 2.3 and Theorem 2.4, we get the following corollary.
Corollary 2.7. If : G GL(V ) is a representation of the Lie group G on a
finite-dimensional vector space V , then the induced homormophism : g gl(V )
is a representation of the Lie algebra g of the Lie group G on the vector space V .
3. From Representations of Lie algebras to Lie groups
In the previous section, we saw that given a representation : G GL(V ) of the
Lie group G on a finite-dimensional vector space V , we obtained an corresponding
representation : g gl(V ) of its Lie algebra on the same vector space V .
However, given a representation of the Lie algebra g of the Lie group G on a finitedimensional vector space V , can we then find an corresponding representation of the
Lie group G on the vector space V ? The answer is yes if we impose the additional
condition that the Lie group G is simply-connected.
In order to reach our ultimate goal, we shall use an important fact that every
Lie subalgebra corresponds to some Lie subgroup.9 The proof of this fact involves
a lot of new machinery which we have not introduced in this paper. Therefore, we
will only state the theorem here without offering the proof.10
9A Lie subalgebra of the Lie algebra g is a linear sub-space h g which is also closed under

brackets. A Lie subgroup of a Lie group G is a subgroup H of G which also has a Lie group
structure and is an immersed submanifold of G.
10
For details of the proof, please refer to either chapter 20 of Lees Introduction to Smooth
Manifolds or Lecture 8 of Representation Theory:A First Course by Fulton and Harris.

12

JIAQI JIANG

Theorem 3.1. Suppose G is a Lie group with Lie algebra g. If h is any Lie
subalgebra of g, then there exists a unique connected Lie subgroup of G whose Lie
algebra is h.
The most important application of Theorem 3.1 is to prove the following theorem.
Theorem 3.2. Suppose G and H are Lie groups with G simply connected, and let
g and h be their Lie algebras. For any Lie algebra homomorphism : g h, there
exists a Lie group homomorphism : G H such that = .
In order to prove Theorem 3.2, we shall first prove the following lemmas.
Lemma 3.3. Suppose G and H are Lie groups with Lie algebras g and h. Then
g h is a Lie algebra with the bracket defined by
[(X, Y ), (X 0 , Y 0 )] = ([X, X 0 ], [Y, Y 0 ]).
The Lie algebra g h defined above is indeed isomorphic to Lie(GH), the Lie
algebra of the Lie group GH.
Proof. The verification for the vector space gh being a Lie algebra with the bracket
defined in the lemma is an immediate consequence of the properties satisfied by the
brackets for g and h.
What remains to show is that Lie(GH) is isomorphic to g h. Let (U, xi )
and (V, y j ) be charts which respectively contain eG and eH . Then, it follows that
(U V, xi y j ) is a chart which contains eGH . So for any tangent vector Z
Te (G H), we have




yj 
xi 
+z
.
Z=z

i
j
x
y 
e

By Theorem 1.21, we know that every left-invariant vector field W on a Lie group
is uniquely determined by its value at e, namely We . Thus, we get a vector space
isomorphism : Lie(G H) g h by considering the composition of maps



 






i

yj 
xi 
yj 
Z Ze = z x
+
z

z
,
z
ZG , ZH ,




i
j
i
j
x
y
x
y
e

where ZH and ZG are the unique left-invariant vector fields determined respectively
i
j
by z x /xi Te G and z y /y j Te H. To complete the proof, we need to show
, V Lie(G
that this isomorphism is indeed a Lie algebra isomorphism. For any W
H),
!
!


xk
yl
xk
yl


xi w
y j w
xi v
y j v

v
v
[W , V ]e = w
+ w
,
xi
xi
xk e
y j
y j
y l e
i

where we have used the fact that wx /y j = 0 and wy /xj = 0 (analogously

j
i
for v x and v y ).11 So we have
, V ]e = ([WG , VG ]e , [WH , VH ]e ).
[W
Then it follows that : Lie(G H) g h is a Lie algebra isomorphism.


 


j
bx
L


11This is because for Z =  , Z
gh = (Lgh ) Z = (Lgh )
=
+



i
i
i
j
x e
x e
x
x gh

yl
b

L

b is the coordinate representation of Lgh . Since Lgh = Lg Lh , then

, where L

xi y l
gh
j
by

L
clearly x
= 0 . From this we can see that wx and wy
i
respectively.

can only depend on xi and y j

REPRESENTATIONS OF LIE GROUPS AND LIE ALGEBRAS

13

Lemma 3.4. Suppose G and H are connected Lie groups with Lie algebras g and
h respectively. Let F : G H be a Lie group homomorphism. Then F is a
smooth covering map if and only if the induced homormorphism F : g h is an
isomorphism.
The proof for Lemma 3.4 requires some extra machinery which we have not
introduced in this paper. So for this reason, we shall not provide the proof here.12
Now ,we are ready to prove Theorem 3.2.
Proof. By Lemma 3.3, the Lie algebra of GH is isomorphic to gh. Let r gh
be the graph of , namely
r = {(X, X) : X g} .
Clearly, r is a vector sub-space of g h. Because is a Lie algebra homomorphism,
we then have
[(X, X), (Y, Y )] = ([X, Y ], [X, Y ]) = ([X, Y ], [X, Y ]) r.
So r is in fact a Lie subalgebra of g h. Then by Theorem 3.1, there exists a unique
connected Lie subgroup R G H whose Lie algebra is r.
Let 1 : G H G and 2 : G H H be the projection maps to G and
H respectively. Obviously, 1 and 2 are Lie group homomorphisms, so it follows
that = 1 |R : R G is also a Lie group homomorphism. We shall show that
is a smooth covering map in the next step. Since G is simply connected, then it
will follow that is bijective, and therefore is a Lie group isomorphism.13
In order to show that is a smooth covering map, it is sufficient to show that
its induced Lie algebra homomorphism is an isomorphism according to Lemma
3.4. Consider the following sequence of maps

1
R , G H
G.

The composition of these maps is exactly . Thus by Lemma 1.5, it follows that
the induced Lie algebra homomorphism is just the composition of the maps
$

1
r , g h
g,

where $1 denotes the projection g h g. This implies that is just the

projection $1 : g h g restricted to r. Since r is the graph of , then r h=0.
So = $1 |r : r g is an isomorphism. Therefore, by Lemma 3.4, is a
smooth covering map of the simply connected Lie group G and thus a Lie group
isomorphism.
Now define a Lie group homomorphism : G H by = 2 |R 1 . Since
= 1 |R , then according to the definition of , we have
2 |R = 1 |R .
Since the Lie algebra homomorphism induced by the projection 2 : G H H
is just the projection $2 : g h h, then by Lemma 1.5, we have
$2 |r = $1 |r : r h.
12For details of the proof, please refer to chapter 9 of Lees Introduction to Smooth Manifolds.
13This follows from the property of covering maps of simply connected spaces: If X is a simply

e X is a homeomorphism. For proofs, please refer

connected space, then any covering map : X
to Lees Introduction To Topological Manifolds.

14

JIAQI JIANG

Therefore, for any X g, we have

X = $2 |r (X, X) = $1 |r (X, X) = X.
This shows that = .

By invoking Theorem 3.2, we shall finally reach the ultimate goal of this section.
Corollary 3.5. Let G be a simply connected Lie group and g be its Lie algebra. If
: g gl(V ) is a representation of the Lie algebra g on a finite-dimensional vector
space V , then there exists a representation : G GL(V ) of the Lie group G on
the same vector space V such that the corresponding representation of G on V
is exactly .
Proof. By Theorem 2.4, we know that the Lie algebra of GL(V ) is exactly gl(V ).
Since the representation : g gl(V ) is a Lie algebra homomorphism, then by
Theorem 3.2, there exists a Lie group homomorphism : G GL(V ) such that
= . Then the map is the desired representation of Lie group G on the vector
space V .

Acknowledgements. I shall thank my mentor Jonathan Gleason, without whose
generous help and support, I would not have possibly picked this interesting topic
and finished this paper. Also I shall thank professor Peter May for organizing this
amazing REU, from which I have truly learned a lot of fun mathematics.
References
[1] John M. Lee. Introduction to Smooth Manifolds. Springer-Verlag New York, Inc. 2003.
[2] William Fulton and Joe Harris. Representation Theory - A First Course. Springer-Verlag New
York, Inc. 1991.
[3] Anthony W. Knapp. Lie groups Beyond An Introduction. Birkh
auser Boston. 2002.