202004032250572537vivek Sahai Lie Algebra 2
202004032250572537vivek Sahai Lie Algebra 2
202004032250572537vivek Sahai Lie Algebra 2
(Remaining Part)
Vivek Sahai
Lucknow University
Lucknow.
April 2, 2020
Lemma
Let V be a vector space over P a field F of characteristic 0, and let
T ∈ L(V ) such that T = ri=1 [Ai , Bi ], Ai , Bi ∈ L(V ) and [T , Ai ] = 0,
i = 1, 2, . . . , r . Then T is nilpotent.
Proof.
Let [T k−1 , Ai ] = 0, then
[T k , Ai ] = T k Ai − Ai T k
= T [T k−1 , Ai ] + TAi T k−1 − [Ai , T ]T k−1 − TAi T k−1
= 0.
r
X r
X
k k−1 k−1
T = T T =T [Ai , Bi ] = (Ai T k−1 Bi − T k−1 Bi Ai )
i=1 i=1
r
X
= [Ai , T k−1 Bi ].
i=1
Corollary
Proof.
Clearly, if L is abelian then L is solvable.
Conversely, let L be solvable and L∗ semi-simple. Then L = L1 ⊕ C , where
C = Z = centre of L and L1 is a semi-simple ideal of L. Therefore L1 is
semi-simple and solvable. But then L1 = {0}. Hence, L = C is abelian.
L+R ∗ L∗
Consider the Lie algebra L+R R . Clearly R = R , which is semisimple.
As R is a homomorphic image of L, so L solvable implies L+R
L+R
R is
L∗
solvable, which gives L+R R is abelian, and hence R is commutative.
Definition
Σ is called indecomposable and V is called Σ-indecomposable if V can not
be written as V = V1 ⊕ V2 , Vi 6= 0 in L(Σ). Clearly, Σ-irreduciblity implies
Σ-indecomposability.
Paper I (Unit IV) M.Sc. Semester IV April 2, 2020 8 / 31
Lecture-2 . . .
Definition
Σ is called completely reducible and V is called Σ-completely reducible if
V = ⊕α Vα , Vα ∈ L(Σ), Vα irreducible.
Assume that dim W < dim V and let the theorem hold for all subspaces
W1 ∈ L(Σ) such that dim W1 > dim W . Since W ( V and V = ⊕α Vα ,
there exists a Vα such that Vα 6⊆ W . Consider Vα ∩ W ∈ L(Σ), As
Vα ∩ W is a subspace of irreducible Σ-subspace Vα , we have either
Vα ∩ W = Vα or Vα ∩ W = {0}. Now Vα ∩ W = Vα is not possible, so
Vα ∩ W = {0}.
Theorem
Let A be an associative algebra of linear transformations in L(V ),
dim V < ∞. If A is completely reducible then A is semi-simple.
Proof.
Let R be the radical of A and let V = ⊕α Vα , Vα ’s irreducible in L(A). Let
R(Vα ) be the subspace spanned by {T (y )|y ∈ Vα , T ∈ R}. Then
R(Vα ) ∈ L(A) and R(Vα ) ⊆ Vα .
Corollary
Proof.
Σ is completely reducible ⇒ L(Σ) is complemented ⇔ L(Σ∗ ) and L(Σ† )
are complemented ⇔ Σ∗ and Σ† are complemented ⇒ Σ∗ , Σ† are
completely reducible ⇒ Σ∗ , Σ† are semisimple.
Definition
An operator T ∈ L(V ) is said to be semi-simple if
mT (x) = p1 (x)p2 (x) · · · pr (x), where each pi (x) is an irreducible
polynomial in F [x], pi (x) 6= pj (x).
Proof.
Let T be not semi-simple. Then mT (x) = p1r1 (x) · · · pkrk (x), where ri > 1
for some i. Let W = p1 (T ) · · · pk (T ). Clearly W ∈ {T }† . If m is the lcm
of {ri }’s, then
W m = p1 (T )m · · · pk (T )m = 0.
Further W 6= 0 because W |mT (x) and deg mT (x) > deg W . So W is
non-zero nilpotent elements of {T }† .
Conversely, if T ∈ L(V ) is semi-simple, then mT (x) = p1 (x) · · · pk (x). Let
W = f (T ) be a nilpotent element of {T }† . Then W r = 0 implies
mT (x)|f r (x), and so mT (x)|f (x). Hence 0 = f (T ) = W .
Paper I (Unit IV) M.Sc. Semester IV April 2, 2020 14 / 31
Lecture-4
Theorem
Let L be a completely reducible Lie algebra of linear operators on a finite
dimensional vector space V over a field of characteristic 0. Then
L = C ⊕ L1 , where C = Z and L1 is a semi-simple ideal. Moreover,
elements of C are semi-simple.
Proof.
We know that if L is completely reducible then L∗ and L† are semi-simple.
So by Theorem of Lecture 1, L = C ⊕ L1 . Further, let T ∈ C be such that
T is not semi-simple. Then there exists a nonzero nilpotent element
W ∈ {T }† . Let k ∈ N be such that W k = 0. Now
{T }† = {a0 + a1 T + a2 T 2 + · · · + an T n |n ∈ N, ai ∈ F }. As T ∈ C , we
have W ∈ {T }† is also in the centre of L. Therefore L† W = WL† is an
ideal in L† and (WL† )k ⊆ W k L† = 0. But 0 6= W ∈ WL† implies WL† is a
non-zero nilpotent ideal in L† , a contradiction as L† is semi-simple.
Paper I (Unit IV) M.Sc. Semester IV April 2, 2020 15 / 31
Lecture-4 . . .
Definition
Let V be a finite dimensional vector space over F and let Σ ⊆ L(V ). A
chain V = V1 ⊃ V2 ⊃ · · · ⊃ Vs ⊃ Vs+1 = {0} of Σ-subspaces of V is a
composition series for V relative to Σ if for every i, there exists no
V 0 ∈ L(Σ) such that Vi ⊃ V 0 ⊃ Vi+1 , that is, VVi+1
i
is irreducible.
Proof.
If T ∈ L, then T has a non-zero characterstic vector x. So T (x) = αx, for
some α ∈ F . Let Vα = {v ∈ V |T (v ) = αv }, the characterstic subspace of
V corresponding to characterstic value α. If U ∈ L then T (U(y )) =
UT (y ) = TU(y ) = U(T (y )) = U(αy ) = αU(y ), for all y ∈ Vα , and so
U(y ) ∈ Vα . This gives Vα is an L-subspace of V . As V is L-irreducible,
we have V = Vα . Hence T = αI .
Thus every T ∈ L is such that T = αI , for some α ∈ F . So every
subspace of V is an L-subspace. Now V is L-irreducible, so V has no
subspaces other than V and {0}. Therefore dimF V = 1.
Paper I (Unit IV) M.Sc. Semester IV April 2, 2020 17 / 31
Lecture-5
Theorem
(Lie’s Theorem) Let V be a finite dimensional vector space over an
algebraic closed field F of characteristic 0 and let L ⊆ L(V ) be such that L
is a solvable Lie algebra. Then there exists a basis B of V such that
[T ]B ∈ Tn (F ) for all T ∈ L.
Proof:
Let V = V1 ⊃ V2 ⊃ · · · ⊃ Vs ⊃ Vs+1 = {0} be a composition series of V
relative to L. If T ∈ L, then Ti = T |Vi ∈ L(Vi ).
Vi Vi
Define T̄i : Vi+1 → Vi+1 by T̄i (x + Vi+1 ) = Ti (x) + Vi+1 . Let
L̄i = {T̄i : T ∈ L, T̄i ∈ L( VVi+1
i
)}.
Hence,
α11 α12 . . . α1s
α22 . . . α2s
. .
[T ]B = .
. .
. .
αss
Theorem
Let L be a finite dimensional solvable Lie algebra over an algebraic closed
field of characterstic zero. Then there exists a chain of ideals
L = Is ⊃ Is−1 ⊃ · · · ⊃ I1 ⊃ {0} such that dim Ij = j.
Proof.
As L is solvable, ad(L) is a solvable Lie algebra of linear transformations
on the finite dimensional vector space L. Let
L = L1 ⊃ L2 ⊃ · · · ⊃ Ls+1 = {0} be a composition series of L relative to
Σ = ad(L). Then ad(L)(Li ) ⊆ Li gives Li is an ideal of L. By Lie’s
theorem dim LLi+1
i
= 1. Hence, dim Ls = 1, dim Ls−1 = 2, ..., dim L1 = s.
Now put Ij = Ls−(j−1) to get dim Ij = dim Ls−(j−1) = j, as required.
Definition
Let L be a Lie algebra. A pair (A, i) where A is an associative algebra and
i a homomorphism of L into AL is called a universal enveloping algebra of
L if for an algebra A and homomorphism θ of L into AL , there exists a
unique homomorphism θ0 of A into A such that θ = iθ0 , that is, the
diagram
A = AL
0
i ↑ & θ (unique)
L −→ A = AL
θ
commutes.
Let (A, i), (B, j) be two universal enveloping algebras for L. Then there
exists a unique isomorphism j 0 of A onto B such that j = ij 0 . We have the
following commutative diagrams:
A = AL
i ↑ & j 0 (unique) ,
L −→ B = BL
j
B = BL
j ↑ & i 0 (unique) .
L −→ A = AL
i
B
j ↑ & iB (unique) ,
L −→ B
j
and
A
i ↑ & iA (unique)
L −→ A
i
Note that T (L) is a vector space over F with usual addition and scalar
multiplication. Define multiplication ⊗ in T (L) by
(x1 ⊗ · · · ⊗ xi ) ⊗ (yi ⊗ · · · ⊗ yj ) = x1 ⊗ x2 ⊗ · · · ⊗ xi ⊗ y1 ⊗ · · · ⊗ yj .
Hence,
{1, uj1 ⊗ uj2 ⊗ · · · ⊗ ujn |n ∈ N,ji ∈ J}
form a basis for T (L).
then θ00 is an algebra homomorphism and θ00 (a) = θ(a) for all a ∈ L.
θ00 ([a, b] − a ⊗ b + b ⊗ a)
= θ00 ([a, b]) − θ00 (a)θ00 (b) + θ00 (b)θ00 (a)
= θ([a, b]) − θ(a)θ(b) + θ(b)θ(a)
= [θ(a), θ(b)] − [θ(a), θ(b)] = 0.
So R ⊆ ker θ00 .
Define θ0 : A → A by θ0 (a + R) = θ00 (a), for all a ∈ T (L). Verify that θ0 is
a well defined homomorphism. Further, for all a ∈ L:
So iθ0 = θ.
Paper I (Unit IV) M.Sc. Semester IV April 2, 2020 28 / 31
Lecture-8 . . .
Let L be a Lie algebra over a field F and let {uj |j ∈ J} be a basis for L,
then {uj1 ⊗ uj2 ⊗ · · · ⊗ ujn |ji ∈ J} form a basis for Ln , n ≥ 1. Here Ln ’s are
as defined in Lecture 7.
Let J be an ordered set. For i, k, i < k, put
(
1 if ji > jk
ηik =
0 if ji ≤ jk .
Theorem
(Poincare-Birkhoff-Witt Theorem) Let L be a Lie algebra over a field F
with basis {uj |j ∈ J}. Let J be an orderderd set, then cosets of 1 and the
standard monomials form a basis for the universal enveloping algebra
A = TR(L) of L.