You are on page 1of 30

REPRESENTATIONS OF C L A S S I C A L LIE SUPERALGEBRAS

V. Kac
D e p a r t m e n t of M a t h e m a t i c s
Massachusetts I n s t i t u t e of Technology, Cambridge, MA 02139

These notes contain an e x p o s i t i o n of the results of [i] and [2]


concerning the finite-dimensional representations of basic classical
Lie superalgebras. Among the new results we note a) the e x p l i c i t con-
struction of typical representations for all superalgebras (modules
V(A)), b) some n e w c h a r a c t e r i z a t i o n s of typical representations (see
Theorem i), c) formula for the m u l t i p l i c i t i e s of the r e s t r i c t i o n of a
representation to the even part (formula (2.8)) and d) s t r u c t u r e theory
of m o d u l e s V(A) (Theorem 3).
The list of c o m p l e x finite-dimensional simple Lie superalgebras
c onsist s of two e s s e n t i a l l y different parts -- c l a s s i c a l and C a r t a n
superalgebras [i]. Classical superalgebras are those for w h i c h the
even subalgebra is reductive. The list of c l a s s i c a l Lie superalgebras
is in turn d i v i d e d into two parts -- basic and strange superalgebras.
The basic classical Lie superalgebras are from many points of v i e w the
closest to the o r d i n a r y simple Lie algebras. For instance they can be
constructed in terms of C a r t a n m a t r i x and c a n o n i c a l generators. In
[i] these Lie superalgebras are called contragredient superalgebras.
F i r s t we d e s c r i b e the s t r u c t u r e of basic classical Lie s u p e r a l g e b -
ras Then we c o n s t r u c t irreducible representations with highest
weight of these algebras and c l a s s i f y all the finite-dimensional irre-
ducible representatons The central part of the notes is the com-
putation of the c h a r a c t e r s and s u p e r c h a r a c t e r s for typical finite-dimen-
sional irreducible representations of basic c l a s s i c a l Lie superalgebras.
We call an i r r e d u c i b l e representation typical, if it splits in any
finite-dimensional representation. An e q u i v a l e n t definition requires
that the i n f i n i t e s i m a l character determines the r e p r e s e n t a t i o n uniquely
among the i r r e d u c i b l e finite-dimensional representations.
We then find n e c e s s a r y and s u f f i c i e n t conditions for a representa-
tion to be typical in terms of h i g h e s t w e i g h t (these c o n d i t i o n s are
expressed in a finite s y s t e m of i n e q u a l i t i e s ) . A crucial point in the
p r oof is the i n v a r i a n t s restriction theorem for Lie superalgebras
(Theorem 2), w h i c h is a g e n e r a l i z a t i o n of the c l a s s i c a l Chevalley
theorem for s e m i s i m p l e Lie algebras. An e s s e n t i a l circumstance here is
that C h e v a l l e y t h e o r e m only holds in a w e a k e n e d form. This is r e l a t e d
to the p r e s e n c e of "degenerate" (non-typical) representations, for
which the p r o b l e m of c o m p u t i n g the c h a r a c t e r s remains unsolved (cf.
598

formula (2.1)).
The ground field is the field C of c o m p l e x numbers.

§l. Basic classical Lie superalgebras

i. Some examples [i]. Let V = V~ • V T be a Zz-graded space,


d i m V~ = m, d i m V T = n. The associative algebra End V becomes an
associative superalgebra if we let

EndiV = {a E E n d V l a Vs ~ Vi+s} , i,s ~ ~z •

The b r a c k e t [a,b] = ab - (_i) (deg a) (deg b ) b a makes End V into a


Lie superalgebra, denoted by Z(V) or £(m,n). If we regard the same
decomposition V = V 0 @ V I as a ~-gradation of V, then the same con-
struction gives a consistent ~-gradation: i(V) = G_I • Z(V)~ @ G I.
On Z(V) there is d e f i n e d the supertrace, a linear function
str : £(V) --> k, w h i c h is u n i q u e l y defined by the p r o p e r t i e s
str([a,b]) = 0, a,b e £(V), and str i~ = m - n.
The subspace

sZ(m,n) = {a ~ £(m,n) Istr a = 0}

is an i d e a l in i(m,n) of c o d i m e n s i o n i. In some (homogeneous) basis


of V, i(m,n) consists of m a t r i c e s of the form (7 ~) and
e ~) = tr ~ - tr
str 17 6 The Z-gradation s£(m,n) = G _ l ~ s i ( m , n ) ~ @ GI
looks as follows: sZ(m,n)~ is the set of m a t r i c e s of the form
0
10 61, w h e r e tr e = tr 8, GI is the set of m a t r i c e s of the form

I~ ~) and G_I of the form I0 ~ (where ~ is an (mxm)-, 6 an


(n×n)-, B an (m×n) -, and ~ an (n×m) -matrix) .
sZ(n,n) contains the o n e - d i m e n s i o n a l ideal consisting of
the scalar matrices ll2n. We set

~(m,n) = si(m+l,n+l) for m + n, m,n > 0 ,

~(n,n) = s~(n+l,n+l)/{ll2n+2 } , n > 0 .

Let F n o w be a n o n - d e g e n e r a t e bilinear f o r m on V, such that


V$ and V_ are o r t h o g o n a l and the restriction of F to V~ is a
1
symmetric and to VT a skew-symmetric form (in p a r t i c u l a r , n = 2r is
even). We define in Z(m,n) the subalgebra
599

osp(m,n) = osp(m,n)~ ~ osp(m,n)~ by setting


osp(m,n) s = {a 6 £(m,n)slF(a(z),y ) = -(-i) s d e g x F(x,a(y) }, s ~ ~2"
The algebra osp(m,n) is called ortho-svmplectic Lie superalgebra.
Let us find the explicit matrix form of the elements of osp(m,n).
We put

~m = I'" "lll' B = I ~0 0 0 -~rO ~r001

and let dr mean the transposition of a m a t r i x d with respect to the


side diagonal. In some homogeneous basis of V the matrix of the form
F is B and osp(m,n) consists in this basis of the matrices of the
form:

L ~ ~ nI
_ r
~r
~
6 -8 r
Y

where ~ = _ r, Y = yr , 6 = 6 r, 8 is any (r×r)-matrix, ~ and n


are any (m×r)-matrices.
By analogy with Cartan's notation we set:

~(m,n) = osp(2m+l,2n) , m > 0, n > 0 ;

~(m,n) = osp(2m,2n) , m ~ 2, n > 0 ;

~(n) = osp(2,2n-2) , n ~ 2.

We now examine the Lie superalgebra ~(n). Subalgebra of the form


~(n)~ consists of the matrices

L
0 -~

6 _~r
Y
where 8, Y, and ~ are (n-lxn-l)-matrices, yr = Y, ~r = 6, and
~ ~. ~(n) has the consistent k-gradation:
~(n) = G_I • ~(n)~ • G1, where G_I and Gl consist of matrices of
the form (respectively):
600

0 0

~2 q2 0 0
r r
-rl 2 0 0 -T1 1
r r
~2 0 0

2. Classification. We call a b i l i n e a r form F on Lie s u p e r a l g e b r a


G = G~ + G T invariant if it s a t i s f i e s the following three conditions:

a) F(a,b) = (-i) ~ BF(b,a), a ~ G , b ~ GB;


b) F(a,b) = 0, a C G~, b ~ G~;
c) F([a,b],c) = F(a, [b,c]).

We call a Lie s u p e r a l g e b r a G = G~ @ G T basic classical if a) G is


simple, b) Lie a l g e b r a Go is a r e d u c t i v e subalgebra and c) there
exists a non-degenerate invariant bilinear form on G.
From [i] one d e r i v e s the f o l l o w i n g statement.

Proposition i.I. a) A c o m p l e t e list of b a s i c classical Lie s u p e r a l -


gebras is as follow s :
i) s i m p l e Lie a l g e b r a s ,
2) ~ ( m , n ) , ~ ( m , n ) , ~(n), ~(m,n) , ~ ( 2 , 1 ; ~ ) , ,~(4), ~ ( 3 ) .
b) For the Lie s u p e r a l g e b r a s ~(m,n) , D(m,n), .~(2,1;~), FF(4),
~(3), the G[-module G? is i r r e d u c i b l e and i s o m o r p h i c to the m o d u l e s
in the f o l l o w i n g list:

G G~ G~IG T G G~

B (m,n) B ~F(4) B 3 • ~,l spiny ® sZ z


~-m @ ~ n S O 2 m + l @ SP2n
D (m,n) 2m s ~Cn S O 2 m ® SP2n ~(3) G2 • A 1 G 2 ® s£ 2

D(2,1;~)
D s£ 2 @ s~ 2 ® s£ 2

C) The Lie superalgebras ~AA(m,n) and ~(n) admit a unique con-


sistent ~-gradation of the form G_~ @ G O @ G I, where the
G0-modules GI and G_I are i r r e d u c i b l e and c o n t r a g r e d i e n t ; they are
isomorphic to the m o d u l e s in the f o l l o w i n g list:
601

G Go G01G_ !

~(m,n) A @A • C Slm+ 1 8 S£n+ 1 ®


--m ~n

A(n,n) An t9 A S£n+ 1 ® Sin+ 1


~'~n

C(n) C n _ 1 ~9 C~ cSP2n_ 2

3. Borel subalgebras. Let G = G~ @ GT be a Lie superalgebra and


B~ be a B o r e l subalgebra (maximal solvable subalgebera of G~). We
call Borel subalgebra of Lie superalgebra G any m a x i m a l solvable
subalgebra B of G, w h i c h contains B~. All Borel subalgebras in a
Lie algebra are conjugate. Lie superalgebra in g e n e r a l contains sev-
eral classes of c o n j u g a c y of B o r e l subalgebras b u t of c o u r s e any class
of c o n j u g a c y contains a Borel subalgebra B with given By. From
this remark we easily obtain the following proposition (see a l s o [i]).

Proposition 1,2. a) There is a o n e - t o - o n e correspondence between


classes of c o n j u g a c y of B o r e l subalgebras in s£(m,n) and the o r d e r -
ings o:e. ,...,e. of a f i x e d h o m o g e n e o u s b a s i s of V, such t h a t
ii lm+ n
the s e q u e n c e of i n d i c e s of e v e n e l e m e n t s and r e s p e c t i v e l y of odd o n e s
increases. This correspondence is g i v e n by:

--> B = {a 6 s Z ( m , n ) l a < e i ..... e i > c <e . . . . . . e i > ,


1 S ii S

s = 1,2,...,m + n }

b) There is a o n e - t o - o n e correspondence between classes of


conjugacy of B o r e l subalgebras in osp(m,n) and o r d e r i n g s
~:e . . . . . . e of a f i x e d h o m o g e n e o u s b a s i s of a f i x e d m a x i m a l
11 i [m+n/2 ]
i s o t r o p i c s u b s p a c e of V w i t h the same p r o p e r t y as in a). The cor-
respondence is g i v e n by:

--> B = {a e o s p ( m , n ) l a < e i l ,...,e i s > c < e i l , . . . , e i s > ,

s = 1,2 ..... [(m+n)/2]}

c) The numbers of c o n j u g a c y classes of B o r e l subalgebras in


D(2,1;e), ~FF(4) and G(3) i__ss 4, 4 and 1 respectively (the cor-
602

responding systems of s i m p l e roots are d e s c r i b e d in [i]).

4. Root system [1]. Let G = G~ ~ G I be a b a s i c classical Lie super-


algebra and let H be a C a r t a n subalgebra of G~. For e~H*, ~ @ 0,
we set

GC~ = { a ~ G I [h,a] = ~(h) a, h E H}.

We c a l l ~ a r o o t if G ~ 0. A root e is c a l l e d e v e n (respective-
ly odd) if G ~ G~ ~ 0 (respectively G ~ Gy ~ 0).
We d e n o t e by A, A0 and AI the sets of all roots, even r o o t s
and o d d r o o t s r e s p e c t i v e l y . We i n t r o d u c e also the f o l l o w i n g two sets
of roots:

B e l o w we d e s c r i b e the s y s t e m s of e v e n a n d odd roots for all b a s i c clas-


s i c a l Lie s u p e r a l g e b r a s .
In all the e x a m p l e s the C a r t a n subalgebra H is a s u b s p a c e of the
space D of d i a g o n a l matrices; the r o o t s are e x p r e s s e d in terms of the
standard basis e. of D* (more p r e c i s e l y the r e s t r i c t i o n s of the
1
~.i to H) .
~(m,n). The r o o t s are e x p r e s s e d in terms of l i n e a r functions

e l , . . . , e m + l , ~ I = S m + 2 , . . . , 6 n + 1 = em+n+ 2.

A 0 . {e i. .
sj; @i . dj}; AI {± (ei - ~j)}, i ~ j.

~(m,n). The r o o t s are e x p r e s s e d in terms of l i n e a r functions

sl,...,em,6 1 = S2m+l,...,6n = S2m+n.

A0 = { ± s i k e j ; ± 2 ~ i ; ± e i ; ± ~ i ± 6 j } ; A I = {±~i;±eik~j}, i ~ j.

~(n). The r o o t s are e x p r e s s e d in t e rms of l i n e a r functions

Sl,d I = s 3 , . . . , ~ n _ 1 = en+ I.

A 0 = {±26i;±6i±~j}; A z = {±sz±~i}-

~(m,n). The roots are e x p r e s s e d in t erms of l i n e a r functions

e f . . . , e m , 6 I = e 2 m + l , . . . , ~ n = S2m+n.
603

A0 = {±ei±ej;±26i;±6i±6j}, AI = {±ei±6j} , i ~ j.

~(2,1;e). The roots are expressed in terms of linear functions


el, e2 and e 3.

A0 = {±2ei} ; AI = {±el±e2±e3}.

F(4). The roots are expressed in t e r m s of l i n e a r functions e 1,


e 2, e3, corresponding to ~3, and 6, corresponding to ~i"

A0 = {±gi±ej;±ei;±6}, i # j; A1 =
i
{~(±el±e2te3
±6)} .

G(3). The roots are e x p r e s s e d in t e r m s of l i n e a r functions e I,


e2, e3, corresponding to ~2, el + ez + e s = 0, and 6, c o r r e s p o n d i n g
to ~i"

A0 = {e i - e j ; ± g i ; ± 2 6 } ; AI = { ± e i ± 6 ; ± 6 }.

We fix an i n v a r i a n t non-degenerate bilinear form ( , ) on G.

Proposition 1.3. A basic classical Lie superalgebra G satisfies the


following properties :
a) G = H • (eAG~)

b) dim G = i, e 6 A, e x c e p t for ~A(I'I)"


c) [G ,G B] ~ 0 if and only if ~, ~, e+~ c A.
d) If ~ e A (or A0, o_rr AI, o__rr ~0, o__rr ~i), then so is -~.
e) (G ,G B) = 0 for ~ 4 -B, the form ( , ) determines a non-
degenerate pairing of G and G and the restriction of
( , ) o__n_n H is n o n - d e g e n e r a t e .
f) [e ,e_e] = (e ,e_ )h , where he is a n o n - z e r o vector de-
termined by (ha,h) = a(h), h ~ H.
g) The bilinear f o r m on H* defined by (~,B) = (h~,h B) i_s a
non-degenerate W-invariant form.
h) k~ ~ A for e ~ 0, k ~ ±i, if and only if ~ eat, and[
(~,~) ~ 0; in this case k = +2.
In p a r t i c u l a r , root ~ belonqs to AI if and only if (~,~) = 0.

Let now B~ be a B o r e l subalgebra of G~, containing H. We fix


a Borel subalgebra B = B~ ~ By of G. Since the adjoint representa-
tion of H in G is d i a g o n a l i z a b l e we obtain the following decompo-
604

s i t i o n of G:

G = N- @ H @ N +, B = H @ N +,

where N- and N+ are s u b a l g e b r a s and [H,N +] ~ N +, [H,N-] c N-.


A root ~ is c a l l e d p o s i t i v e (respectively negative) if
- + +
G~ n N + ~ 0 (respectively G~ n N ~ 0 . We d e n o t e by A +, A0, At,
-+ -+
A 0, A I the s u b s e t s of p o s i t i v e r o o t s in the sets A, A 0 etc. res-
pectively.
We let P0 (respectively PI) denote the h a l f - s u m of all of the
even (respectively odd) positive roots. We set P = P0 - Pl- We let
-+
also ~0 denote the h a l f - s u m of all the r oots from A 0.
A positive roo t ~ is c a l l e d simple if it c a n n o t be d e c o m p o s e d
into a sum of two p o s i t i v e roots. Let ~ = {~i .... 'er } be the set
of all s i m p l e roots.

Proposition 1.4. For a basic classical Lie s u p e r a l g e b r a G the fol-


lowing properties hold.
a) A l l the s u b s p a c e s G ~ N± are o n e - d i m e n s i o n a l .
b) One can c h o o s e n o n - z e r o elements e.1 ~ G ~ , ~ N +, f.1 6 G --~, ~ N
1 1
and h i 6 H, i = l,...,r, such that ei' fi' hi' i = l,...,r,
is the s y s t e m of g e n e r a t o r s of G, s a t i s f y i n g the f o l l o w i n g re-
lations:

(i.i) [ei,f j] = 6ijh i, [hi,h j] = 0, [hi,e j] = a i j e j, [hi,f j] = - a i j f j

1
c) (p, ei ) = [ ( ~ i , ~ i ) ; in p a r t i c u l a r , (p,~i) = 0 i_~f ~i 6 [i"

We can (and will) assume t h a t the m a t r i x A = (ai/) satisfies


the f o l l o w i n g normalizing conditions:

a.. = 2 or 0, i = l,...,r;
ii

( 1 . 2 ) { if aii = 0, t h e n the f i r s t n o n - z e r o element among aii+k,


k = 1,2,..., is i.

The m a t r i x A is c a l l e d the C a r t a n m a t r i x of Lie s u p e r a l g e b r a G. We

emphasize that A depends on the c h o i c e of B.


Let T be a s u b s e t of {l,...,r}, consisting of those i for
which ~ is an o d d root.
605

It is c o n v e n i e n t to d e s c r i b e the p a i r (A,T) by a D y n k i n diagram.


It c o n s i s t s of r nodes, of the form O, @ and 0, w h i c h are called
white, grey and b l a c k respectively. The i-th node is w h i t e if i 6 T
and grey or b l a c k if i ~ T and a.. = 0 or 2 respectively. The
ll
i-th and j-th n o d e s are joined by laijaji I lines unless the case
D(2,1;e). We n o t e that always a..a.. = 0 implies a. = a.. = 0
l] ]i l] ]l
and that in the i-th row of A all the entries are n o n - p o s i t i v e
integers if a. = 2.
ii

Proposition 1.5. Let G be a basic classical Lie super algebra, let


H be a C a r t a n subalgebra and let B be a Borel subalgebra in G,
containing H.
a) The elements ei' fi' hi' i = i, .... r, g e n e r a t e G.
b) The elements hl,...,h r span H. They are linearly independent
for all G except for G = ~(n,n). For ~(n,n) there is a
unique linear dependence: (h1+h2n+l) + 2(h2+h2n) + .-. + (n-l)+
(hn_ 1 + hn+ I) + nh n = 0.
c) L e t G --> G be the u n i v e r s a l central extension of G. Then
= G if G ~ ~(n,n) and G = sZ(n+l,n+l) if G = ~(n,n).
d) G is u n i q u e l y determined up to i s o m o r p h i s m by the p a i r (A,T) .
Basic classical Lie superalgebras admit a Borel subalgebra B,
for w h i c h the c o r r e s p o n d i n g Dynkin diagram has the form represented in
Table i. The labels ci in the T a b l e are the coefficients of the de-
composition of the h i g h e s t root with respect to s i m p l e roots, s being
the n u m b e r of the e x c e p t i o n a l nonwhite node, r being the total num-
ber of nodes.
We c a l l this B a distinguished Borel subalgebra. The p a i r (A,T)
is u n i q u e l y determined by the D y n k i n diagram except for the case
~(2,n) and ~(2,1;~). The C a r t a n matrix of ~(2,1;~) is

0 1
D = -i 2 0
-i 0 2

and (3x3)-submatrix corresponding to the last 3 n o d e s of D y n k i n diag-


r a m of ~(2,n) is D I.
606

Table 1

G Diagram S r

1 1 1 1 1
A (m, n) O---O . . . . O---O .- .---O m+l m+n+l

2 2 2 2 2 2
B(m,n) ,m > 0 O---O •- .---O---O . . . . O=>O n m + n

2 2 2 2 2
B (0,n) O---O ••• O---O=>@ n n

1 2 2 2 1
C(n) ,n > 2 0---<3 ... O---O<=O 1 n

DD (m,n) O---O ... ---O---O . . . . n m + n

2 3 2 1
F (4 ) 0-----O <=O---O 1 4

2 4 2
G~( 3 ) O----<) <-~:O 1 3

D (2, i; ~) 20/O1
1 3
~ O 1

Let B be the distinguished Borel subalgebra of a basic classical


Lie superalgebra G. Let ei,fi, hi, i = 1 .... ,r, be the correspond-
ing generators. We introduce a distinguished Zz g r a d a t i o n G = O Gi
ieZ
of G by putting :

deg hi = 0, i = 1 ..... r; deg e i = deg fi = 0, i ~ s;

deg es = -deg fs = i.

We call Lie superalgebras ~(m,n) and ~(n) superalgebras of I


type and all the other basic classical Lie superalgebras - superalgeb-
ras of II type.
607

Proposition 1.6. The distin@uished ~-@radation G = ~ G i of a b a s i c


i~Z
classical Lie superalgebra G satisfies the following ~roperties.
a) Gi = 0 for lil > 1 for superalgebras of I type and Gi = 0
for lil > 2 for s u p e r a l q e b r a s of II type.
b) GO is a d i r e c t s u m of a o n e - d i m e n s i o n a l center C (except for
~(n,n), when C = 0) and a semisimple Lie algebra [G0,G0], whose
Dynkin diaqram is a s u b d i a g r a m of the D y n k i n diagram of L i e s u p e r -
algebra G with a non-white (~-th) node removed.
c) The representations of Go and Gi and G_i are dual. The
lowest weight of G0-module Gi is ~s" The lowest w e i g h t of
r
G0-module G2 in case of superalgebras of II type is ~ = .Z ci~i,
l=S
where c. are the labels in T a b l e i.
l
d) The system ~' = {~i' i ~ s} for s u p e r a l g e b r a s of I t y p e and
~' = {ai' i ~ s; d} for superalqebras of II type is a s y s t e m
of simple roots for G~.

5. Weyl group and functions L and K[2]. Let W be the W e y l group


of the e v e n part G[ of a b a s i c classical Lie superalgebra G = G~G[.
Every element w ~ W can be w r i t t e n as a p r o d u c t of a c e r t a i n number
of r e f l e c t i o n s with respect to the even roots. We introduce two h o m o -
morphisms g and g' : W --> {±i} as f o l l o w s . Put g (w) = 1 if the
number of r e f l e c t i o n s in the expression of w is e v e n and g (w) = -i
if the n u m b e r ~ odd; put g' (m) = 1 if the n u m b e r of r e f l e c t i o n s with
respect to the roots of ~: in this expression is e v e n and g' (w) = - i
if the n u m b e r is odd. g and g' are w e l l - d e f i n e d . It f o l l o w s from
Proposition 1.7 below.
For A ~ H* let D(A) be the set of linear functions on H* of
the form: A ~ ~ n ~, n ~ ~+.
X = A ~ e-A +En ~ e D(A) deg
For w e set ~(1) = (-l)Zn~
Let E be the algebra of functions on H* that vanish outside
of the u n i o n of finitely many sets of the form
D(A), w i t h the convolu-
tion operation, eI ~ E is d e f i n e d by X
e (X) = i, e l (P) = 0 for
p ~ X. Clearly, e X ' e ~ = e X+D. We set:

L = ~+(e ~/2 - e - ~ / 2 ) / ~ + ( e ~/2 + e -~/2)


~ &0 ~EA i

L' = ~c-A:
- -(ea/2 e-~/2)/~+(ea/2-a~Al e-~/2)

We d e f i n e the K o s t a n t function K on H* by s e t t i n g K(I) equal


+
to the n u m b e r of sets {n~}, n d E ~+ for ~ e A 0 and n ~ = 0,1 for
608

+
6 Al such that [ + n e = -I. We set K' (I) = ~ ( 1 ) K ( 1 ) .
~A

Proposition 1.7. a) The sets A, A 0, At' 50' ~i ar___~e W - i n v a r i a n t .


b) W acts transitively on ~l and the function ~ e i_ss W - i n -
-+
variant. ~6A i
+
c) For superalgebras of I type the set AI is W-invariant; In
particular, w(01) = 01 for w ~ W.
d) Let T be a s u b g r o u p in W generated by r e f l e c t i o n s with res-
+ -+
pect to the r o o t s from A0 - A0
(i) W is a s e m i d i r e c t product of the n o r m a l subgroup T and
a subgroup W0 such that e' (w0t) = e(w 0) for w 0 ~ W 0,
t ~T.
(ii) ~ (e ~/2 - e -~/2) i__ss T - i n v a r i a n t ;
-+
~eA 0
(iii) the o r d e r of T is g i v e n by the following formula:
#T = 1 i_~f G ~ ~(m,n) or G(3),
#T = 2 n for B(m,n) ,
#T = 2 for ~(3).
e) w(L ) = e(w)L, w(L') = s' (w)L' for w E W.
f) L . K = L'-K' = e 0.

Proof. a) is e v i d e n t , b) and c) can be c h e c k e d case by case verifica-


tion. Also from the list of the root systems A0 it is c l e a r that
A0 ~ ~0 only for G = ~(m,n) or ~(3), i.e., (W:W 0) ~ 1 only for
these superalgebras. For t h e s e two cases d) f o l l o w s f r o m the structure
of W e y l group of Lie algebra ~n" For the p r o o f of e) and f) see [4].

§2. Finite-dimensional representations of b a s i c


classical Lie superalgebras.

1. Universal enveloping superalgebra and induced representations ([i],


[3]). For associative .~2-graded algebra A a bracket is d e f i n e d by:

[a,b] = ab - (1) ( d e g a ) ( d e g b ) b a

This bracket defines a Lie superalgebra A L-


A pair (U(G),i) where U(G) is an a s s o c i a t i v e 12-graded al-
gebra and i : G --> U(G) L is a h o m o m o r p h i s m of Lie superalgebras, is
called a universal enveloping superalgebra of G if for any o t h e r
pair (U',i') there is a u n i q u e homomorphism @ : U --> U' for w h i c h
i' = e o i.
609

The universal enveloping superalgebra of G = G~ @ Gy is con-


structed as follows. Let T(G) be the tensor superalgebra over the
space G with the induced ~2-gradation, and R the ideal of T(G)
generated by the elements of the form:

[a,b] - a @ b + (-i) (deg a)(deg b) b @ a .

We set U(G) = T(G)/R. The natural map G --> ~(G) evidently induces
a homomorphism i : G --> ~(G)L, and the pair (U(G),i) is the re-
quired enveloping superalgebra.
In Lie superalgebra case the Poincar4-Birkhoff-Witt theorem has
the following form.

The Poincar4-Birkhoff-Witt Theorem. Let G = G~ @ Gy be a Lie super-


algebra, al,...,a m a basis of G~, bl,...,b n a basis of G 7. Then
the elements of the form

kl
kmb .... b , where k. > 0 and 1 < i < .-- < i < n ,
ai • •- am 1I is 1 -- -- s --

form a basis of ~(G).

Let V = V~ @ VT be a ~2-graded linear space. Alinear represen-


tation ~ of a Lie superalgebra G = G~ @ GT in V is a homomorphism
~:G --> ~ ( V ) .
For brevity we often say in this case that V is a G-module, and
instead of ~(g) (v) we write g(v), g ~ G, v e V. Note that, by de-
finition, G i ( V j) ~ V i + j, i,j 6 ~z, and [gl,g2] ( ) = gl (gz (v)) -

- (-i) ( d e g g l ) ( d e g g z ) g 2 ( g l (v)). Note also that the map a d : G --> %(G)


for which (ad g)(a) = [g,a] is a linear representation of G. It is
called the adjoint representation.
A submodule of a G-module V is assumed to be ~2-graded; a
G-module V is said to be irreducible if it has no non-trivial sub-
modules. By a homomorphism of G-modules ~ : V --> V' we mean one that
preserves the ~2-gradation in the sense that ¢ ( V i) = Vi(i), where
is a bijection !2 --> ~ 2 -
Schur's lemma in Lie superalgebra case has the following form.

Schur's Lemma. Let V = V~@Vi, ~ an irreducible family of operators


from ~(V) , a n d C(~) = {a ~ ~(V) I [a,m] = 0, m ~ ~}. Then either
C~) = <i> o_rr dimV~ = dimVy and C~) = <I,A>, where A is a non-
degenerate operator in V permuting V[ and Vy, and Az = I.
610

By the symmetric superalgebra over ~2-graded space V = V~ • Vy


k
we m e a n an a l g e b r a S(V) = ~ sk(v), w h e r e sk(v) = ~ (sr(v~) @ A k - r ( v y ) )
k~0 r=0
is a Z -graded space
l2 with a natural gradation induced from V. Ana-
l o g o u s l y the e x t e r i o r superalgebra over V is an a l g e b r a
A(V) = • Ak(v), where Ak(v) = ~ (Ar(V~) @ sk-r(vy)) is a ~2-graded
k~0 r=0
space.
Any representation of Lie superalgebra G in V produces repre-
sentations of G in sk(v) and Ak(v) ; they are called symmetric and
exterior powers of G-module V. The definition of t h e s e representa-
tions in "super" case differs from the usual case only by a m i n u s sign
which we put each time there is a p e r m u t a t i o n of two n e i g h b o u r i n g odd
elements.
Let G be a Lie superalgebra, ~(G) its u n i v e r s a l enveloping
superalgebra, H a subalgebra of G, and V an H-module. V can be
extended to a ~(H)-module. We consider the ~2-graded space
~(G)@~(H)V (this is the factor space of U(G) @ V by the linear span
of the e l e m e n t s of the form gh @ v - g 8 h(v), g e ~(G), h ~ H(H),
with an o b v i o u s ~2-gradation). This space can be e n d o w e d with the
structure of a G-module as follows: g(u@v) = gu@v, g ~ G, u ~ U(G),
v V. The so c o n s t r u c t e d G-module is said to be i n d u c e d from the
H-module V and is d e n o t e d by Ind~V.
We list some of the simplest properties of induced modules, which
follow from the P o i n c a r 4 - B i r k h o f f - W i t t theorem.

Proposition 2.1. a) Let G be a Lie superal@ebra, H a subal@ebra,


V a simple G-module, and W an H-submodule of V considered as ,an
H-module. Then V is a f a c t o r module of the G-module Ind~W.
b) If H z ~H I C G are subalgebras of G and W is an
H2-module, then I n d G 1 (IndH2W)HI -~ I n d G 2W.
c) Let H C G be a s u b a l g e b r a of G containing G~, and
gl,---,gt be o d d elements of G whose projections onto G/H form a
basis, Let W be an H-module. Then
indGW = • ''<i <t gi '''''gi W is a d i r e c t s u m of s u b s p a c e s ; in
i<ii<
-- 1
s--G s
particular, dim I n d H W = 2 t d i m W.

2. Representations with the h i g h e s t weight [i]. From n o w on G=G~@G T


is one of t h e b a s i c classical Lie superalgebras ~(m,n), m ~ n, ~(m,n),
~(n), ~ ( m , n ) , ~(2,1;~), ~(4), ~(3). We e x c l u d e the e x c e p t i o n a l c a s e of
~(n,n) (see P r o p o s i t i o n 1.5c)) for simplicity of e x p o s i t i o n . All the
statements below hold for the central extension si(n+l,n+l) of A(n,n).
611

Let H be a C a r t a n subalgebra of G~. From now on we fix a B o r e l sub-


algebra G, containing H; then B = H (9 N +.
Let A 6 H* be a linear function on H. We define a one-dimen-
sional even B-module <vA> by:

h ( v A) = A ( h ) v A, h e H; N+(VA) = 0; deg vA = O .

G
We set V(A) = IndB<VA>. This G-module contains a unique maximal sub-
module I(A); we set V(A) = V(A)/I(A). G-module V(A) is c a l l e d an
irreducible representation with highest weight.

Proposition 2.2. a) v
+
A is . a . u n. i q.u e . v .e c t. o r
.
in V(A) up to a c o n s t a n t
factor, for which N (vA) = 0. In p a r t i c u l a r , for G-module V(A) the
first case of S c h u r ' s lemma always takes place.
b) The G-modules V(Ai) and V(A~) are isomorphic if a n d o n l y
if A = A .
- - i 2

c) If V is a q u o t i e n t module of the G-module 9(A) then the


weight decomposition takes place:

V = (9 Vl, where V1 = {v 6 V l h ( v ) = l(h)v, h ~ H},


teD(A)
all Vl being finite-dimensional.

d) Any finite-dimensional irreducible representation of G is


one of V(A).

Proposition 2.3. Let B be a distinguished Borel subalgebra of G


(the c o r r e s p o n d i n g Dynkin diagram is c o n t a i n e d in T a b l e i). Let e i,
fi' hi' i = l,...,r, be the g e n e r a t o r s of G described in sec. 2.1.
Let A ~ H*; we set a i = A(hi) , i = 1 ..... r. The representation V(A)
is f i n i t e - d i m e n s i o n a l if a n d o n l y if t h e following conditions are sa-
tisfied:
i) a i e ~+ for i ~ s;
2) for s u p e r a l g e b r a s of II t y p e k @ ~+, where k is g i v e n by
the following table:
612

Table 2

G k b

B(0,n) ½ an 0

B(m,n) , m > 0
a n - an+ 1 ..... am+n_ 1 - ~am+ n m

D~(m, n) a . . . . . . . ½( m
n an+ 1 am+n- 2 am+n- l+am+n )

m(2,1;~) (i+~)-i(2ai - a 2 - e a 3) 2

F(4) i / 3 ( 2 a I - 3a 2 - 4a 3 - 2a 4) 4

G(3) ½(a I - 2a 2 - 3a 3) 3

3) for k < b (in table 2) there are the supplementary conditions:

~(m,n) : an+k+ 1 . . . . . am+ n = 0.

0~(m,n) : an+k+ 1 . . . . . am+ n = 0, k ~ m-2; a m + n _ 1 = a m + n, k = m-l.

~(2,1;~) : all ai = 0 i_~f k = 0 ; (ao+l)~ = ±(a2+l) if k = i.

F(4) : all a = 0 if k = 0; k ~ i; a2 = a 4 = 0 if k = 2;
a2 = 2a~ + 1 if k = 3.

G(3) : all ai = 0 if k = 0; k ~ i; a2 = 0 if k = 2.

Remarks. a) Properties l) and 2) (in p r o p o s i t i o n 2.3) are equivalent


to the property of A being dominant for Lie algebra G-.
0

b) For Lie superalgebras of I type a is an arbitrary complex num-


s
ber, i.e. there exist parameter families of finite-dimensional repre-
sentations. For Lie superalgebras of II type there is n o such
family. Note however, that for both cases the set of A for which
dim V(A) < ~ is a Zariski dense set in H*.
Now we give a more explicit construction of the finite-dimensional
G-modules V(A). Let G = i~Gi be the distinguished ~-gradation
of G. We set P = i@0Gi , P+ = i~0Gi . Let A E H* be a linear for~
which satisfies the conditions i) - 3) of Proposition 2.3.
Let V°(1) denote a simple G0-module with highest weight I.
613

By c o n d i t i o n i) G 0 - m o d u l e V°(A) is f i n i t e dimensional. We extend


V0(A) to a P-module by p u t t i n g P+(V°(A)) = 0. We take the induced
module Ind~V°(A). When G is of II type this module contains a sub-
module,

k+l
M = U ( G ) G _ ~ v A.

Here G_6 is a (1-dimensional) root space, w h i c h c o r r e s p o n d s to


r
the e v e n root 6 = E c.e. (see P r o p o s i t i o n 1.6 c)) k = 2 ( A , ~ ) / ( d , 6 )
l=s 1 1
is an i n t e g e r g i v e n by T a b l e 2 and vA is the h i g h e s t w e i g h t v e c t o r
of V°(A) (see the p r o o f of t h e o r e m 8 in [i]). We set

V(A) = Ind~V°(A) if G is of I type,

V(A) = Ind~V°(A)/M if G is of II type.

From the p r o o f of t h e o r e m 8 in [i] we obtain,

Proposition 2.4. Suppose that A ~ H* satisfies conditions i) - 3)


of P r o p o s i t i o n 2.3. Then G-module ~(A) is f i n i t e - d i m e n s i o n a l ;
V(A) contains a unique maximal submodule T(A), and V(A) =V(A)/~(A).

3. Characters, supercharacters, and infinitesimal characters. Let V


be a quotient module of G-module V(A) and let V = I~D(A)VI be the
weight decomposition of V. We set

chV : [ (dimVl)e I , schV = [ ~(1)(dimVl)el


l~D (A) leD (A)

The functions ch V and sch V are called the c h a r a c t e r and super-


character, respectively, of G-module V.
Let Z(G) be the center of the enveloping superalgebra U(G).
An element z ~ Z(G) can be u n i q u e l y written in the form:

z = u z + E u ? u °+u. , where Uz,U° c U(H), u +. E N-+U(N-+) •


i 1 1 1 1

The m a p z --> u ° gives a monomorphism (Harish-Chandra homomorphism) :


z

B : ZCG)--> SCH) (: ~ [ H * ] ) .
614

We d e f i n e an a u t o m o r p h i s m e:C[H*] --> ~ [ H * ] by s e t t i n g ~(P(1)) =


P(l-p) and set I = doB: Z(G) --> S(H).
We n o t e t h a t if v~ is an e i g e n v e c t o r of the B o r e l subalgebra
B, then

z(v X) = U z ( X ) v I •

Therefore all the e l e m e n t s of Z(G) act as s c a l a r o p e r a t o r s (Uz(A) I)


in V(A) and any q u o t e n t m o d u l e V of V(A). We have:

z(v) = XA(Z)V , z ~ Z(G), v ~ V,

where XA:Z(G) --> ~ is a h o m o m o r p h i s m . XA is c a l l e d the i n f i n i t e s -


imal c h a r a c t e r of G-module V (it d e p e n d s only on A).
We set B(A) = D(A) N {I ~ H * I X 1 = XA}.

Proposition 2.5. Let V be a q u o t i e n t module of G-module V(A),


A ~ H*.
a) For V = 9(A) the f o l l o w i n g formulas hold

ch V(A) = [ K(I-A)e I t sch V(A) = [ K' (l-A)e 1


X6D (A) I&D (A)

L ' c h V(A) = L ' - s c h V(A) = e A+P

b) L.chV = ~ clel+P , L'.schV = ~ c~e


I+p,^ where CA=C~=l.
16B(A) I~B(A)

c) (l+p,l+p) = (A+p,A+p) for k ~ B(A).


d) We d e f i n e the L a p l a c e operator A o__nn E by the f o r m u l a
A(e l) = (l,l)e I. Then L.chV and L'-schV are e i g e n f u n c t i o n s
of A with eigenvalue (A+p,A+p). In p a r t i c u l a r :

A(L) = (p,p)L, A(L') = (p,p)L'

Let 6:S(G) --> U(G) be the s u p e r s y m m e t r i z a t i o n map (it g e n e r a l -


izes the s y m m e t r i z a t i o n map in Lie a l g e b r a case; the d i f f e r e n c e is
that e a c h p e r m u t a t i o n of
!
two n e i g h b o u r i n g odd e l e m e n t s produces a
minus sign). Let S(G) ~ denote a subalgebra in S(G) of e l e m e n t s
which are a n n i h i l a t e d by the a c t i o n of G. Then 6 : S(G) ~ --> Z(G) is
an i s o m o r p h i s m . Finally let i:S(G) --> S(H) be the h o m o m o r p h i s m
i n d u c e d by the o r t h o g o n a l projection of G onto H. It i n d u c e s a
homomorphism i:S(G) ~ --> S(H) W
615

Proposition 2.6. a) y(Z(G)) ~ S(H) w.


b) i = yo6 is a m o n o m o r p h i s m S(G) ~ --> S(H) W. The fields of
fractions of i(S(G) 9) and S(H) W coinside.
c) Xw(A+p ) = XA+p for w ~ W, A ~ H*.
d) If A ~ H* and (A+p,~) = 0, ~ ~ ~i, then XA = X A _ t ~ for
any t 6 ~.

Proof. For a), b) a n d c) see [2]. To prove d) w e first note that


d) is true if the root ~ = as is simple. Indeed, let vA be the
highest vector of V(A) and (A+p,~ s) = (A,~ s) = 0. Then e_~ v A is
s
obviously an eigenvector for B. Therefore z(e_~ v R) = Uz(A-as)V A =
s
= XA_~ (z)e_a v A, z ~ Z(G), which proves that XA = X A _ ~ .
s s s
Now w(~) = ~ for some w E W. We have:
s

XA_ ~ = Xw(A_m+p)_ p = X(w(A+p)_p)_ a


s

We set M = w(A+p)-p ; then (M+p,~ s) = ( w ( A + p ) , a s) = (A+p,a) = 0.


Therefore XM_~ = XM and using again c) w e o b t a i n : X A _ e = XA.
s
Since (a,~) = 0 we obtain by iteration that X A _ t e = XA, t ~ Z+,
and therefore X A _ t ~ = XA for a n y t e ~C"

4. Typical representations. Our first main result is the following

Theorem I. Let A be the h i g h e s t weight of a finite-dimensional ir-


reducible module V(A) of a basic classical Lie superalgebra G.
Then the following statements are equivalent.
a) If X A = Xl for s o m e X ~ H*, then l+p = w ( A + p ) for some
w~W.
b) If a l l the p o l y n o m i a l s of i(S(G) @) have the same value at
A and at some X ~ H*, then l+p = w ( A + p ) for s o m e w ~ W.
c) V(A) is u n i q u e l y determined u p to an i s o m o r p h i s m among the
irreducible finite-dimensional G-modules by its i n f i n i t e s i m a l charac-
ter XA.
d) V(A) splits in a n y finite-dimensional representation (i.e.
if V(A) is a s u b m o d u l e or a factor-module of a finite-dimensional
G-module, then V(A) is a d i r e c t summand).
e) V(A) = V(A).
f) V(A) is a n irreducible G-module.
616

fai+l
g) VCA) = V(A)/I(A), where I(A) = E U(G) VA for G of
i~s i
ai+l k+l
I type and I(A) = E U(G) f i VA + U(G)G_~ VA for G of II type
igs
(a i = A(hi), k = 2(A,6)/(6,6)).
h) ch V(A) = L -I ~ e(w) e W ( A + 0 ) o
w&W
i) ch V(A)(I) = E K(l+p-w(A+p)).
wEW
j) sch V(A) = L '-I ~ E' (w)e w(A+0)
w~W
k) sch V(A) (l) = E K' (l+p-w(A+p)) •
weW
-+
£) (A+o,~) ~ 0 for any e ~ AI

Definition. A finite-dimensional irreducible G-module V(A) with


highest weight A is c a l l e d typical if one of the e q u i v a l e n t condi-
tions a) - £) of T h e o r e m 1 is s a t i s f i e d .
An essential part of T h e o r e m 1 is c o n t a i n e d in [2]. In t h e s e
notes ~nis theorem follows from Propositions 2.6, 2.7, 2.8, 2.12 and
Theorem 2.

Proposition 2.7. Let V be a f i n i t e - d i m e n s i o n a l G-module with high-


est w e i g h t A. We set B+(A) = {I ~ B(A) IX # A and l+p is d o m i n a n t
for G~}. Then the following formulas take place:

(2.1) LchV = weW[ g ( w ) e W ( A + P ) + %eB[+(A) cX(V)waW[ E(w) eW(l+P)

(2.1') L'schV = [ g' (w)e w(A+p) + [+ c~(V) [ ~' (w)e w(l+p) .


w~ W I~B (A) w~W

Proof. The action of G~ on V can be e x t e n d e d to the a c t i o n of the


corresponding connected simply connected Lie group G o , and in p a r t -
icular to the n o r m a l i z e r of H in G o. It p r o v e s that ch V and
schV are W-invariant functions. Since any p 6 H* can be trans-
formed by W into a dominant for G~ function, by P r o p o s i t i o n 2.5 b)
and 1.7 e) we o b t a i n (2.1) and 2.1')o
617

Proposition 2.8. Let V be a f i n i t e - d i m e n s i o n a l G-module with high-


est w e i g h t A, w h i c h satisfies property a) of T h e o r e m i. Then

(2.2) ch V = L -I [ E(w) e w(A+p)


w6W

(2.2') sch V = L '-I [ e' (w)e w(A+p)


weW

(2.3) d i m Vl = [ K(l+p-w(A+p))
weW

(2.3') '~(1)dim Vl = [ K' (l+p-w(A+p)) o

w&W

Proof. Since B+(A) = ¢ if a) is s a t i s f i e d , (2.2) and (2.2') follow


from (2.1) and (2.1'). Formulas (2.3) and (2.3') are e q u i v a l e n t forms
of (2.2) and (2.2') (cf. Proposition 1.7 f).

The following implications in T h e o r e m 1 are e i t h e r evident or


follow from Propositions 2.6, 2.7, 2.8:

Z) <-- a) <--> b) --> c) - - > d) - - > e) <--> f) <--> g)

h) <--> i) <--> j) <--> k)

We also show how to p r o v e that f) <--> b) for Lie superalgebras


G of I type. Together with preceding remarks and T h e o r e m 2 below for
G [2] it p r o v e s Theorem 1 for superalgebras of I type.

Proposition 2.9. Let G be a Lie superalgebra of I type. Then G-


-+
module V(A) is i r r e d u c i b l e if and only if (A+p,~) ~ 0, for ~ 6 A I.

+ -+
Proof. We n o t e that AI = AI for Lie s u p e r a l g e b r a s of I type, and
+
that [G ,G B] = 0, if e,B ~ A I. We set: T = ~ e , T+ = ~ + e ~ .
- ~A~ -~ ~A~
Clearly any submodule of V(A) contains the e l e m e n t T_v A and there-
fore V(A) is i r r e d u c i b l e if a n d o n l y if T + T _ v A ~ 0. We consider
G-module V(A) for any A ~ H*. We h a v e the following relation:
T + T _ v A = P ( A ) v A, where P(A) is a p o l y n o m i a l of d e g r e e equals to
#A~. From Proposition 2.6 w e o b t a i n that P(A) ~ 0 and from the
p r o o f of this proposition that P(A) is d i v i s i b l e by the linear form
¢(A) = (A,~ s) = (Aep,e s) where ~s is the unique odd simple root.
618

Suppose that A(hi) C !+ for i ~ s. Since evidently fi,i ~ s,


commutes with T+T_ , we o b t a i n that P(A) = P(A-(A(hi)+l)~i), i ~ s,
and t h e r e f o r e that P(A) = P(w(A+p)-p), w ~ W. Since the last
equality holds for Zariski dense set in H*, it h o l d s for any A E H*.
+
Since W acts transitively on Ai we o b t a i n now that P(A) is
divisible by
~+(A+p,e). Since P(A) is a p o l y n o m i a l of d e g r e e
~A i
#AT, we o b t a i n t h a t P(A) = c~+(A+p,e), c ~ C* , w h i c h proves the
d6~ 1
proposition.
The following theorem, which is a g e n e r a l i z a t i o n of C h e v a l l e y
invariants restriction theorem to Lie superalgebra case, is a c r u c i a l
point in the p r o o f of T h e o r e m i.

Theorem 2. Any element of S(H) W can be r e p r e s e n t e d in the form


p/Qt, where p e i(S(G) ~) Q = ~ h e i(S(G) ~) t E ~+.
ae~
The proof of a s l i g h t l y weakened f o r m of T h e o r e m 2 can be found
in [2]. We n o t e that Theorem 2 implies that if the property b) of
Theorem 1 holds then the W-orbit of A can be s e p a r a t e d from any
other W-orbit by p o l y n o m i a l s from i(S(G)@). Therefore, by P r o p o s i -
tion 2.6 we o b t a i n the implication Z) --> a) of T h e o r e m i.

Remark. It f o l l o w s from the results above that all the finite-


dimensional representations of G are t y p i c a l if and only if
G = B(0,n) (since Ai = ¢ <=> G = ~(0,n)) . This result is c o n s i s t e n t
with Djokovic-Hochshild Theorem [6] t h a t any finite-dimensional rep-
resentation of a Lie superalgebra G is c o m p l e t e l y reducible if and
only if G is a d i r e c t sum of a s e m i s i m p l e Lie algebra and several
copies of superalgebras of type ~(0,n).

Now we will compute the d i m e n s i o n s of t y p i c a l representations and


their even and odd parts [2]. We introduce a homomorphism
fp: E --> ~ [ [ t ] ] , where t is an i n d e t e r m i n a t e , U ~ H*, by s e t t i n g

f (e x) = e t ( X , P )

Two cases of f are especially important:

f = f and f = f-
P0 P0
619

ProPosition 2.10. Let G be a basic classical Lie superalgebra. We


set d = #AT , d = #~T • Let V by a typical G-module with highest
weight A. Then

(2.4) dim V = 2 d ~ + (A+p,~)


~A ° (P0 '~)

(2.5)* f(sch V) = t ~ ~ (A+p,e) ( ~ o , c~) + o(t d)


-+o (Fo,~) -+
c~6A c~cA 1

I dim V~ - dim V T = 0 if G / B(O,n),

(2.6) dim VE - dim Vy = (A+p, cO


if G = B(0,n) .
-+
c~&Ao (5o ,~)
-U
Proof. Clearly, dim V = lira f(ch V). Let c = #A'0 . We have:
t+0
f[ ~ e(w) eW(A+P)] = ~ ~(w) e(W(A+p),P0)t = ~ ~(w)e(A+p,w(P0) )t
w6W wcW w~W

fi+@I~ [ s(w)eW(@0)] ) = (by Weyl denominator formula) =


w~W

= fA+p ~ + (e~/2 - e-~/2] : tc[~+(A+P'~)+~6A0 O(I)]" On the other hand

f(L) = tel2 -d ~+(p0,~) + o(1)]. Hence f(eh V) = 2 d ~+(A+p,e)/


~6A 0
(O0,~) + o(1), which proves (2.4).

For the proof of formula (2.5) we use the Weyl denominator for-
mula for the root system [+0:

~(w) eW(~°) = ~-I (e~/2 - e-~/2)

From Proposition 1.7 d) we obtain the following form of this identity:

(2.7) [ e' (w) e w(~°) = #T ~ (e ~/2 - e -e/2]


w6W -+
~eA 0

Now we apply f to both sides of formula (2.2'). The same argu-


ment as before together with (2.7) gives (2.5).
Since dim V 0 - dim V I = lim sch V and d = 0 if and only if
t÷0
G = B(0,n), formula (2.5) gives (2.6).

* There is a misprint in the formula for f(sch V) in the article


[2] (in the case ~(m,n)).
620

-+
Remark. If G = B(0,n), then clearly (P,e) = (~0 ,~) for ~ e A0
and w e can rewrite (2.6) as follows (see [4]):

d i m V~ - dim V 7 = ~ (A+p,e)
-+ (p,~)
~6A o
Examples. i. ~(m,n). A finite-dimensional representation V(A) is
typical if and o n l y if*

i m
am+ 1 ~ at - ~ a t - 2m - 2 + i + j ,
t=m+2 t=l

for 1 < i < m + 1 < j < m + n + i. Under these conditions

dim~A)=2 (m+l)(n+l) ~ ai+ai+l +..-+aj+j-i+l ~ ai+...+aj+j-i+l


j-i+l 1 1 j-i+l
3 _<i
_ 3<_''<m m+2<i_< j < m + n + 1

For a non-typical finite-dimensional representation d i m V(A) < right


h a n d - side.
2. C(n). A finite-dimensional representation V(A) is
typical if and o n l y if

i
al ~ [ at + i - 1
t=2

i n

a1 [ at + 2 [ a t + 2n - i - i,
t=2 t=i+l

for 1 < i < n-1. Under these conditions

dimV(A) = 2 2n-2 ~ a'+°''+aj+j-i+ij


j-i+l ~ ai+'''+aj-l+2aj+°''+2an2n-i-j+2
2<i<j<_n-i 2<i<j<n

For a non-typical finite-dimensional representation d i m V(A) < right


hand-side.
3. ~(m,n). A finite-dimensional representation V(A) is
typical if and o n l y if

n j
at - Z a t + 2n - i - j ~ 0
t=i t=n+l

J =
* We assume that t~i 0 if j < i.
621

n j m+n-i
[ at - [ at - 2 [ at - i + j - 2m + 1 ~ 0
t=i t=n+l t=j+l - am+n

for 1 < i < n < j < m + n - i. Under these conditions

dimV(A) = 2( 2 m + l ) n ~ ai+'''+aj+j-i+lj-i+l ~ ai+'''+aj+j-i+lj-i+l

l~i<j~n-i n+l<i<j<m+n-i

( a i + ° ' ' + a j _ I) + 2 ( a j + ' . . + a n - a n + 1 ..... am+n_ ~ - am+n +2n-2m+l-i-j

l<i<j<n 2n + 2 - i - j

( a i + ' - ' + a j _ I) + 2(aj+--.+am+n_ 1 + am+ n + 2m- i -j + 1

n+l~i~j~m+n-i 2m - i - j + 1

~(0,n). All the finite-dimensional representations V(A)


are typical.

( a i + . . - + a j) + 2(aj+l+.-.+an_l)+an+2n-i-j
dimV~(A) - dimVy(A) =
2n - i - j
l < i < j_<n

+ 2n-2i+l
2 ( a i + .... ~an_ I) + a
dimV(A) = (dimV~(A)-dimy(A)) n
2n - 2 i + 1
l<i<n

~. ~(m,n). A finite-dimensional representation V(A) is


typical if a n d only if

n j
[ at - [ at + 2n - i - j ~ 0
t=i t=n+l

for 1 < i < n < j < m+n-l,

n m+n- 2

I at - a t - am+ n + n-m-i + 1 ~ 0
t=l t=n+l

for 1 < i < n, and

n j m+n- 2
[ at - [ at - 2 [ a t - am+n_l-am+n-i+j-2m ~ 0
t=i t=n+l t=j+l

for i ! i ! n ! j ! m+n-2o
622

The formula for dimV(A) is similar to o n e for B(m,n) .


5. ~(2,1;~)D . A finite-dimensional representation V(A) is
typical if a n d only if a I # 0, a I ~ a 2 + i, a I ~ ~a 3 + ~,
a I ~ a 2 + ea 3 + 1 + ~. Under these conditions

dim V(A) = 16(a 2 + i)(a 3 + i)I(l+~)-l(2a 1 - a2 - ~a3)-l)"

6. G(3). A finite-dimensional representation V(h) is


typical if a n d only if a I ~ 0, a I ~ a s + I, a z ~ a 2 + 3a 3 + 4,
a I # 3a 2 + 3a 3 + 6, a I # 3a 2 + 6a 3 + 9, a z ~ 4a 2 + 6a 3 + 10.
Under these conditions

dim V(A) = l~(a2+l) (a3+l) ( a 2 + a 3 + 2 ) ( a 2 + 3 a 3 + 4 ) ( a 2 + 2 a 3 + 3 ) ×

(2a2+3a3+5) (al-2a2-3a3-5)

7. ~(4). A finite-dimensional representation V(A) is


typical if and only if

aI ~ 0, a I ~ a2+l , a I ~ a2+2a +3 , a I ~ 2a2+2a3+4

a I ~ a2+2a3+2a4+5, a I ~ 2a2+2a3+2a4+6,

a z ~ 2a2+4a3+2a4+8, a I ~ 3a2+4a3+2a~+9.

Under these conditions

32
dim V(A) = ~(a2+l) (a3+l) (a4+l) ( a 2 + a 3 + 2 ) ( a 3 + a 4 + 2 ) ( a 2 + 2 a 3 + 3 ) ×

(az+a3+a4+3) (a2+2aa+2a4+5) (a2+2a3+a4+4) (2al-3a2-4a3-2a4-9)

Proposition 2.11. Let V be a typical G-module with highest weight


A. Let mA(~) denote the multiplicity of an irreducible G~-module
with (dominant) highest weight ~ inn G~-module V, and let KI (1)
denote the number of partitions of I ~ H* into a sum of distinct
+
roots from A I. Then the following formula takes place:

(2.8) mA(P) = [ s(w)K1(w(A+p) - (p+p))o


w~W
623

Proof. Let V° (U) denote an i r r e d u c i b l e G~-module with highest


weight ~. By f o r m u l a (2.2) we have:

ch V(A) = ~+(l+e -~) ~ + ( l - e - ~ ) -I [ g ( w ) e w ( A + p ) - p =


~A ~ ~6A 0 weW

= [ mA(v)~+(l-e-~) -I [ e w ( A + p 0 ) - p 0
v ~ ~A 0 weW

= I
~+
eA 0
(l-e-a) ]i I g(w) l m A [ w - l ( ~ + p 0 ) - P 0 ) e ~ =
weW

= I+0 ]i (1-e -s ) [ g(w) [ m A ( ~ ) e ~


w6W
,

because mA(w-l(~+p0)-P0) = 0 for w ~ e, since w-l(~+P0)-P0 is


not a d o m i n a n t function for w # e. Therefore we have:

mA(u)e~ = ~+(l+e -e) [ g(w)e w ( A + p ) - p ,


~A i w

which is e q u i v a l e n t to (2.8).

Remark. T h e p r o o f of P r o p o s i t i o n 2.11 is the same as K o s t a n t ' s proof


of his f o r m u l a of m u l t i p l i c i t i e s with respect to a r e g u l a r subalgebra
of a s e m i - s i m p l e Lie a l g e b r a . There is an o b v i o u s generalization of
both formula (2.8) and K o s t a n t ' s formula for the m u l t i p l i c i t i e s with
respect to a r e g u l a r subalgebra of G~.

Example. Let G = ~(i,0) and let V = V(A), where A(hl) = k 6 Z+,


A(h2) 6 ~, be a G-module. Formula (2.8) gives the following decom-
position of the G-module V with respect to G~ = g£2:

V % = V°(A) @ V ° ( A - a l - 2 a 2)

V] = V°(A-~2) @ v°(A-al-a2) for k > 1

and V~ = V° ( A - a 2 ) for k = i.

Here V °(I) denotes a G~-module with highest weight I. Their dimen


sions are r e s p e c t i v e l y :
dim V°(A) = dim V°(A-~1-2~2) = k + i,
d i m V ° ( A - ~ 2 ) = k + 2, dim V°(A-~1-~2) = k.
624

A non-typical G-module sks~(2,1) is i s o m o r p h i c to V(A)/~(A),


where A(h I) = k, A(h 2) = 0 and w h e r e [(A) is an i r r e d u c i b l e
G-module with highest weight A - ~2"

5. Bilinear f o r m on U(G) and the s t r u c t u r e of m o d u l e s V(A). In


the smme w a y as in [7] we i n t r o d u c e a bilinear form
A:U(G) ® ~ U(G) ~ U(H). For t h a t we d e n o t e by a an i n v o l u t i v e anti-
automorphism of U(G), w h i c h is i d e n t i c a l on H and for w h i c h
o(N +) = N-. We h a v e d e c o m p o s i t i o n U(G) = U(H) @ (N_U(G) + U(G)N+).
Denote by B the p r o j e c t i o n B:U(G) ~ U(H) . We set:

A(x,y) = B(o(x)y) , x,y e U(G) .

We h a v e a w e i g h t decomposition of U(N ) with respect to H:


i

U(N_) =@ U(N )-q. Let A be the r e s t r i c t i o n of A to U(N )-~


n6H* - ~
= Zkie i, k i ¢ Z+. The same p r o o f as in [8] g i v e s the f o l l o w i n g
result (cf. [5] and [7]).

Theorem 3. a) U p to a n o n - z e r o constant factor

det A n I I I i ha+ (P,e)-n~r-LJ [ I NB+(P,B)-


n>0 0

b) If G - m o d u l e V(A) has a subquotient isomorphic to V(~), then


there exists a chain A = ~0,~l,...,~k = ~ such t h a t ~i+l=~i-niSi ,
where B i 6 A +, ni is a p o s i t i v e integer, and

2(~i+ P,8 i) = n i ( S i , S i ) .
c) If t h e r e e x i s t s a chain A = ~0,~i,. "''~k = ~ with properties
+
described in b), and in a d d i t i o n 8i ~ Bi+l' ni = 1 for ~i c AI,
then there exists a homomorphism of G - m o d u l e s : V(~) -~V(A).

Theorem 3 h a s the f o l l o w i n g consequence for f i n i t e - d i m e n s i o n a l


representations.

Proposition 2.12. The statements f), h) , and ~) of T h e o r e m 1 are


equivalent.
625

6. Some r e m a r k s and open questions.


a) For s u p e r a l g e b r a s G of I type the e x p r e s s i o n for ch V(A)
is g i v e n by the r i g h t h a n d - s i d e of (2.2) (it f o l l o w s easily f r o m the
fact t h at w(p I) = Pl, w 6 W, in this case). Therefore, ch V(A) is
g i v e n by r i g h t h a n d - s i d e of (2.4) and for a n o n - t y p i c a l G-module
V(A) we h a v e (cf. (2.4) and (2.8)):

(2.9) d i m V(A) < 2d ~ (A+p,~)


~A+ (P0 '~)

(2.10) mA(~) < [ s(w)K1 (w(A+p) -- (~+p))


weW

One can s h o w t h a t (2.9) and (2.10) take p l a c e for s u p e r a l g e b r a s of II

type if 2(A+p,6)
(~,~) is a p o s i t i v e integer. It is i n t e r e s t i n g to find

the e x p r e s s i o n for ch V(A) for s u p e r a l g e b r a s of II type and c o r r e s -


p o n d i n g a n a l o g u e s of (2.9) and (2.10) in g e n e r a l case.
b) It is i n t e r e s t i n g to c o m p u t e cA(V) in f o r m u l a (2.1) for any
finite-dimensional G-module V; I do n o t k n o w the a n s w e r even for
the o n e - d i m e n s i o n a l V. By T h e o r e m 3 cx(V(A)) = 0 if A and ~ can
no t be c o n n e c t e d by a chain, described in T h e o r e m 3.
c) F i n d the i r r e d u c i b l e subquotients of G-modules V(A) (in the
non-typical case) (cf. T h e o r e m 3).
d) F i n d the f o r m u l a s of d i m V(A) and dim V0 (A) - d i m VI (A)
for n o n - t y p i c a l representations.
e) In the s i m p l e s t c a s e of G = ~(i,0) a complete list of non-
typical representations is sksz(2,1) a n d its dual, k ~ ~+. It in-
dicates t h a t all n o n - t y p i c a l representations can be c o n s t r u c t e d from
the s t a n d a r d representation by s o m e tensor operations. We note that
one-dimensional G-module for G ~ ~(0,n), standard representations
si(m,n) and osp(m,n), m ~ 0, a n d the a d j o i n t representations of
G ~ ~(i,0) or ~(0,n) are n o n - t y p i c a l (because d i m G~ ~ d i m Gy).
f) The supertrace form (a,b) = str a b of a G - m o d u l e V = V(A)
V
is n o n d e g e n e r a t e if and o n l y if (dim V~ - d i m V7) (A,A+2p) ~ 0, e x c e p t
for the case G = A(I,0) [i] ; in the l a t t e r case (a,b) v is non-
degenerate for all V. Proposition 2.8 shows that for G ~ A~(I,0) and
typical G-module V the f o r m (a,b) V is 0. F i n d the c o n d i t i o n s in
t e r m s of A fol (a,b)v(A) to be n o n - d e g e n e r a t e .
g) Conjecture (cf. P r o p o s i t i o n 2.6). X X = X~ if and only if
w(~+p) = X+p+~ for c e r t a i n w ~ W, w h e r e B = ZciB i, 8i ~ ~i,
ci ~ C and (X + p + ciB I + ... + c i B i ~ i +l ) = 0 i = D,~,
626

h) It follows from T h e o r e m l, that HI (G,V(A)) = 0 for a typi-


cal G-module V(A). One can be shown that dimH1(A(l,0),si(2,1)) = 1
and H I (~(i,0) ,V) = 0 for all t h e o t h e r modules V = V(A) . Compute
H I (G,V(A)) in general case.
i) Classify all the finite-dimensional indecomposable
G-modules. All the i r r e d u c i b l e quotients of any such r e d u c i b l e module
are e v i d e n t l y non-typical. Note that the m o d u l e s V(A) are indecom-
posable.

References
[i] V. G. Kac, Lie superalgebras, Advances in Math., 26, no. 1 (1977),
8-96.
[2] V. G. Kac, Characters of typical representations of c l a s s i c a l Lie
superalgebras, Communications in Algebra, 5(8) (1977), 889-897.
[3] L. Corwin, Y. Ne'emen, S. Sternberg, Graded Lie a l g e b r a s in
mathematics and physics, Rev. Mod. Phys. 47 (1975), 573-604.
[4] V. G. Kac, Infinite-dimensional algebras, Dedekind's q-function,
classical M6bius function and the very strange formula, Advances
in Math., to appear.
[5] I. N. Bernstein, I. M. Gelfand, S. I. Gelfand, Structure of rep-
resentations generated by v e c t o r s of h i g h e s t weight, Funk. Anal.
Appl., 5 (1971), 1-9.
[6] D. Z. D j o k o v i 6 and G. Hochschild, Semi-simplicity of z~graded
Lie a l g e b r a s II, Illinois J. Math. 20 (1976), 134-143.
[7] N. N. Shapovalov, On a b i l i n e a r form on the u n i v e r s a l e n v e l o p i n g
algebra of a c o m p l e x semisimple Lie algebra, Funk. Anal. Appl.
6 (1972), 307-312.
[8] V. G. Kac, D. A. Kazhdan, Representations with highest weight of
infinite-dimensional Lie algebras, to appear.

You might also like