Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras
We prove an explicit formula for a projection of singular vectors in the Verma module over a rank 2 Kac-Moody Lie algebra onto the universal enveloping algebra of the Heisenberg Lie algebra and of sl2 (Theorem 3). The formula is derived from a more general but less explicit formula due to Feigin, Fu...
Збережено в:
| Опубліковано в: : | Symmetry, Integrability and Geometry: Methods and Applications |
|---|---|
| Дата: | 2008 |
| Автори: | , |
| Формат: | Стаття |
| Мова: | Англійська |
| Опубліковано: |
Інститут математики НАН України
2008
|
| Онлайн доступ: | https://nasplib.isofts.kiev.ua/handle/123456789/149013 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Цитувати: | Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras / D. Fuchs, C. Wilmarth // Symmetry, Integrability and Geometry: Methods and Applications. — 2008. — Т. 4. — Бібліогр.: 6 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1860236893238591488 |
|---|---|
| author | Fuchs, D. Wilmarth, C. |
| author_facet | Fuchs, D. Wilmarth, C. |
| citation_txt | Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras / D. Fuchs, C. Wilmarth // Symmetry, Integrability and Geometry: Methods and Applications. — 2008. — Т. 4. — Бібліогр.: 6 назв. — англ. |
| collection | DSpace DC |
| container_title | Symmetry, Integrability and Geometry: Methods and Applications |
| description | We prove an explicit formula for a projection of singular vectors in the Verma module over a rank 2 Kac-Moody Lie algebra onto the universal enveloping algebra of the Heisenberg Lie algebra and of sl2 (Theorem 3). The formula is derived from a more general but less explicit formula due to Feigin, Fuchs and Malikov [Funct. Anal. Appl. 20 (1986), no. 2, 103–113]. In the simpler case of A11 the formula was obtained in [Fuchs D., Funct. Anal. Appl. 23 (1989), no. 2, 154–156].
|
| first_indexed | 2025-12-07T18:25:26Z |
| format | Article |
| fulltext |
Symmetry, Integrability and Geometry: Methods and Applications SIGMA 4 (2008), 059, 11 pages
Projections of Singular Vectors of Verma Modules
over Rank 2 Kac–Moody Lie Algebras?
Dmitry FUCHS and Constance WILMARTH
Department of Mathematics, University of California, One Shields Ave., Davis CA 95616, USA
E-mail: fuchs@math.ucdavis.edu, wilmarth@math.ucdavis.edu
Received June 29, 2008, in final form August 24, 2008; Published online August 27, 2008
Original article is available at http://www.emis.de/journals/SIGMA/2008/059/
Abstract. We prove an explicit formula for a projection of singular vectors in the Verma
module over a rank 2 Kac–Moody Lie algebra onto the universal enveloping algebra of the
Heisenberg Lie algebra and of sl2 (Theorem 3). The formula is derived from a more general
but less explicit formula due to Feigin, Fuchs and Malikov [Funct. Anal. Appl. 20 (1986),
no. 2, 103–113]. In the simpler case of A1
1 the formula was obtained in [Fuchs D., Funct.
Anal. Appl. 23 (1989), no. 2, 154–156].
Key words: Kac–Moody algebras; Verma modules; singular vectors
2000 Mathematics Subject Classification: 17B67
1 Introduction
Let G(A) be the complex Kac–Moody Lie algebra corresponding to an n × n symmetrizable
Cartan matrix A, let N−,H, N+ ⊂ G(A) be subalgebras generated by the groups of standard
generators: fi, hi, ei, i = 1, . . . , n. Then G(A) = N− ⊕ H ⊕ N+ (as a vector space), the Lie
algebras N− and N+ are virtually nilpotent, and H is commutative. Let λ : H −→ C be a linear
functional and let M be a G(A)-module. A non-zero vector w ∈M is called a singular vector of
type λ if gw = 0 for g ∈ N+ and hw = λ(h)w for h ∈ H. Let
Jλ = {α ∈ U(N−)| ∃ a G(A)-module M and a singular vector w ∈M
of type λ such that αw = 0}.
Obviously Jλ is a left ideal of U(N−). It has a description in terms of Verma modules M(λ).
Let Iλ be a one-dimensional (H ⊕ N+)-module with hu = λ(h)u, gu = 0 for g ∈ N+ and
arbitrary u ∈ Iλ. The Verma module M(λ) is defined as the G(A)-module induced by Iλ; as
a U(N−)-module, M(λ) is a free module with one generator u; this “vacuum vector” u is, with
respect to the G(A)-module structure, a singular vector of type λ. It is easy to see that M(λ)
has a unique maximal proper submodule and this submodule L(λ) is, actually, Jλu.
This observation demonstrates the fundamental importance of the following two problems.
1. For which λ is the module M(λ) reducible, that is, L(λ) 6= 0?
2. If M(λ) is reducible, then what are generators of L(λ) (equivalently, what are generators
of Jλ)?
Problem 1 is solved, in a very exhaustive way, by Kac and Kazhdan [5]. They describe
a subset S ⊆ H∗ such that the module M(λ) is reducible if and only if λ ∈ S; this subset is
?This paper is a contribution to the Special Issue on Kac–Moody Algebras and Applications. The full collection
is available at http://www.emis.de/journals/SIGMA/Kac-Moody algebras.html
mailto:fuchs@math.ucdavis.edu
mailto:wilmarth@math.ucdavis.edu
http://www.emis.de/journals/SIGMA/2008/059/
http://www.emis.de/journals/SIGMA/Kac-Moody_algebras.html
2 D. Fuchs and C. Wilmarth
a countable union of hyperplanes. (See a precise statement in Section 2 below.) Actually, λ ∈ S
if and only if M(λ) contains a singular vector not proportional to u.
A formula for such a singular vector in a wide variety of cases is given in the work of Feigin,
Fuchs and Malikov [6]. This formula is short and simple, but it involves the generators fi
raised to complex exponents; when reduced to the classical basis of U(N−) the formula becomes
very complicated (as shown in [6] in the example G(A) = sln). There remains a hope that the
projection of these singular vectors onto reasonable quotients of U(N−) will unveil formulas that
possess a more intelligible algebraic meaning, and this was shown to be the case by Fuchs with
the projection over the algebra A1
1 into U(sl2) and U(H), where H is the Heisenberg algebra [3],
work which took its inspiration from the earlier investigation of Verma modules over the Virasoro
algebra by Feigin and Fuchs [2].
In this note we extend these results by providing projections to U(sl2) and U(H) of the
singular vectors over the family of Kac–Moody Lie algebras G(A) of rank 2 (see Theorem 3 in
Section 4 and a discussion in Section 5). As in [3] and [2], our formulas express the result in the
form of an explicit product of polynomials of degree 2 in U(H) and U(sl2).
It is unlikely that this work can be extended to algebras of larger rank.
2 Preliminaries
Let A = (aij) be an integral n × n matrix with aij = 2 for i = j and aij ≤ 0 for i 6= j. We
assume that that A is symmetrizable, that is, DA = Asym, where D = [d1, . . . , dn] is diagonal,
di 6= 0, and Asym is symmetric. To A is associated a Kac–Moody Lie algebra G(A) defined in
the following way.
G(A) is a complex Lie algebra with the generators ei, hi, fi, i = 1, . . . , n and the re-
lations [hi, hj ] = 0, [hi, ej ] = aijej , [hi, fj ] = −aijfj , [ei, ej ] = δijhi, (ad ei)−aij+1ej = 0,
(ad fi)−aij+1fj = 0. There is a vector space direct sum decomposition G(A) = N− ⊕ H ⊕ N+
where N−,H, N+ ⊂ G(A) are subalgebras generated separately by {fi}, {hi}, {ei}. Actually,
H is a commutative Lie algebra with the basis {hi}. We introduce in H a (possibly, degenerate)
inner product by the formula 〈hi, hj〉 = diaij .
Fix an auxiliary n-dimensional complex vector space T with a basis α1, . . . , αn; Let Γ denote
a lattice generated by α1, . . . , αn, and let Γ+ be the intersection of Γ with the (closed) positive
octant. For an integral linear combination α =
∑n
i=1 miαi, denote by Gα the subspace of G(A)
spanned by monomials in ei, hi, fi such that for every i, the difference between the number of
occurrences of ei and fi equals mi. If α 6= 0 and Gα 6= 0, then α is called a root of G(A). Every
root is a positive, or a negative, integral linear combination of αi; accordingly the root is called
positive or negative (and we write α > 0 or α < 0). Obviously, N+ = ⊕α>0Gα, N− = ⊕α<0Gα.
Remark that Verma modules have a natural grading by the semigroup Γ+.
For α =
∑
kiαi, let hα =
∑
kid
−1
i hi. We can carry the inner product from H to T using
the formula 〈α, β〉 = 〈hα, hβ〉. If 〈α, α〉 6= 0, then we define a reflection sα : H∗ → H∗ by the
formula
(sαλ)(h) = λ(h)− 2λ(hα)
〈α, α〉
〈hα, h〉.
The similar formula
sαβ = β − 2〈α, β〉
〈α, α〉
α
defines a reflection sα : T → T . “Elementary reflections” si = sαi generate the action of the
Weyl group W (A) of G(A) in H∗ and in H. In H∗ we consider, besides the reflections sα the
reflections sρ
α, sρ
α(λ) = sα(λ+ρ)−ρ where ρ ∈ H∗ is defined by the formula ρ(hi) = 1, 1 ≤ i ≤ n.
Projections of Singular Vectors of Verma Modules 3
The Kac–Kazhdan criterion for reducibility of Verma modules M(λ), mentioned above, has
precise statement:
Theorem 1. M(λ) is reducible if and only if for some positive root α and some positive inte-
ger m,
(λ + ρ)(hα)− m
2
〈α, α〉 = 0. (1)
Moreover, if λ satisfies this equation for a unique pair α, m, then all non-trivial singular vectors
of M(λ) are contained in M(λ)mα.
For m and α satisfying this criterion, Feigin, Fuchs and Malikov [6] give a description for the
singular vector of degree mα in M(λ). In the case when α is a real root, that is, 〈α, α〉 6= 0, their
description is as follows. Let sα = siN · · · si1 be a presentation of sα ∈ W (A) as a product of
elementary reflections. For λ ∈ H∗, set λ0 = λ, λj = sij (λj−1 +ρ)−ρ for 0 < j ≤ N . Obviously,
the vector
−−−−→
λj−1λj is collinear to αij (or, rather, to 〈αij , 〉); let
−−−−→
λj−1λj = γj . Let α satisfies (for
some m) the equation (1). Then
F (sα;λ)u where F (sα;λ) = f−γN
iN
· · · f−γ1
i1
is a singular vector in M(λ)mα. Notice that the exponents in the last formula are, in general,
complex numbers. It is explained in [6] why the expression for F (sα;λ) still makes sense.
3 The case of rank two
In the case n = 2, a (symmetrizable) Cartan matrix given by
A =
(
2 −q
−p 2
)
,
where p > 0, q > 0. Since for pq ≤ 3, the algebra G(A) is finite-dimensional, we consider below
the case when pq ≥ 4.
Simple calculations show that sα1(α1) = −α1, sα1(α2) = pα1 + α2, sα2(α1) = α1 + qα2,
and sα2(α2) = −α2, and it is easy to check that the orbit of the root (1, 0) lies in the curve
qx2 − pqxy + py2 = q and the orbit of (0, 1) lies in qx2 − pqxy + py2 = p. (If pq < 4, these two
curves are hyperbolas sharing asymptotes, in the (degenerate) case of pq = 4, they are pairs of
parallel lines with a slope of q
2 .)
Define a sequence recursively by a0 = 0, a1 = 1, and an = san−1−an−2 where s2 = pq. Then
for σ2 = q
p we can calculate
(1, 0) = (a1, σa0),
s2((1, 0)) = (1, q) = (a1, σa2),
s1s2((1, 0)) = (pq − 1, q) = (a3, σa2),
s2s1s2((1, 0)) = (pq − 1, q(pq − 2)) = (a3, σa4).
More generally, the following is true.
Proposition 1. The orbit of (1, 0) consists of points
· · · (a2n−1, σa2n−2), (a2n−1, σa2n), (a2n+1, σa2n), (a2n+1, σa2n+2) · · ·
determined by the sequence {an} above; while the orbit of (0, 1) consists of points
· · · (σ−1a2n−2, a2n−1), (σ−1a2n, a2n−1), (σ−1a2n, a2n+1), (σ−1a2n+2, a2n+1) · · ·
for n ≥ 1.
4 D. Fuchs and C. Wilmarth
Proof. The proof is by induction on n. �
Obtaining explicit coordinates for the real orbits is straightforward in the affine case, because
of the simpler geometry. For pq > 4 an explicit description of the sequence {an} is possible using
an argument familiar to Fibonnaci enthusiasts:
Proposition 2. The nth term is
an =
1√
pq − 4
(√
pq +
√
pq − 4
2
)n
− 1√
pq − 4
(√
pq −
√
pq − 4
2
)n
.
Proof. Direct computation. �
With these real roots now labelled by the sequence {an}, we present the singular vectors
indexed by them in the Verma modules over G(A). Write λ = xλ1 + yλ2 where λi(hj) = δij , so
that λ(h1) = x and λ(h2) = y. Let us define the numbers Γk
1, Γk
2 by the formulas:
Γ2m
1 = q
m−1∑
i=0
(−1)i
(
2m− i− 1
2m− 2i− 1
)
(pq)m−i−1,
Γ2m
2 =
m−1∑
i=0
(−1)i
(
2(m− 1)− i
2(m− 1)− 2i
)
(pq)m−i−1,
Γ2m+1
1 =
m∑
i=0
(−1)i
(
2m− i
2m− 2i
)
(pq)m−i,
Γ2m+1
2 = p
m−1∑
i=0
(−1)i
(
2m− i− 1
2m− 2i− 1
)
(pq)m−i−1.
(Note that Γ0
1 = Γ0
2 = 0.)
The formula from [6] takes in our case the following form.
Theorem 2. For the algebra G(A) with Cartan matrix
A =
(
2 −p
−q 2
)
the singular vectors are as follows:
1. For the root α = (a2n−1, σa2n−2), with m ∈ N, and t ∈ C arbitrary,
F (sα;λ) = f
Γ4n−3
1 m
a2n−1
+Γ2n−1
2 t
1 f
Γ4n−4
1 m
a2n−1
+Γ2n−2
2 t
2 f
Γ4n−5
1 m
a2n−1
+Γ2n−3
2 t
1
· · · f
Γ2n
1 m
a2n−1
+Γ2
2t
2 fm
1 f
Γ2n−2
1 m
a2n−1
−Γ2
2t
2 · · · f
Γ2
1m
a2n−1
−Γ2n−2
2 t
2 f
Γ1
1m
a2n−1
−Γ2n−1
2 t
1
and the vector F (sα, λ)u is singular in M(λ) = M
(
m−Γ2n−1
1
Γ2n−1
1
− Γ2n−1
2 t, Γ2n−1
1 t− 1
)
.
2. For α = (a2n−1, σa2n),
F (sα;λ) = f
Γ4n
2 m
σ−1a2n
+Γ2n
2 t
2 f
Γ4n−1
2 m
σ−1a2n
+Γ2n−1
2 t
1 f
Γ4n−2
2 m
σ−1a2n
+Γ2n−2
2 t
2
· · · f
Γ2n+2
2 m
σ−1a2n
+Γ2
2t
2 fm
1 f
Γ2n
2 m
σ−1a2n
−Γ2
2t
2 · · · f
Γ3
2m
σ−1a2n
−Γ2n−1
2 t
1 f
Γ2
2m
σ−1a2n
−Γ2n
2 t
2
and the vector F (sα, λ)u is singular in M(λ) = M
(
Γ2n
2 t− 1,
m−Γ2n
2
Γ2n
2
− Γ2n
1 t
)
.
Projections of Singular Vectors of Verma Modules 5
3. For α = (σ−1a2n−2, a2n−1),
F (sα, λ) = f
Γ4n−2
2 m
a2n−1
+Γ2n−2
1 t
2 f
Γ4n−3
2 m
a2n−1
+Γ2n−3
1 t
1 f
Γ4n−4
2 m
a2n−1
+Γ2n−4
1 t
2
· · · f
Γ2n+1
2 m
a2n−1
+Γ1
1t
1 fm
2 f
Γ2n−1
2 m
a2n−1
−Γ1
1t
1 · · · f
Γ3
2m
a2n−1
−Γ2n−3
1 t
1 f
Γ2
2m
a2n−1
−Γ2n−2
1 t
2
and the vector F (sα, λ)u is singular in M(λ) = M
(
Γ2n−1
2 t− 1,
m−Γ2n−1
2
Γ2n−1
2
− Γ2n−1
1 t
)
.
4. For α = (σ−1a2n, a2n−1),
F (sα;λ) = f
Γ4n−1
1 m
σa2n
+Γ2n−1
1 t
1 f
Γ4n−2
1 m
σa2n
+Γ2n−2
1 t
2 f
Γ4n−3
1 m
σa2n
+Γ2n−3
1 t
1
· · · f
Γ2n+1
1 m
σa2n
+Γ1
1t
1 fm
2 f
Γ2n−1
1 m
σa2n
−Γ2n−3
1 t
1 · · · f
Γ2
1m
σa2n
−Γ2n−2
1 t
2 f
Γ1
1m
σa2n
−Γ2n−1
1 t
1
and the vector F (sα, λ)u is singular in M(λ) =
(
m−Γ2n
1
Γ2n
1
− Γ2n
2 t, Γ2n
1 t− 1
)
.
Proof. It must be checked that the vectors given above actually correspond to the Feigin–
Fuchs–Malikov (FFM) procedure for obtaining singular vectors, and also that the Kac–Kazhdan
criterion for reducibility is satisfied. For λ = xλ1 + yλ2 and the reflection sα = siN · · · si1 (pro-
duct of simple reflections) the algorithm requires successive application of the transformations
sρ
1 := s1(λ + ρ)− ρ and sρ
2 := s2(λ + ρ)− ρ. One generates the list
λ0 = xλ1 + yλ2, λj = sij (λ
j−1 + ρ)− ρ
and the auxiliary sequence {
−−−−→
λj−1λj}j≥1. The algorithm then gives
F (sα;λ) = fθN
iN
· · · fθ1
i1
,
where
−−−−→
λj−1λj = −θjαij (here αij is the functional 〈hαij
, ·〉).
So we first need to know the decomposition of sα into elementary reflections for α in the
orbit of (1, 0) or (0, 1). Let Si(m) denote the word in H∗ beginning and ending with si, and
containing m si’s. For example, S1(3) = s1s2s1s2s1.
Lemma 1. For real α as above, sα is the word
(a2n−1, σa2n−2)←→ S1(2n− 1),
(a2n−1, σa2n)←→ S2(2n),
(σ−1a2n, a2n−1)←→ S1(2n),
(σ−1a2n−2, a2n−1)←→ S2(2n− 1).
Proof. This is an easy induction on n. �
The coefficients of collinearity θj have the following description.
Lemma 2. For λk = Λk
1λ1 + Λk
2λ2 we have
(i)
−−−−−−−−→
λ2n+1λ2n+2 = −(Λ2n+1
2 + 1)〈hα2 , ·〉,
(ii)
−−−−−−→
λ2nλ2n+1 = −(Λ2n
1 + 1)〈hα1 , ·〉.
6 D. Fuchs and C. Wilmarth
Proof. One easily computes that 〈hα1 , ·〉 = 2λ1 + qλ2, 〈hα2 , ·〉 = −pλ1 + 2λ2, and further that
sρ
1(xλ1 + yλ2) = (−x− 2)λ1 + (y + q(x + 1))λ2,
sρ
2(xλ1 + yλ2) = (x + p(y + 1))λ1 + (−y − 2)λ2.
Then for (i) it is verified that
−−→
λ1λ2 = −(y + qx + q + 1)(−pλ1 + 2λ2) = −(Λ1
2 + 1)〈hα2 , ·〉. For
n > 0,
λ2n+2 = sρ
2(s
ρ
1(λ
2n)) = sρ
2((−Λ2n
1 − 2)λ1 + (Λ2n
2 + qΛ2n
1 + q)λ2)
= (−Λ2n
1 − 2 + pΛ2n
2 + pqΛ2n
1 + pq + p)λ1 + (−Λ2n
2 − qΛ2n
1 − q − 2)λ2
while
λ2n+1 = sρ
1(λ
2n) = (−Λ2n
1 − 2)λ1 + (Λ2n
2 + qΛ2n
1 + q)λ2.
So
−−−−−−−−→
λ2n+1λ2n+2 = −(Λ2n
2 + qΛ2n
1 + q + 1)(−pλ1 + 2λ2). Since λ2n+1 = sρ
1(λ
2n) = (−Λ2n
1 − 2)λ1 +
(Λ2n
2 + qΛ2n
1 + q)λ2, we have −Λ2n+2
2 − 1 = −(Λ2n
2 + qΛ2n
1 + q + 1) as desired. The argument
for (ii) is similar. �
Let us put Γk = Γk
1x + Γk
2y. We will also need
Lemma 3.
(i) Γ2n+1 = pΓ2n − Γ2n−1,
(ii) Γ2n+2 = qΓ2n+1 − Γ2n.
Proof. These can be verified directly. �
We are now in a position to show by induction that the FFM-exponents correspond to the Γk
in the statement of Theorem 2. It suffices by the second lemma to show that
Γ2n+1 = Λ2n
1 + 1 and Γ2n+2 = Λ2n+1
2 + 1.
Making a change of variable x + 1→ x and y + 1→ y one can calculate that
−−→
λ0λ1 = −x〈hα1 , ·〉,
−−→
λ1λ2 = −(qx + y)〈hα2 , ·〉,
−−→
λ2λ3 = −((pq − 1)x + py)〈hα1 , ·〉,
. . . . . . . . . . . . . . . . . . . . . . . . . . .
The FFM exponents are just the coefficients of these functionals with reversed sign. For the
base case of our induction, n = 0, observe that λ0 = (x−1)λ1 +(y−1)λ2, hence Λ0
1 +1 = x = Γ1
(using the binomial definition of Γ1), while Λ1
2 + 1 = y + qx since sρ
1(λ
0) = (−x − 1)λ1 + (y −
1 + qx)λ2, agreeing with the binomial sum Γ2 = y + qx.
Inductively assume that for some (n− 1) > 0
Γ2(n−1)+1 = Λ2n−2
1 + 1 and Γ2(n−1)+2 = Λ2n−1
2 + 1.
We need to show that
Γ2n+1 = Λ2n
1 + 1 and Γ2n+2 = Λ2n+1
2 + 1.
Projections of Singular Vectors of Verma Modules 7
Just observe that
Λ2n
1 + 1 = Λ2n−1
1 + p(Λ2n−1
2 + 1) + 1 = (Λ2n−1
1 + 1) + p(Λ2n−1
2 + 1)
= (−Λ2n−2
1 − 2 + 1) + p
[
Λ2n−2
2 + q(Λ2n−2
1 + 1) + 1
]
= −(Λ2n−2
1 + 1) + p
[
(Λ2n−2
2 + 1) + q(Λ2n−2
1 + 1)
]
= −Γ2n−1 + p(Λ2n−1
2 + 1) = −Γ2n−1 + pΓ2n = Γ2n+1,
where the last equality comes from Lemma 3 and the inductive hypothesis is used in the preceding
two lines. The same tack proves that Γ2n+2 = Λ2n+1
2 + 1.
We now know that the FFM-exponents are as given Theorem 2. It only remains to check
that the Kac–Kazhdan criterion (Theorem 1) is satisfied. For m and α satisfying this criterion,
[6] give the prescription for the singular vector F (sα;λ)u of degree mα in M(λ); so we need to
verify the existence of such α and m.
For α = aα1 + bα2 and λ = xλ1 + yλ2, hα = ad−1
1 h1 + bd−1
2 = a
ph1 + b
qh2, so the criterion can
be restated as 2(xλ1 + yλ2 + ρ)
(
a
ph1 + b
qh2
)
= m〈aα1 + bα2, aα1 + bα2〉. After the calculations
this is
2
(
x
a
p
+ y
b
q
+
a
p
+
b
q
)
= m
(
2a2
p
− 2ab +
2b2
q
)
or
(x + 1)
a
p
+ (y + 1)
b
q
= m
(
a2
p
− ab +
b2
q
)
.
So after change of variable x + 1→ x and y + 1→ y the Kac–Kazhdan criterion becomes
x
a
p
+ y
b
q
= m
(
a2
p
− ab +
b2
q
)
.
We show that the integral exponent of the centermost element in the singular vectors in the
statement of the theorem precisely meets the integrality requirement of the criterion. This is
a case by case check, and somewhat tedious and technical; let us verify it for roots of type
(a2n+1, σa2n), whose singular vector comprises 2n + 1 f1’s and 2n f2’s raised to appropriate
powers; the centermost exponent is then the 2n + 1-st coefficient of collinearity in the FFM
procedure, or what we have called Γ2n+1.
A remark, a lemma, and a corollary will show that Γ2n+1 does what it is supposed to.
Remark 1. qΓ2n+1
2 = pΓ2n
1 , as is transparent from the definitions of Γ2n and Γ2n+1.
The next lemma will relate the root sequence {an} to the exponents of the singular vectors.
Lemma 4. The following is true for n ≥ 0:
Γ2n+1
1 = a2n+1, Γ2n+2
1 = σa2n+2.
Proof. Induction on n. �
Corollary 1. Γ2n+1 = a2n+1x + p
q σa2ny.
Proof. Γ2n+1 = Γ2n+1
1 x+Γ2n+1
2 y = a2n+1x+Γ2n+1
2 y = a2n+1x+ p
q Γ2n
1 y = a2n+1x+ p
q ua2ny. �
8 D. Fuchs and C. Wilmarth
Finally the Kac–Kazhdan criterion for (a2n+1, ua2n) is
a2n+1
x
p
+ ua2n
y
q
= m
(
(a2n+1)2
p
− ua2na2n+1 +
(ua2n)2
q
)
or equivalently
a2n+1x +
p
q
ua2ny = m((a2n+1)2 − pua2na2n+1 +
p
q
(ua2n)2)
= m((a2n+1)2 − pua2na2n+1 +
p
q
q
p
(a2n)2) = m′ ∈ N
since a2n+1 and ua2n are integral (polynomial in p and q). But the left-hand side here is by
the corollary exactly the exponent of the centermost letter in the singular vector, which by the
formula given in the theorem is an integer; so the Kac–Kazhdan criterion is indeed satisfied in
this case. One can check in similar fashion that the remaining three cases also fit the integrality
requirement.
The singular vectors in the statement of the theorem appear as follows. The exponent of the
centermost vector in all four cases must be integral: setting this expression in x and y equal
to m one then solves for x (or y) in terms of m and y (respectively, x); t is then introduced as
a scalar multiple of y (resp., x) to minimize notational clutter. This completes the proof of the
theorem. �
4 Projections
We next obtain projections of the singular vectors into the Heisenberg algebra, where they
factor as products. While the theorem gives a simple and perhaps the most natural expression
for the singular vectors in terms of the Γk a change of variable is advantageous in the projection
and factoring of these vectors in the Heisenberg algebra. In each case this involves setting the
exponent of the vector immediately to the left of the centermost letter equal to a complex variab-
le α (which will then depend on n). For example the root γ = (u−1a4, a3) = (p(pq − 2), pq − 1)
has from the theorem the corresponding singular vector:
f
((pq)3−5(pq)2+6pq−1)m
q(pq−2)
+(pq−1)t
1 f
q((pq)2−4pq+3)m
q(pq−2)
+qt
2 f
(pq)2−3pq+1)m
q(pq−2)
+t
1 fm
2
× f
(pq−1)m
q(pq−2)
−t
1 f
qm
q(pq−2)
−qt
2 f
m
q(pq−2)
−(pq−1)t
1 .
Taking
α =
((pq)2 − 3pq + 1)m
q(pq − 2)
+ t
the singular vector becomes
f
(pq−1)α−pm
1 f qα−m
2 fα
1 fm
2 fpm−α
1 f
q(pm−α)−m
2 f
(pq−1)(pm−α)−pm
1
which can in turn be rewritten in terms of the Γk as:
f
Γ4
2α−Γ3
2m
1 f
Γ2
1α−Γ1
1m
2 f
Γ2
2α−Γ1
2m
1 fm
2 f
−Γ2
2α+Γ3
2m
1 f
−Γ2
1α+Γ3
1m
2 f
−Γ4
2α+Γ5
2m
1 .
Proposition 3. Under this change of variable the singular vectors take the following form:
1. For α = (a2n−1, σa2n−2) the corresponding F (sα;λ) is
f
Γ2n−1
2 α−Γ2n−2
2 m
1 f
Γ2n−3
1 α−Γ2n−4
1 m
2 f
Γ2n−3
2 α−Γ2n−4
2 m
1 f
Γ2n−5
1 α−Γ2n−6
1 m
2 f
Γ2n−5
2 α−Γ2n−6
2 m
1
Projections of Singular Vectors of Verma Modules 9
· · · fΓ3
1α−Γ2
1m
2 f
Γ3
2α−Γ2
2m
1 f
Γ1
1α−Γ0
1m
2 fm
1 f
−Γ1
1α+Γ2
1m
2 f
−Γ3
2α+Γ4
2m
1 f
−Γ3
1α+Γ4
1m
2
· · · f−Γ2n−3
2 α+Γ2n−2
2 m
1 f
−Γ2n−3
1 α+Γ2n−2
1 m
2 f
−Γ2n−1
2 α+Γ2n
2 m
1 .
2. For γ = (a2n−1, ua2n) the corresponding singular vector F (sγ ;λ) is
f
Γ2n−1
1 α−Γ2n−2
1 m
2
(
F (s(a2n−1,ua2n−2);λ
)
f
−Γ2n−1
1 α+Γ2n
1 m
2 .
3. For γ = (u−1a2n−2, a2n−1) the singular vector F (sγ ;λ) becomes
f
Γ2n−2
1 α−Γ2n−3
1 m
2 f
Γ2n−2
2 α−Γ2n−3
1 m
1 f
Γ2n−4
1 α−Γ2n−5
1 m
2 f
Γ2n−4
2 α−Γ2n−5
2 m
1 · · ·
f
Γ2
1α−Γ1
1m
2 f
Γ2
2α−Γ1
2m
1 fm
2 f
−Γ2
2α+Γ3
2m
1 f
−Γ2
1α+Γ3
1m
2 · · · f−Γ2n−2
2 α+Γ2n−1
2 m
1 f
−Γ2n−2
1 α+Γ2n−1
1 m
2 .
4. For γ = (u−1a2n, a2n−1) the singular vector is
f
Γ2n
2 α−Γ2n−1
2 m
1
(
F (s(u−1a2n−2,a2n−1)
;λ)
)
f
−Γ2n
2 α+Γ2n+1
2 m
1 .
Proof. This can be established by induction on the number of pairs transformed. �
We now project the singular vectors into the universal enveloping algebra of the three-
dimensional Heisenberg algebra H. Recall that this is generated by f1, f2, with [f1, f2] =: h,
[f1, h] = [f2, h] = 0. Thus, the projection U(N−)→ U(H) is the factorization over the (two-sided)
ideal generated by [f1, h], [f2, h].
Let Hu := f2f1+uh for u ∈ C. Observe that HuHv = HvHu, u, v ∈ C. The following relations
also hold in the Heisenberg (for positive integers, and hence for arbitrary complex numbers α,
β, u).
Lemma 5. For α, β, u ∈ C,
1) fα
2 Hu = Hu−αfα
2 ;
2) fβ
1 Hu = Hu+βfβ
1 ;
3) fα
1 fn
2 fn−α
1 = HαHα−1 · · ·Hα−(n−1);
4) fα
2 fn
1 fn−α
2 = H1−αH2−α · · ·Hn−α.
Proof. The calculations follow readily from the complex binomial formula given in [6]: for
g1, g2 ∈ G a Lie algebra, γ1, γ2 ∈ C, we have
gγ1
1 gγ2
2 = gγ2
2 gγ1
1 +
∞∑
j1=1
∞∑
j2=1
(
γ1
j1
)(
γ2
j2
)
Qj1j2(g1, g2)g
γ2−j2
2 gγ1−j1
1 ,
where
(
γ
j
)
= γ(j)
j! with γ(j) = γ(γ − 1) · · · (γ − j + 1) and the Lie polynomials Qj1j2 can be
calculated explicitly, for example using the recursion
Qj1j2(g1, g2) = [g1, Qj1−1,j2(g1, g2)] +
j2−1∑
v=0
(
j2
v
)
Qj1−1,v(g1, g2)[g2, . . . , [g2, [g2︸ ︷︷ ︸
j2−v
, g1]] . . . ]
with Q00 = 1 and Q0,v = 0 for v > 0. �
10 D. Fuchs and C. Wilmarth
Now set, for r, s ∈ N
Hr,s
0 = 1 (the empty product),
Hr,s
1 =
(Γ2
2−Γ0
1)m∏
k=1
H(−Γ2
2+Γ3
2−···±Γr
2)α+(−Γ1
1+Γ2
1−···±Γs
1)m +k.
The for j ≥ 1 define
Hr,s
2j =
(Γ2j
1 −Γ2j
2 )m∏
k=1
H
(Γ2j+1
2 −Γ2j+2
2 +···±Γr
2)α+(−Γ2j−1
1 +Γ2j
1 −···±Γs
1)m− (k−1)
,
Hr,s
2j+1 =
(Γ2j+2
2 −Γ2j
1 )m∏
k=1
H
(−Γ2j+2
2 +Γ2j+3
2 −···±Γr
2)α+(Γ2j
1 −Γ2j+1
1 +···±Γs
1)m +k
.
We will also need the following, not dissimilar, but warranting its own notation:
H̃r,s
0 = 1,
H̃r,s
1 =
(Γ2
2−Γ1
2)m∏
k=1
H(Γ1
1−Γ2
1+···±Γr
1)α+(Γ2
2−Γ3
2+···±Γs
2)m−(k−1).
For j ≥ 1 define
H̃r,s
2j =
(Γ2j+1
2 −Γ2j
2 )m∏
k=1
H
(−Γ2j
1 +Γ2j+1
1 −···±Γr
1)α+(Γ2j
2 −Γ2j+1
2 +···±Γs
2)m +k
,
H̃r,s
2j+1 =
(Γ2j+2
2 −Γ2j+1
2 )m∏
k=1
H
(Γ2j+1
1 −Γ2j+2
1 +···±Γr
1)α+(−Γ2j+1
2 +Γ2j+2
2 −···±Γs
2)m−(k−1)
.
Theorem 3. The singular vectors whose words F (sα;λ) were given in the preceding theorem
project to the Heisenberg algebra as the following products:
1. The singular vector corresponding to (a2n−1, ua2n−2) projects to
2n−2∏
w=1
H2n−1,2n−3
w f
(Γ2n−1
1 −Γ2n−2
1 )m
1 .
2. The singular vector corresponding to (a2n−1, ua2n) projects to
2n−1∏
w=1
H2n,2n−2
w f
(Γ2n
1 −Γ2n−1
1 )m
2 .
3. The singular vector for (u−1a2n−2, a2n−1) projects to
2n−2∏
w=1
H̃2n−2,2n−2
w f
(Γ2n
2 −Γ2n−1
2 )m
2 .
4. The singular vector for (u−1a2n, a2n−1) projects to
2n−1∏
w=1
H̃2n−1,2n−1
w f
(Γ2n+1
2 −Γ2n
2 )m
1 .
Proof. Induction on n. This is completely straightforward using the change of variables for the
singular vectors given in the proposition above. �
Projections of Singular Vectors of Verma Modules 11
5 Other projections
One might ask whether projections into other algebras, for instance into U(sl2), are equally
possible. A homomorphism from the universal enveloping algebra of N− to U(sl2) is defined
when p > 1 and q > 1. It is the factorization over the (two-sided) ideal generated by [h, f1]− f1
and [h, f2] + f2 (where, as before, h = [f1, f2]). Set, for u ∈ C,
Ju = f2f1 + uh− u(u− 1)
2
.
It can then be checked that for β ∈ C we have the following in U(sl2) (putting e = f1, f = f2) :
fβ
1 Ju = Ju+βfβ
1 , fβ
2 Ju = Ju−βfβ
2
as well as, for n ∈ N,
fβ
1 fn
2 fn−β
1 = JβJβ−1 · · · Jβ−(n−1),
fβ
2 fn
1 fn−β
2 = J1−βJ2−β · · · Jn−β,
fn
1 fn
2 = J1 · · · Jn.
These properties permit the formal manipulations that afford the factorization results we have
already detailed. Simply substitute J for H in the statement of the projection theorem.
6 Concluding remarks
In [3] it is observed that information about singular vectors of Verma modules can be used to
obtain information about the homologies of nilpotent Lie algebras. Namely, the differentials
of the Bernstein–Gel’fand–Gel’fand resolution BGG of C over N− (see [1]) are presented by
matrices whose entries are singular vectors in the Verma modules. Thus, if V is an N−-module
that is trivial over the kernel of the projection of U(N−) onto U(H) or U(sl2) considered above,
then our formulas give an explicit description of BGG⊗ V and Hom(BGG, V ).
References
[1] Bernstein I.N., Gel’fand I.M., Gel’fand S.I., Structure of representations generated by highest weight vectors,
Funktsional. Anal. i Prilozhen. 5 (1971), no. 1, 1–9 (English transl.: Funct. Anal. Appl. 5 (1971), no. 1,
1–8).
[2] Feigin B.L., Fuchs D.B., Verma modules over the Virasoro algebra, in Topology (Leningrad, 1982), Lecture
Notes in Math., Vol. 1060, Springer, Berlin, 1984, 230–245.
[3] Fuchs D., Two projections of singular vectors of Verma modules over the affine Lie algebra A1
1, Funktsional.
Anal. i Prilozhen. 23 (1989), no. 2, 81–83 (English transl.: Funct. Anal. Appl. 23 (1989), no. 2, 154–156).
[4] Kac V., Infinite-dimensional Lie algebras, 3rd ed., Cambridge University Press, Cambridge, 1990.
[5] Kac V., Kazhdan D., Structure of representations with highest weight of infinite-dimensional Lie algebras,
Adv. in Math. 34 (1979), 97–108.
[6] Malikov F.G., Feigin B.L., Fuchs D.B., Singular vectors in Verma modules over Kac–Moody algebras,
Funktsional. Anal. i Prilozhen. 20 (1986), no. 2, 25–37 (English transl.: Funct. Anal. Appl. 20 (1986), no. 2,
103–113).
1 Introduction
2 Preliminaries
3 The case of rank two
4 Projections
5 Other projections
6 Concluding remarks
References
|
| id | nasplib_isofts_kiev_ua-123456789-149013 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 1815-0659 |
| language | English |
| last_indexed | 2025-12-07T18:25:26Z |
| publishDate | 2008 |
| publisher | Інститут математики НАН України |
| record_format | dspace |
| spelling | Fuchs, D. Wilmarth, C. 2019-02-19T12:57:31Z 2019-02-19T12:57:31Z 2008 Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras / D. Fuchs, C. Wilmarth // Symmetry, Integrability and Geometry: Methods and Applications. — 2008. — Т. 4. — Бібліогр.: 6 назв. — англ. 1815-0659 2000 Mathematics Subject Classification: 17B67 https://nasplib.isofts.kiev.ua/handle/123456789/149013 We prove an explicit formula for a projection of singular vectors in the Verma module over a rank 2 Kac-Moody Lie algebra onto the universal enveloping algebra of the Heisenberg Lie algebra and of sl2 (Theorem 3). The formula is derived from a more general but less explicit formula due to Feigin, Fuchs and Malikov [Funct. Anal. Appl. 20 (1986), no. 2, 103–113]. In the simpler case of A11 the formula was obtained in [Fuchs D., Funct. Anal. Appl. 23 (1989), no. 2, 154–156]. This paper is a contribution to the Special Issue on Kac–Moody Algebras and Applications. en Інститут математики НАН України Symmetry, Integrability and Geometry: Methods and Applications Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras Article published earlier |
| spellingShingle | Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras Fuchs, D. Wilmarth, C. |
| title | Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras |
| title_full | Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras |
| title_fullStr | Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras |
| title_full_unstemmed | Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras |
| title_short | Projections of Singular Vectors of Verma Modules over Rank 2 Kac-Moody Lie Algebras |
| title_sort | projections of singular vectors of verma modules over rank 2 kac-moody lie algebras |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/149013 |
| work_keys_str_mv | AT fuchsd projectionsofsingularvectorsofvermamodulesoverrank2kacmoodyliealgebras AT wilmarthc projectionsofsingularvectorsofvermamodulesoverrank2kacmoodyliealgebras |