KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians
We discuss a relation between the characteristic variety of the KZ equations and the zero set of the classical Calogero-Moser Hamiltonians.
Збережено в:
| Дата: | 2012 |
|---|---|
| Автори: | , , |
| Формат: | Стаття |
| Мова: | English |
| Опубліковано: |
Інститут математики НАН України
2012
|
| Назва видання: | Symmetry, Integrability and Geometry: Methods and Applications |
| Онлайн доступ: | http://dspace.nbuv.gov.ua/handle/123456789/148698 |
| 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: | KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians / E. Mukhin, V. Tarasov, A. Varchenko // Symmetry, Integrability and Geometry: Methods and Applications. — 2012. — Т. 8. — Бібліогр.: 23 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
irk-123456789-148698 |
|---|---|
| record_format |
dspace |
| spelling |
irk-123456789-1486982019-02-19T01:23:44Z KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians Mukhin, E. Tarasov, T. Varchenko, A. We discuss a relation between the characteristic variety of the KZ equations and the zero set of the classical Calogero-Moser Hamiltonians. 2012 Article KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians / E. Mukhin, V. Tarasov, A. Varchenko // Symmetry, Integrability and Geometry: Methods and Applications. — 2012. — Т. 8. — Бібліогр.: 23 назв. — англ. 1815-0659 2010 Mathematics Subject Classification: 82B23; 17B80 DOI: http://dx.doi.org/10.3842/SIGMA.2012.072 http://dspace.nbuv.gov.ua/handle/123456789/148698 en Symmetry, Integrability and Geometry: Methods and Applications Інститут математики НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| description |
We discuss a relation between the characteristic variety of the KZ equations and the zero set of the classical Calogero-Moser Hamiltonians. |
| format |
Article |
| author |
Mukhin, E. Tarasov, T. Varchenko, A. |
| spellingShingle |
Mukhin, E. Tarasov, T. Varchenko, A. KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians Symmetry, Integrability and Geometry: Methods and Applications |
| author_facet |
Mukhin, E. Tarasov, T. Varchenko, A. |
| author_sort |
Mukhin, E. |
| title |
KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians |
| title_short |
KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians |
| title_full |
KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians |
| title_fullStr |
KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians |
| title_full_unstemmed |
KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians |
| title_sort |
kz characteristic variety as the zero set of classical calogero-moser hamiltonians |
| publisher |
Інститут математики НАН України |
| publishDate |
2012 |
| url |
http://dspace.nbuv.gov.ua/handle/123456789/148698 |
| citation_txt |
KZ Characteristic Variety as the Zero Set of Classical Calogero-Moser Hamiltonians / E. Mukhin, V. Tarasov, A. Varchenko // Symmetry, Integrability and Geometry: Methods and Applications. — 2012. — Т. 8. — Бібліогр.: 23 назв. — англ. |
| series |
Symmetry, Integrability and Geometry: Methods and Applications |
| work_keys_str_mv |
AT mukhine kzcharacteristicvarietyasthezerosetofclassicalcalogeromoserhamiltonians AT tarasovt kzcharacteristicvarietyasthezerosetofclassicalcalogeromoserhamiltonians AT varchenkoa kzcharacteristicvarietyasthezerosetofclassicalcalogeromoserhamiltonians |
| first_indexed |
2025-07-12T20:00:34Z |
| last_indexed |
2025-07-12T20:00:34Z |
| _version_ |
1837472615672840192 |
| fulltext |
Symmetry, Integrability and Geometry: Methods and Applications SIGMA 8 (2012), 072, 11 pages
KZ Characteristic Variety as the Zero Set
of Classical Calogero–Moser Hamiltonians
Evgeny MUKHIN †, Vitaly TARASOV †‡ and Alexander VARCHENKO §
† Department of Mathematical Sciences, Indiana University – Purdue University Indianapolis,
402 North Blackford St, Indianapolis, IN 46202-3216, USA
E-mail: mukhin@math.iupui.edu, vtarasov@math.iupui.edu
‡ St. Petersburg Branch of Steklov Mathematical Institute,
Fontanka 27, St. Petersburg, 191023, Russia
E-mail: vt@pdmi.ras.ru
§ Department of Mathematics, University of North Carolina at Chapel Hill,
Chapel Hill, NC 27599-3250, USA
E-mail: anv@email.unc.edu
Received June 09, 2012, in final form October 04, 2012; Published online October 16, 2012
http://dx.doi.org/10.3842/SIGMA.2012.072
Abstract. We discuss a relation between the characteristic variety of the KZ equations and
the zero set of the classical Calogero–Moser Hamiltonians.
Key words: Gaudin Hamiltonians; Calogero–Moser system; Wronski map
2010 Mathematics Subject Classification: 82B23; 17B80
1 Statement of results
1.1 Motivation
This paper is motivated by the Givental and Kim observation [10] that the characteristic variety
of the quantum differential equation of a flag variety is a Lagrangian variety of the classical Toda
lattice. The quantum differential equation is a system of differential equations ~∂iψ = bi ◦ ψ,
i = 1, . . . , r, defined by the quantum multiplication ◦ and depending on a parameter ~. The
system defines a flat connection for all nonzero values of ~. Givental and Kim, in particular,
observe that the characteristic variety of this system is the Lagrangian variety of the classical
Toda lattice, defined by equating to zero the first integrals of the Toda lattice.
In this paper we describe a similar relation between the KZ equation and the classical
Calogero–Moser system. On numerous relations between the KZ equations and quantum Calo-
gero–Moser systems see [2, 3, 6, 7, 14].
1.2 Classical Calogero–Moser system
Fix an integer n > 2. Denote ∆ = {z = (z1, . . . , zn) ∈ Cn | za = zb for some a 6= b}, the union of
diagonals. Consider the cotangent bundle T ∗(Cn −∆) with symplectic form ω =
n∑
a=1
dpa ∧ dza,
where p1, . . . , pn are coordinates on fibers. The classical Calogero–Moser system on T ∗(Cn−∆)
is defined by the Hamiltonian
H =
n∑
a=1
p2a −
∑
16a<b6n
2
(za − zb)2
.
mailto:mukhin@math.iupui.edu
mailto:vtarasov@math.iupui.edu
mailto:vt@pdmi.ras.ru
mailto:anv@email.unc.edu
http://dx.doi.org/10.3842/SIGMA.2012.072
2 E. Mukhin, V. Tarasov, and A. Varchenko
The system is completely integrable. For
Q =
p1
1
z1 − z2
1
z1 − z3
. . .
1
z1 − zn
1
z2 − z1
p2
1
z2 − z3
. . .
1
z2 − zn
. . . . . . . . . . . . . . .
1
zn − z1
1
zn − z2
1
zn − z3
. . . pn
, (1.1)
let det(u−Q) = un −Q1u
n−1 + · · · ±Qn be the characteristic polynomial. Then Q1, . . . , Qn is
a complete list of commuting first integrals, and H = Q2
1 −Q2.
We will be interested in the subvariety L0 ⊂ T ∗(Cn −∆) defined by the equations
L0 =
{
(z,p) ∈ T ∗(Cn −∆) | Qa(z,p) = 0, a = 1, . . . , n
}
. (1.2)
Theorem 1.1 ([23]). For any n, the subvariety L0 is smooth and Lagrangian.
See propositions in Section 6 of [23]. Another proof of Theorem 1.1 will be given in Section 2.4.
1.3 Gaudin Hamiltonians and KZ characteristic variety
Fix an integer N > 2. Denote V = CN the vector representation of the Lie algebra glN .
The Hamiltonians of the quantum Gaudin model are the linear operators H1, . . . ,Hn on the
space V ⊗n,
Ha(z) =
N∑
i,j=1
∑
b 6=a
e
(a)
ij e
(b)
ji
za − zb
, (1.3)
where eij are the standard generators of glN , e
(a)
ij is the image of 1⊗(a−1) ⊗ eij ⊗ 1⊗(n−a), and
z1, . . . , zn are distinct complex numbers, see [9]. The operators commute, [Ha(z), Hb(z)] = 0 for
all a, b. The operators commute with the glN -action on V ⊗n.
Let λ = (λ1, . . . , λN ) ∈ ZN>0 be a partition of n with at most N parts, λ1 > · · · > λN ,
|λ| = λ1 + · · ·+ λN = n. Denote
Sing V ⊗n[λ] =
{
v ∈ V ⊗n | eiiv = λiv, i = 1, . . . , N ; eijv = 0 for all i < j
}
,
the subspace of singular vectors of weight λ. The Gaudin Hamiltonians preserve Sing V ⊗n[λ].
We define the spectral variety of the Gaudin model on Sing V ⊗n[λ],
SpecN,λ =
{
(z,p) ∈ T ∗(Cn −∆) | ∃ v ∈ Sing V ⊗n[λ] with Ha(z)v = pav, a = 1, . . . , n
}
.
The spectral variety is a Lagrangian subvariety of T ∗(Cn−∆), see, for example, Proposition 1.5
in [20].
The Gaudin Hamiltonians are the right hand sides of the KZ equations,
κ∂ziI(z) = Hi(z)I(z), i = 1, . . . , n,
where κ ∈ C× is a parameter. The spectral variety SpecN,λ is, by definition, the characteristic
variety of the κ-dependent D-module defined by the KZ equations with values Sing V ⊗n[λ].
Example. If λ = (n, 0, . . . , 0), then SpecN,λ is given by the equations pa =
∑
b 6=a
(za−zb)−1, a = 1,
. . . , n. If N = n and λ = (1, . . . , 1), then SpecN,λ is given by the equations pa = −
∑
b 6=a
(za−zb)−1,
a = 1, . . . , n.
KZ Characteristic Variety as the Zero Set of Classical Calogero–Moser Hamiltonians 3
Theorem 1.2.
(i) The variety SpecN,λ does not depend on N . Namely, consider λ as a partition of n with
at most N + 1 parts. Then SpecN,λ = SpecN+1,λ. From now on we denote SpecN,λ by
Specλ.
(ii) For any n, the variety L0 is the disjoint union of the varieties Specλ, where the union is
over all partitions λ of n.
By Theorem 1.1, each Specλ is smooth and Lagrangian.
Part (i) of Theorem 1.2 is proved in Section 2.2, and part (ii) is proved in Section 2.4.
1.4 Master function generates Specλ
Let λ be a partition of n with at most N parts. Denote la =
N∑
b=a+1
λb, a = 1, . . . , N − 1. Denote
l = l1 + · · ·+ lN−1. Consider the set of l variables
t =
(
t
(1)
1 , . . . , t
(1)
l1
, . . . , t
(N−1)
1 , . . . , t
(N−1)
lN−1
)
and the affine space Cn × Cl with coordinates z, t. The function ΦN,λ : Cn × Cl → C,
ΦN,λ(z, t) =
∑
16a<b6n
log(za − zb)−
n∑
a=1
la∑
i=1
log
(
t
(1)
i − za
)
+ 2
N−1∑
k=1
∑
16i<j6lk
log
(
t
(k)
i − t
(k)
j
)
−
N−2∑
k=0
lk∑
i=1
lk+1∑
j=1
log
(
t
(k)
i − t
(k+1)
j
)
is called the master function, see [21, 22].
The master function depends on λ, but not on N . Namely, consider λ as a partition of n
with at most N + 1 parts. Then ΦN,λ = ΦN+1,λ. From now on we denote ΦN,λ by Φλ.
Critical points of Φλ with respect to t are given by the equation dtΦλ = 0. Denote by Critλ
the critical set of Φλ with respect to t,
Critλ =
{
(z, t) ∈ Cn × Cl | dtΦλ(z, t) = 0
}
.
This is an algebraic subset of the domain of Cn × Cl, where the master function is a regular
(multivalued) function. Denote by Lλ ⊂ T ∗(Cn −∆) the image of the map
Critλ → T ∗(Cn −∆), (z, t) 7→ (z,p), where pa =
∂Φλ
∂za
(z, t), a = 1, . . . , n.
Theorem 1.3. For any n and a partition λ of n, we have Lλ ⊂ Specλ and Specλ is the closure
of Lλ in T ∗(Cn −∆).
Theorem 1.3 is proved in Section 2.1.
1.5 Calogero–Moser space Cn and cotangent bundle T ∗(Cn − ∆)
The Calogero–Moser system has singularities if some of z1, . . . , zn coincide. These singularities
can be resolved and the Calogero–Moser system can be lifted by the map ξ given by (1.4) below
to a regular completely integrable Hamiltonian system on the Calogero–Moser space Cn, see [13].
Denote
C̃n =
{
(Z,Q) ∈ gln × gln | rank([Z,Q] + 1) = 1
}
.
4 E. Mukhin, V. Tarasov, and A. Varchenko
The group GLn of complex invertible matrices acts on C̃n by simultaneous conjugation. The
action is free and proper, see [23]. The quotient space Cn is called the n-th Calogero–Moser
space. The Calogero–Moser space Cn is a smooth affine variety of dimension 2n, see [23].
The group Sn freely acts on Cn − ∆ by permuting coordinates. The action lifts to a free
action on T ∗(Cn −∆). Define the map
ξ : T ∗
(
Cn −∆
)
→ T ∗
(
Cn −∆
)
/Sn → Cn (1.4)
by the rule: (z,p) is mapped to (Z,Q), where Z = diag(z1, . . . , zn) and Q is defined by (1.1).
The map ξ induces an embedding T ∗(Cn −∆)/Sn → Cn whose image is Zariski open in Cn.
Set C(n) = Cn/Sn and let spec(X) ∈ C(n) stand for the point given by the eigenvalues of
a square matrix X. The canonical map
π : Cn → C(n) × C(n), (Z,Q) 7→ (spec(Z), spec(Q)),
is a finite map of degree n!, see [4]. This map and its fiber over 0× 0 were studied, for example,
in [4, 8, 11].
Let C0
n be the subvariety π−1(C(n) × 0) ⊂ Cn. Identifying C(n) × 0 with C(n) we get a map
π0 : C0
n → C(n), (Z,Q) 7→ spec(Z),
induced by π. We will describe π0 in Section 1.8.
1.6 Wronski map
For a partition λ of n, introduce λ̃ = {λ̃1, . . . , λ̃n} by λ̃i = λi + n− i. Denote
fi(u) = uλ̃i +
λ̃i∑
j=1
λ̃i−j /∈λ̃
fiju
λ̃i−j , i = 1, . . . , n.
Denote Xλ the n-dimensional affine space of n-tuples {f1, . . . , fn} of such polynomials. The
polynomial algebra C[Xλ] = C[fij , i = 1, . . . , n, j ∈ {1, . . . , λ̃i}, λ̃i − j /∈ λ̃] is the algebra of
regular functions on Xλ.
If z = (z1, . . . , zn) are coordinates on Cn, then σ = (σ1, . . . , σn), where σa is the a-th
elementary symmetric function of z1, . . . , zn, are coordinates on C(n) = Cn/Sn.
For arbitrary functions g1(u), . . . , gn(u), introduce the Wronskian determinant by the formula
Wr(g1(u), . . . , gn(u)) = det
g1(u) g′1(u) . . . g
(n−1)
1 (u)
g2(u) g′2(u) . . . g
(n−1)
2 (u)
. . . . . . . . . . . .
gn(u) g′n(u) . . . g
(n−1)
n (u)
.
We have
Wr(f1(u), . . . , fn(u)) =
∏
16i<j6n
(
λ̃j − λ̃i
)
·
(
un +
n∑
a=1
(−1)aWau
n−a
)
with W1, . . . ,Wn ∈ C[Xλ]. Define an algebra homomorphism
Wλ : C[C(n)]→ C[Xλ], σa 7→Wa.
The corresponding map Wrλ : Xλ → C(n) is called the Wronski map.
KZ Characteristic Variety as the Zero Set of Classical Calogero–Moser Hamiltonians 5
Denote
X0
λ = Xλ ∩Wr−1λ
((
Cn −∆
)
/Sn
)
.
Irreducible representations of the symmetric group Sn are labeled by partitions λ of n. Denote
by dλ the dimension of the irreducible representation corresponding to λ. The Wronski map is
a finite map of degree dλ, see, for example, [18].
1.7 Universal differential operator on Xλ
Given an n×n matrix A with possibly noncommuting entries aij , we define the row determinant
to be
rdetA =
∑
σ∈Sn
(−1)σa1σ(1)a2σ(2) · · · anσ(n).
Let x = (f1, . . . , fn) be a point of Xλ. Define the differential operator Dλ,x by
Dλ,x =
∏
16i<j6n
(
λ̃j − λ̃i
)−1 · rdet
f1(u) f ′1(u) . . . f
(n)
1 (u)
f2(u) f ′2(u) . . . f
(n)
2 (u)
. . . . . . . . . . . .
1 ∂ . . . ∂n
,
where ∂ = d/du. It is a differential operator in variable u,
Dλ,x =
∑
06i6j6n
Pij(x)un−j ∂n−i, (1.5)
with Pij ∈ C[Xλ]. By formulae (2.11) and (2.3) in [18], we have
n∑
i=1
Pii
n∏
j=i+1
(s+ j) =
n∏
j=1
(s− λj + j), (1.6)
where s is an independent formal variable.
Let x ∈ X0
λ. Fix zx = (z1,x, . . . , zn,x) ∈ Cn corresponding to Wrλ(x) ∈ C(n). Then
Dλ,x =
n∏
a=1
(u− za,x)
(
∂n −
n∑
a=1
1
u− za,x
∂n−1
+
n∑
a=1
1
u− za,x
−pa,x +
∑
b 6=a
1
za,x − zb,x
∂n−2 + · · ·
for suitable numbers px = (p1,x, . . . , pn,x), see Lemma 3.1 in [16].
Lemma 1.4. The map
ψλ : X0
λ → T ∗
(
Cn −∆
)
/Sn, x 7→
(
zx,px
)
,
is an embedding whose image is Specλ /Sn.
Lemma 1.4 is proved in Section 2.3.
6 E. Mukhin, V. Tarasov, and A. Varchenko
1.8 Description of π0
Theorem 1.5.
(i) The irreducible components of the subvariety C0
n ⊂ Cn are naturally labeled by partitions λ
of n, C0
n = ∪λC0
λ, where C0
λ is the closure of ξ(Specλ) in Cn.
(ii) For any λ, the equations Qa = 0, a = 1, . . . , n, define C0
λ in Cn with multiplicity dλ.
(iii) The irreducible components of C0
n do not intersect. Each component is an n-dimensional
submanifold of Cn isomorphic to an n-dimensional affine space, [23].
(iv) Let λ be a partition of n. Then there is an embedding ϕλ : Xλ → C0
n whose image is C0
λ
and such that the following diagram is commutative:
Xλ C0
λ
-
ϕλ
C(n)
A
A
AU
Wrλ
�
�
��
π0
C.f. the statements (i)–(iii) with results in [8].
The map ϕλ is given by the following construction. The restriction of ϕλ to X0
λ is the
composition ξ ◦ ψλ, where ξ is given by (1.4). This map extends from X0
λ to an embedding
Xλ → C0
n, see Section 2.5.
Parts (i) and (ii) of Theorem 1.5 are proved in Section 2.4. Parts (iii) and (iv) of Theorem 1.5
are proved in Section 2.5.
Remark. It follows from Theorem 1.5 that π−1(0 × 0) consists of points labeled by partitions
and the multiplicity of the point corresponding to a partition λ equals (dλ)2.
The fact that the points of π−1(0×0) are labeled by partitions was explained in [4]. The fact
that the multiplicity equals (dλ)2 was formulated in [4] as Conjecture 17.14 and proved in [8].
2 Proofs
2.1 Proof of Theorem 1.3
Assume that a point z = (z1, . . . , zn) has distinct coordinates. The Bethe ansatz construction
assigns an eigenvector ω(z, t) of Gaudin Hamiltonians Ha(z) on Sing V ⊗n[λ] to a critical point
(z, t) of the master function Φλ(z, t), see [1, 12, 19, 20],
Ha(z)ω(z, t) =
∂Φλ
∂za
(z, t)ω(z, t), a = 1, . . . , n. (2.1)
Formula (2.1) shows that Lλ ⊂ Specλ. By Theorem 6.1 in [19] the Bethe vectors form a basis
of Sing V ⊗n[λ] for generic z ∈ Cn −∆. This proves Theorem 1.3.
2.2 Proof of part (i) of Theorem 1.2
It is easy to see that Sing V ⊗n[λ] and the action on it of the Hamiltonians Ha(z) do not depend
on N . Hence, SpecN,λ does not depend on N .
Another (less straightforward) proof of part (i) follows from formula (2.1) and the fact
that ΦN,λ does not depend on N .
KZ Characteristic Variety as the Zero Set of Classical Calogero–Moser Hamiltonians 7
2.3 Proof of Lemma 1.4
Lemma 2.1. For every λ, the spectral variety Specλ /Sn ⊂ T ∗(Cn − ∆)/Sn is smooth. For
different λ’s the spectral subvarieties do not intersect.
Proof. Let x ∈ X0
λ and Wrλ(x) be a projection of zx = (z1,x, . . . , zn,x). By [18], the points
x ∈ X0
λ are in a one-to-one correspondence with the eigenvectors of the Gaudin Hamiltonians on
Sing V ⊗n[λ]. Denote vx the eigenvector corresponding to x. By Lemma 3.1 in [16] the numbers
px = (p1,x, . . . , pn,x) are eigenvalues of H1(zx), . . . ,Hn(zx) on vx. Hence, the image of ψλ is
Specλ /Sn.
By Theorem 3.2 in [16], the coordinates z1,x, . . . , zn,x, p1,x, . . . , pn,x generate all functions
on X0
λ. This proves that ψλ is an embedding of X0
λ to T ∗(Cn −∆)/Sn.
By Theorem 3.2 in [16], the coefficients Pij(x) in (1.5) are given by some universal functions
in zx, px independent of λ. Hence formula (1.6) implies that the spectral varieties for different
λ’s do not intersect. �
Lemma 2.2. For every λ the spectral variety Specλ lies in the variety L0 defined in (1.2).
Proof. Let x ∈ X0
λ and zx = (z1,x, . . . , zn,x) corresponds to Wrλ(x). By Theorem 3.2 in [16],
det
(
(u− Zx)(v −Qx)− 1
)
=
∑
06i6j6n
Pij(x)un−jvn−i,
where Zx = diag(z1,x, . . . , zn,x), Qx is given by (1.1) in terms of zx and px, and Pij(x) are given
by (1.5). This implies that det(v −Qx) = vn and hence Specλ ⊂ L0. �
2.4 Proofs of parts (i) and (ii) of Theorem 1.5,
part (ii) of Theorem 1.2, and Theorem 1.1
Consider the Lie algebra gln with standard generators eij . Fix a set of complex numbers q =
(q1, . . . , qn). Consider the weight subspace
V ⊗n[1, . . . , 1] =
{
w ∈ V ⊗n | eiiw = w, i = 1, . . . , n
}
and the Gaudin Hamiltonians
Ha(z, q) =
n∑
i=1
qie
(a)
ii +
n∑
i,j=1
∑
b 6=a
e
(a)
ij e
(b)
ji
za − zb
,
which generalize the Gaudin Hamiltonians in (1.3). The generalized Gaudin Hamiltonians act
on V ⊗n[1, . . . , 1]. By Theorem 5.3 in [17], for generic (z, q) the generalized Gaudin Hamilto-
nians H1(z, q), . . . ,Hn(z, q) have an eigenbasis in V ⊗n[1, . . . , 1]. By Theorem 4.3 in [15] the
eigenvectors are in one-to-one correspondence with the preimages of the point (z, q) under the
map π : Cn → C(n) × C(n). This identification sends an eigenvector w with Ha(z, q)w = paw,
a = 1, . . . , n, to the point (Z,Q), where Z = diag(z1, . . . , zn) and Q is defined by (1.1).
The gln-action on every eigenvector of the Gaudin Hamiltonians Ha(z, q = 0) on the space
Sing V ⊗n[λ] generates a dλ-dimensional subspace in V ⊗n[1, . . . , 1] of eigenvectors of Ha(z, q=0).
The identification of Theorem 4.3 in [15] implies that Specλ has multiplicity dλ when defined by
the equations Qa(z,p) = 0, a = 1, . . . , n. More precisely, the identification tells that dλ points
of the π−1(z, q) collide to one point of Specλ, when q → 0. This proves part (ii) of Theorem 1.5.
Since
∑
λ d
2
λ = n!, we conclude that L0 = ∪λ Specλ. This proves part (i) of Theorem 1.5,
part (ii) of Theorem 1.2, and Theorem 1.1.
8 E. Mukhin, V. Tarasov, and A. Varchenko
2.5 Proof of parts (iii) and (iv) of Theorem 1.5
Every point of Xλ is a point of Wilson’s adelic Grassmannian Grad(n), which is identified with
the Calogero–Moser space Cn by the main Theorem 5.1 in [23]. For generic points of Xλ the
map is x 7→ (Zx, Qx). The induced map of functions C[Cn] → C[Xλ] is constructed as follows.
Consider P (u, v) = det((u−Z)(v−Q)−1) as a polynomial in u, v whose coefficients are functions
on Cn, P (u, v) =
∑
i,j P̃iju
n−jvn−i. The coefficients P̃ij generate C[Cn], see Lemma 4.1 in [15].
The map C[Cn]→ C[Xλ] is defined by the formula P̃ij 7→ Pij , where Pij are given by (1.5), see
Theorem 4.3 in [15]. By Lemma 3.4 in [18] the image of this map is C[Xλ]. Hence Xλ → Cn
is an embedding. By formula (1.6) the images do not intersect for different λ’s. This proves
parts (iii) and (iv) of Theorem 1.5.
3 Further remarks
Fix distinct complex numbers q = (q1, . . . , qn). Let σa(q), a = 1, . . . , n, be the elementary
symmetric functions of q. Define
Lq =
{
(z,p) ∈ T ∗(Cn −∆) | Qa(z,p) = σa(q), a = 1, . . . , n
}
.
Theorem 3.1. The subvariety Lq is irreducible, smooth, and Lagrangian.
Define the spectral variety of the Gaudin Hamiltonians Ha(z, q), a=1, . . . , n, on V ⊗n[1, . . . , 1]
by the formula
Specq =
{
(z,p) ∈ T ∗(Cn −∆) | ∃ v ∈ V ⊗n[1, . . . , 1] with Ha(z, q)v = pav, a = 1, . . . , n
}
.
Theorem 3.2. We have Lq = Specq.
Consider the set of n(n− 1)/2 variables
t =
(
t
(1)
1 , . . . , t
(1)
n−1, . . . , t
(n−2)
1 , t
(n−2)
2 , t
(n−1)
1
)
and the affine space Cn × Cn(n−1)/2 with coordinates z, t. Consider the master function Φq :
Cn × Cn(n−1)/2 → C,
Φq(z, t) =
∑
16a<b6n
log(za − zb)−
n∑
a=1
n−1∑
i=1
log
(
t
(1)
i − za
)
+ 2
n−1∑
k=1
∑
16i<j6n−k
log
(
t
(k)
i − t
(k)
j
)
−
n−2∑
k=0
n−k∑
i=1
n−k−1∑
j=1
log
(
t
(k)
i − t
(k+1)
j
)
+
n−1∑
k=1
n−k∑
i=1
(qk+1 − qk)t
(k)
i + q1
n∑
a=1
za,
see [5]. Denote by Critq the critical set of Φq with respect to t,
Critq =
{
(z, t) ∈ Cn × Cn(n−1)/2 | dtΦq(z, t) = 0
}
.
Denote by L̃q ⊂ T ∗(Cn −∆) the image of the map
Critq → T ∗(Cn −∆), (z, t) 7→ (z,p), where pa =
∂Φq
∂za
(z, t), a = 1, . . . , n.
KZ Characteristic Variety as the Zero Set of Classical Calogero–Moser Hamiltonians 9
Theorem 3.3. We have L̃q ⊂ Specq and Specq is the closure of L̃q in T ∗(Cn −∆).
Consider the canonical map π : Cn → C(n)×C(n), (Z,Q) 7→ (spec(Z), spec(Q)). Denote by Cqn
the subvariety π−1(C(n)×q) ⊂ Cn. Identifying C(n)×q with C(n) we get a map πq : Cqn → C(n),
(Z,Q) 7→ spec(Z), induced by π.
Denote
fi(u) = eqiu(u+ fi1), i = 1, . . . , n.
Denote by Xq the n-dimensional affine space of n-tuples {f1, . . . , fn} of such quasiexponentials.
The polynomial algebra C[Xλ] = C[f11, . . . , fn1] is the algebra of regular functions on Xq. We
have
Wr(f1(u), . . . , fn(u)) = e(q1+···+qn)u
∏
16i<j6n
(qj − qi) ·
(
un +
n∑
a=1
(−1)aWau
n−a
)
with W1, . . . ,Wn ∈ C[Xq]. Define an algebra homomorphism
Wq : C[C(n)]→ C[Xq], σa 7→Wa.
Let Wrq : Xq → C(n) be the corresponding map of spaces.
Theorem 3.4.
(i) The equations Qa = qa, a = 1, . . . , n, define Cqn in Cn with multiplicity 1.
(ii) There is an embedding ϕq : Xq → Cqn whose image is Cqn and such that the following
diagram is commutative:
Xq Cqn
-
ϕq
C(n)
A
A
AU
Wrq
�
�
��
π0
The map ϕq is given by the following construction. For x = (f1, . . . , fn) ∈ Xq, define the
differential operator Dq,x by
Dq,x = e−(q1+···+qn)u
∏
16i<j6n
(qj − qi)−1 rdet
f1(u) f ′1(u) . . . f
(n)
1 (u)
f2(u) f ′2(u) . . . f
(n)
2 (u)
. . . . . . . . . . . .
1 ∂ . . . ∂n
.
Then
Dq,x =
n∑
i=0
n∑
j=0
Pij(x)un−j∂n−i,
where Pij ∈ C[Xq].
Denote X0
q = Xq ∩Wr−1λ
(
(Cn −∆)/Sn
)
and consider the map
ψq : X0
q → T ∗
(
Cn −∆
)
/Sn, x 7→
(
zx,px
)
,
10 E. Mukhin, V. Tarasov, and A. Varchenko
where zx ∈ Cn projects to Wrq(x) ∈ C(n) and px = (p1,x, . . . , pn,x),
pa,x = − Res
u=za,x
n∑
j=0
P2,j(x)un−j
n∏
i=1
(u− qi)
+
∑
b6=a
1
za,x − zb,x
.
Then the restriction of ϕq to X0
q is the composition ξ ◦ ψq, where ξ is given by (1.4). This map
extends from X0
q to an embedding Xq → Cqn .
The proofs of Theorems 3.1–3.4 are basically the same as the proofs of Theorems 1.1–1.5.
Acknowledgments
The authors thank V. Schechtman and D. Arinkin for useful discussions. The third author thanks
the MPIM, HIM, and IHES for the hospitality. E. Mukhin was supported in part by NSF grant
DMS-0900984. V. Tarasov was supported in part by NSF grant DMS-0901616. A. Varchenko
was supported in part by NSF grants DMS-0555327 and DMS-1101508.
References
[1] Babujian H.M., Off-shell Bethe ansatz equations and N -point correlators in the SU(2) WZNW theory,
J. Phys. A: Math. Gen. 26 (1993), 6981–6990.
[2] Cherednik I., Integration of quantum many-body problems by affine Knizhnik–Zamolodchikov equations,
Adv. Math. 106 (1994), 65–95.
[3] Cherednik I., Lectures on Knizhnik–Zamolodchikov equations and Hecke algebras, in Quantum Many-Body
Problems and Representation Theory, MSJ Mem., Vol. 1, Math. Soc. Japan, Tokyo, 1998, 1–96.
[4] Etingof P., Ginzburg V., Symplectic reflection algebras, Calogero–Moser space, and deformed Harish-
Chandra homomorphism, Invent. Math. 147 (2002), 243–348, math.AG/0011114.
[5] Felder G., Markov Y., Tarasov V., Varchenko A., Differential equations compatible with KZ equations,
Math. Phys. Anal. Geom. 3 (2000), 139–177, math.QA/0001184.
[6] Felder G., Veselov A.P., Polynomial solutions of the Knizhnik–Zamolodchikov equations and Schur–Weyl
duality, Int. Math. Res. Not. 2007 (2007), no. 15, 21 pages, math.RT/0610383.
[7] Felder G., Veselov A.P., Shift operators for the quantum Calogero–Sutherland problems via Knizhnik–
Zamolodchikov equation, Comm. Math. Phys. 160 (1994), 259–273.
[8] Finkelberg M., Ginzburg V., Calogero–Moser space and Kostka polynomials, Adv. Math. 172 (2002), 137–
150, math.RT/0110190.
[9] Gaudin M., La fonction d’onde de Bethe, Collection du Commissariat à l’Énergie Atomique: Série Scien-
tif ique, Masson, Paris, 1983.
[10] Givental A., Kim B., Quantum cohomology of flag manifolds and Toda lattices, Comm. Math. Phys. 168
(1995), 609–641, hep-th/9312096.
[11] Gordon I., Baby Verma modules for rational Cherednik algebras, Bull. London Math. Soc. 35 (2003), 321–
336, math.RT/0202301.
[12] Jurčo B., Classical Yang–Baxter equations and quantum integrable systems, J. Math. Phys. 30 (1989),
1289–1293.
[13] Kazhdan D., Kostant B., Sternberg S., Hamiltonian group actions and dynamical systems of Calogero type,
Comm. Pure Appl. Math. 31 (1978), 481–507.
[14] Matsuo A., Integrable connections related to zonal spherical functions, Invent. Math. 110 (1992), 95–121.
[15] Mukhin E., Tarasov V., Varchenko A., Bethe algebra, Calogero–Moser space and Cherednik algebra,
arXiv:0906.5185.
[16] Mukhin E., Tarasov V., Varchenko A., Gaudin Hamiltonians generate the Bethe algebra of a tensor power
of the vector representation of glN , St. Petersburg Math. J. 22 (2011), 463–472, arXiv:0904.2131.
http://dx.doi.org/10.1088/0305-4470/26/23/037
http://dx.doi.org/10.1006/aima.1994.1049
http://dx.doi.org/10.1007/s002220100171
http://arxiv.org/abs/math.AG/0011114
http://dx.doi.org/10.1023/A:1009862302234
http://arxiv.org/abs/math.QA/0001184
http://dx.doi.org/10.1093/imrn/rnm0046
http://arxiv.org/abs/math.RT/0610383
http://dx.doi.org/10.1007/BF02103276
http://dx.doi.org/10.1006/aima.2002.2083
http://arxiv.org/abs/math.RT/0110190
http://dx.doi.org/10.1007/BF02101846
http://arxiv.org/abs/hep-th/9312096
http://dx.doi.org/10.1112/S0024609303001978
http://arxiv.org/abs/math.RT/0202301
http://dx.doi.org/10.1063/1.528305
http://dx.doi.org/10.1002/cpa.3160310405
http://dx.doi.org/10.1007/BF01231326
http://arxiv.org/abs/0906.5185
http://dx.doi.org/10.1090/S1061-0022-2011-01152-5
http://arxiv.org/abs/0904.2131
KZ Characteristic Variety as the Zero Set of Classical Calogero–Moser Hamiltonians 11
[17] Mukhin E., Tarasov V., Varchenko A., Generating operator of XXX or Gaudin transfer matrices has
quasi-exponential kernel, SIGMA 3 (2007), 060, 31 pages, math.QA/0703893.
[18] Mukhin E., Tarasov V., Varchenko A., Schubert calculus and representations of the general linear group,
J. Amer. Math. Soc. 22 (2009), 909–940, arXiv:0711.4079.
[19] Mukhin E., Varchenko A., Norm of a Bethe vector and the Hessian of the master function, Compos. Math.
141 (2005), 1012–1028, math.QA/0402349.
[20] Reshetikhin N., Varchenko A., Quasiclassical asymptotics of solutions to the KZ equations, in Geometry,
Topology, & Physics, Conf. Proc. Lecture Notes Geom. Topology, Vol. IV, Int. Press, Cambridge, MA, 1995,
293–322, hep-th/9402126.
[21] Schechtman V.V., Varchenko A.N., Arrangements of hyperplanes and Lie algebra homology, Invent. Math.
106 (1991), 139–194.
[22] Schechtman V.V., Varchenko A.N., Hypergeometric solutions of Knizhnik–Zamolodchikov equations, Lett.
Math. Phys. 20 (1990), 279–283.
[23] Wilson G., Collisions of Calogero–Moser particles and an adelic Grassmannian, Invent. Math. 133 (1998),
1–41.
http://dx.doi.org/10.3842/SIGMA.2007.060
http://arxiv.org/abs/math.QA/0703893
http://dx.doi.org/10.1090/S0894-0347-09-00640-7
http://arxiv.org/abs/0711.4079
http://dx.doi.org/10.1112/S0010437X05001569
http://arxiv.org/abs/math.QA/0402349
http://arxiv.org/abs/hep-th/9402126
http://dx.doi.org/10.1007/BF01243909
http://dx.doi.org/10.1007/BF00626523
http://dx.doi.org/10.1007/BF00626523
http://dx.doi.org/10.1007/s002220050237
1 Statement of results
1.1 Motivation
1.2 Classical Calogero–Moser system
1.3 Gaudin Hamiltonians and KZ characteristic variety
1.4 Master function generates Spec
1.5 Calogero-Moser space Cn and cotangent bundle T*(Cn-)
1.6 Wronski map
1.7 Universal differential operator on X
1.8 Description of 0
2 Proofs
2.1 Proof of Theorem 1.3
2.2 Proof of part (i) of Theorem 1.2
2.3 Proof of Lemma 1.4
2.4 Proofs of parts (i) and (ii) of Theorem 1.5, part (ii) of Theorem 1.2, and Theorem 1.1
2.5 Proof of parts (iii) and (iv) of Theorem 1.5
3 Further remarks
References
|