# Admissible representations, multiplicity-free representations and visible actions on non-tube type Hermitian symmetric spaces.

1. Introduction. Let us begin this paper with two (apparently, quite different) facts on the relationship between multiplicities of representations and geometry.The first fact is proved by T. Kobayashi and T. Oshima in [16]. Let [[??].sub.ad] be the set of equivalence classes of irreducible admissible representations of a real reductive Lie group G. Here, a representation [pi] of G is admissible if dim [Hom.sub.K]([mu], [pi]|K) < [infinity] for any irreducible representation [mu] of a maximal compact subgroup K of G. For a pair G [contains] H of algebraic reductive groups, the homogeneous space G/H is real spherical if the dimension of intertwiners for any irreducible admissible representation [pi] [member of] [[??].sub.ad] into the space [C.sup.[infinity]]G/H) of continuous functions on G/H is finite, namely, dim [Hom.sub.G]([pi], [C.sup.[infinity]](G/H)) < [infinity], and vice versa ([16, Theorem A]). Here, G/H is real spherical if there exists an open P-orbit in G/H where P is a minimal parabolic subgroup of G ([9]). Moreover, the complexification [G.sub.C]/[H.sub.C] of G/H is spherical, namely, [G.sub.C]/[H.sub.C] has an open Borel orbit, if and only if the multiplicity is uniformly bounded in the sense of sup [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] dim Hom ([pi], [C.sup.[infinity]](G/H)) < [infinity] ([16, Theorem B]).

The second fact is concerned with the complex geometry. Let H be a Lie group. The space O(D, V) of holomorphic sections of an H-equivariant Hermitian holomorphic vector bundle V [right arrow] D over a complex manifold D defines a continuous representation of H with respect to the Frechet topology. Let H be a unitary representation of H which is realized in O(D, V), namely, there exists a continuous and injective H-homomorphism from the Hilbert space H into O(D, V). Now, we consider a general setting where the H-action on D is not transitive, and also a basic question when H is multiplicity-free. In general, the property of multiplicity-freeness of H is not fulfilled even though each fiber [V.sub.x] (x [member of] D) is multiplicity-free as a representation of the isotropy subgroup [H.sub.x]. However, this does hold if H acts on the base space D in a strongly visible fashion in the sense of [11]. We say that this theory is propagation theory of multiplicity-freeness which is established by T. Kobayashi (see [12, 17]). A part of the idea of proof goes back to Gelfand-Kazhdan, S. Kobayashi [7], and Faraut-Thomas [4].

Among irreducible bounded symmetric domain, there are two types: Hermitian symmetric spaces of tube type; Hermitian symmetric spaces of non-tube type. The Hermitian symmetric spaces G/K = SU(n, n)/S(U(n) x U(n)), S[O.sup.*](4n)/U(2n), S[O.sub.0](n, 2)/(SO(n) x SO(2)), Sp(n, R)/U(n), and [E.sub.7(-25)]/([E.sub.6] x T) are of tube type, whereas G/K = SU(p, q)/S(U(p) x U(q)) with p [not equal to] q, S[O.sup.*](4n + 2)/ U(2n + 1), and [E.sub.6(-14)]/(Spin(10) x T) are of nontube type. We shall see in Theorem 1 below that aforementioned two theories applied to the associated Stein manifolds [G.sub.C]/[[K.sub.C], [K.sub.C]] will reveal sharp differences between tube and non-tube types, giving new characterization of tube type domains from the viewpoint of multiplicities in some branching laws and also from the viewpoint of visibility of holomorphic actions.

2. Visible actions on complex manifolds. Let us review from [11] (see also [12]) the notion of strongly visible actions on complex manifolds. A holomorphic action of a Lie group H on a connected complex manifold D is called strongly visible if there exist a real submanifold S in D and an anti-holomorphic diffeomorphism a on D satisfying the following conditions:

(V.1) S meets every H-orbit in D,

(5.1) [sigma][|.sub.S] = [id.sub.S];

(5.2) [sigma] preserves each H-orbit in D.

We say that the submanifold S is a slice. The slice S is automatically a totally real submanifold ([12, Remark 3.3.2]).

We allow that S meets every H-orbit twice and more than twice, namely, S is not necessary a complete representative of H-orbits in D.

In Kobayashi's original definition [12, Definition 3.3.1], the concept of strongly visible actions is slightly wider, namely, he calls that this action is strongly visible if a complex manifold D contains an open set satisfying the conditions (V.1)-(S.2). For an application to multiplicity-free representations, this wider definition is sufficient. However, for simplicity, we adopt the narrower one throughout this paper.

3. Multiplicity-freeness and visible action. Taking a pair [G.sub.u] [contains] K of compact Lie groups as an example, the theory of visible actions gives a geometric explanation for multiplicity-free representations as follows: We want to understand which irreducible representation [mu] of K the multiplicity-freeness holds in the sense that dim [Hom.sub.K]([mu], [lambda][|.sub.K]) [less than or equal to] 1 for any irreducible representation [lambda] of [G.sub.u]. By the Frobenius reciprocity, this dimension is nothing but the one of intertwiners from [lambda] to the space O(D, V) of holomorphic sections for the [G.sub.C]-equivariant Hermitian holomorphic vector bundle V = [G.sub.C] x [K.sub.C] [mu] on D = [G.sub.C]/[K.sub.C]. Then, the multiplicity-freeness holds if the [G.sub.u]-action on D is strongly visible and [mu] is multiplicity-free as a representation of M, where M is the stabilizer of a generic element of a slice for the strongly visible [G.sub.u]-action on D. If ([G.sub.u], K) is a symmetric pair, then a slice can be taken as the A-orbit under the Cartan decomposition [G.sub.C] = [G.sub.u]A[K.sub.C] for symmetric [G.sub.C]/[K.sub.C]. Thus, M is the centralizer of A in K.

Not only for finite-dimensional representations of a compact Lie group but also for infinite-dimensional representations of a non-compact real form, we give an explanation of the multiplicity-freeness by the complex geometric viewpoint. In fact, by switching a compact real form [G.sub.u] by a non-compact one [G.sub.R] in the above example, we can show that the Hilbert space [L.sup.2]([G.sub.R]/K, [mu]) of square integrable sections on the non-compact [G.sub.R]/K is multiplicityfree as a representation of [G.sub.R].

4. Characterization of tube type Hermitian symmetric spaces. We are ready to state our main results of this paper.

Let G/K be a non-compact irreducible Hermitian symmetric space. Then, K has a one-dimensional center, and hence the commutator subgroup [K.sup.s] := [K, K] is of codimension one in K. Therefore, the homogeneous space G/[K.sup.s] is not a symmetric space. We note that the complexified [G.sub.C]/[K.sup.s.sub.C] of G/[K.sup.s] is a Stein manifold by Matsushima's theorem.

Our main result characterizes tube type (or non-tube type) among Hermitian symmetric spaces by visible actions, and also by multiplicities in branching laws, and is stated as follows:

Theorem 1. The following six conditions are equivalent for a non-compact irreducible Hermitian symmetric space G/K:

(i) G/K is of non-tube type.

(ii) [G.sub.C]/[K.sup.s.sub.C] is spherical.

(iii) The action of a compact real form [G.sub.u] of [G.sub.C] on [G.sub.C]/[K.sup.s.sub.C] is strongly visible.

(iv) The [K.sup.s]-action on the Hermitian symmetric space G/K is strongly visible.

(v) The restriction [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is [K.sup.s]-admissible for a (equivalently, for any) holomorphic discrete series representation [pi] of G.

(vi) For a (equivalently, for any) holomorphic discrete series representation [pi] of G of scalar type, the restriction [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is multiplicity-free.

Concerning to (v) of Theorem 1, it follows from the corollary of [10, Theorem 2.4.5] that the restriction [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is [K.sup.s]-admissible, namely, the irreducible decomposition of [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] contains only discrete spectra and dim [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] holds for any [mu] [member of] [[??].sup.s].

Our strategy of the proof of Theorem 1 is as follows: Kramer's classification of spherical affine irreducible complex homogeneous spaces [18] shows the equivalence (i) [??] (ii). The equivalence between (i) and (iii) is proved in [19]. We discuss the equivalence (i) [??] (v) in Section 4.1; the implication (vi) [??] (i) in Section 4.2; (i) [??] (iv) in Section 4.3; and (iv) [??] (vi) in Section 4.4. We summarize the strategy as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

4.1. Proof of (i) [??] (v). Our proof of the equivalence (i) [??] (v) is based on Kobayashi's criterion for admissible restrictions of representations [10, 13], namely, criterion for discretely decomposability of the restriction with finite-multiplicities. We remark that the relation between (i) and (v) was announced by Duflo-Vargas [1].

First, we summarize his criterion briefly, see [13, Section 6.2]. Let [[??].sub.0] be the Lie algebra of a compact Lie group K. We fix a maximal torus [t.sub.0] of [[??].sub.0] and a positive system [DELTA].sup.+]([[??].sub.0], [t.sub.0]). We write [C.sub.+] [subset] [square root of -1][t.sup.*.sub.0] for the corresponding closed Weyl chamber. We regard the unitary dual [??] as a lattice of [square root of -1][t.sup.*.sub.0] and put [[LAMBDA].sub.+] := [??] [intersection] [C.sub.+]. For a representation [??] of G, we define the K-support of [??] by

[Supp.sub.K]([??]) := {[lambda] [member of] [[LAMBDA].sub.+] : [Hom.sub.K]([[tau].sub.[lambda]], [??][|.sub.K]) [not equal to] 0},

and the asymptotic K-support of [??] introduced by Kashiwara-Vergne [6], see also [10],

(1) A[S.sub.K]([??]) := [Supp.sub.K]([??])[infinity]

which is a closed cone in [C.sub.+]. Here, the asymptotic cone S[infinity] for a subset S in a vector space [R.sup.N] is defined by

S[infinity] := {y [member of] [R.sup.N] : there exists a sequence {([y.sub.n], [e.sub.n])} [subset] S x [R.sub.+] such that

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

nn nn

Let L be a closed subgroup of K and [[??].sub.0] the Lie algebra of L. The inclusion [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] defines the natural projection pr : [[??].sup.*.sub.0] [right arrow] [[??].sup.*.sub.0]. We set [[??].sup.[perpendicular to].sub.0] := ker pr and define a closed cone [C.sub.K](L) in [square root of -1][t.sup.*.sub.0] by

(2) [C.sub.K](L) := [C.sub.+] [intersection] [square root of -1] [Ad.sup.*](K)[[??].sup.[perpendicular to].sub.0].

The criterion for admissible restrictions of representations is written by two closed cones (1), (2) in [C.sub.+] as follows:

Lemma 2 ([13, Theorem 6.3.3]). The restriction [??][|.sub.L] is L-admissible if and only if A[S.sub.K]([??]) [intersection] [C.sub.K](L) = {0}.

Next, we return to our setting of (v). Let [[??].sup.s.sub.0] be the Lie algebra of [K.sup.s], which coincides with the derived ideal [[[??].sub.0], [[??].sub.0]] of [[??].sub.0]. In view of the natural projection pr : [[??].sup.*.sub.0] [right arrow] [([[??].sup.s.sub.0]).sup.*], the kernel [([[??].sup.s.sub.0]).sup.[perpendicular to]] is isomorphic to the dual c[([[??].sub.0]).sup.*] of the center c([[??].sub.0]) in [[??].sub.0]. Then, [square root of -1] [Ad.sup.*](K)[([[??].sup.s.sub.0]).sup.[perpendicular to]] = [square root of - 1]c[([[??].sub.0]).sup.*], from which we obtain

(3) [C.sub.K]([K.sup.s]) = [C.sub.+] [intersection] [square root of -1]c[([[??].sub.0]).sup.*].

An explicit formula of the asymptotic K-support A[S.sub.K]([pi]) is given for a holomorphic discrete series representation [pi] of G of scalar type as follows: Let Z [member of] c([[??].sub.0]) be the characteristic element such that g := [g.sub.0] [[cross product].sub.R] C = [??] + [p.sub.+] + [p.sub.-] is the eigenspace decomposition of ad(Z) with eigenvalues 0, [square root of -1], - [square root of -1], respectively. Let [v.sub.1], ..., [v.sub.r] be strongly orthogonal roots in [DELTA]([p.sub.+]) such that [v.sub.1] is the highest root among [DELTA]([p.sub.+]) and that [v.sub.j+1] is the highest root in [DELTA]([p.sub.+]) strongly orthogonal to [v.sub.1], ..., [v.sub.j] where r = rank G/K. By using the K-type formula [20] of [pi] and the stability of A[S.sub.K]([pi]) [10, Lemma 3.1] under the tensor product, we have:

Lemma 3. A[S.sub.K]([pi]) is expressed by

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

Combining [2,3] with Lemma 3, we have:

Lemma 4. {[summation] [a.sub.i][v.sub.i] : [a.sub.1] [greater than or equal to] [a.sub.2] [greater than or equal to] ... [greater than or equal to] [a.sub.r] [greater than or equal to] 0} [intersection] [square root of -1]c[([??]).sup.*] = {0} if and only if G/K is of non-tube type.

Now, we are ready to give a proof of the equivalence (i) [??] (v).

Proof of (i) [??] (v). Since the [K.sup.s]-admissibility is presented by taking the tensor product with finite-dimensional representations [8, Corollary 1.3], it is sufficient for the proof to deal with the case where [pi] is of scalar type. By Lemma 2, the restriction [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is [K.sup.s]-admissible if and only if A[S.sub.K]([pi]) [intersection] [C.sub.K]([K.sup.s]) = {0}. It follows from the equality (3) and Lemma 3 that A[S.sub.K]([pi]) [intersection] [C.sub.K]([K.sup.s]) = {[summation] [a.sub.i][v.sub.i] : [a.sub.1] [greater than or equal to] [a.sub.2] [greater than or equal to] ... [greater than or equal to] [a.sub.r] [greater than or equal to] 0} [intersection] [square root of -1]c[([??]).sup.*] [intersection] [C.sub.+]. Applying Lemma 4 to the right-hand side, we conclude that A[S.sub.K]([pi]) [intersection] [C.sub.K]([K.sup.s]) = {0} if and only if G/K is of non-tube type. Therefore, the equivalence (i) [??] (v) has been proved.

4.2. Proof of (vi) [??] (i). The equivalence (i) [??] (v) brings us to the implication (vi) [??] (i) as follows:

Proof of (vi) [??] (i). Suppose that G/K is of tube type. By the equivalence (i) [??] (v), the restriction [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is not [K.sup.s]-admissible, in particular, not multiplicity-free for any holomorphic discrete series representation [pi] of G. This is the contraposition of the implication (vi) [??] (i).

4.3. Proof of (i) [??] (iv). The key of the proof for the implication (i) [??] (iv) is to construct a slice for the [K.sup.s]-action on G/K explicitly.

Let [g.sub.0], [[??].sub.0], and [[??].sup.2.sub.0] be the Lie algebras of G, K, and [K.sup.s], respectively. We write [g.sub.0] = [[??].sub.0] + [p.sub.0] for the corresponding Cartan decomposition. Let [a.sub.0] be a maximal abelian subspace in [p.sub.0] and A := exp [a.sub.0]. Then, we have the Cartan decomposition

(4) G = KAK.

Let [m.sub.0] be the centralizer of [a.sub.0] in [[??].sub.0]. We recall:

Lemma 5 (cf. [5, Lemma 3.1]). G/K is of tube type if and only if [m.sub.0] [subset] [[??].sup.s.sub.0].

Proof of (i) [??] (iv). Suppose that G/K is of non-tube type. By Lemma 5, [m.sub.0] is not contained in [[??].sup.s.sub.0]. We take X [member of] [m.sub.0] such that X [not member of] [[??].sup.s.sub.0] Then,

(5) [[??].sub.0] = [[??].sup.s.sub.0] + RX,

because [[??].sup.s.sub.0] is of codimension one in [[??].sub.0]. Thus, we obtain

(6) K = [K.sup.s](exp RX) = (exp RX)[K.sup.s].

Combining (4) and (6), we get

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

This implies that the real submanifold S := AK/K meets every [K.sup.s]-orbit in G/K.

The existence of an anti-holomorphic diffeomorphism a on G/K satisfying (S.1) and (S.2) for the [K.sup.s]-action on G/K with S follows from [15, Lemmas 2.2 and 2.4].

Hence, the [K.sup.s]-action on G/K is strongly visible. In particular, one can take a slice S for this action to be dim S = rank G/K.

As a corollary of our proof, we get a new decomposition for the non-symmetric pair (G, [K.sup.s]) as follows:

Theorem 6. For a non-tube type Hermitian symmetric space G/K, one can find an abelian subgroup A of G with dim A = rank G/K such that the following group decomposition holds:

G = [K.sup.s]AK.

4.4. Proof of (iv) [??] (vi). The idea of the proof of the implication (iv) [??] (vi) is based on that of [15, Corollary 6.3].

Let ([pi], H) be a holomorphic discrete series representation of G. It is known that there is a natural injective continuous G-homomorphism from the Hilbert space H to the Frechet space O(G/K, V) consisting of holomorphic sections over a holomorphic line bundle V = G x [sub.K] [mu] for some [mu] [member of] [??]. In order to prove the multiplicity-freeness property of the restriction [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], it is sufficient to show that O(G/K, V) is multiplicity-free as a representation of [K.sup.s].

Proof of (iv) [??] (vi). Let [pi] be of scalar type. Then, each fiber [V.sub.x] is one-dimensional. In particular, the representation of the isotropy subgroup [K.sup.s.sub.x] on the fiber [V.sub.x] is obviously multiplicity-free. If the [K.sup.s]-action on G/K is strongly visible, then the assumption of propagation theory of multiplicity-freeness property [17] is satisfied, from which we conclude that O(G/K, V) is multiplicity-free as a representation of [K.sup.s].

Therefore, the implication (iv) [??] (vi) has been proved.

As a conclusion, the proof of Theorem 1 has been completed.

4.5. Remark. Here is a direct proof of (ii) [??] (v).

Suppose that [G.sub.C]/[K.sup.s.sub.C] is spherical. It follows from the theory of spherical manifolds [16, Theorem B] that there exists a constant C > 0 such that

dim [Hom.sub.G]([pi], [C.sup.[infinity]](G/[K.sup.s], G x [sub.K] [mu])) [less than or equal to] C

for any [pi] [member of] [[??].sub.ad] and [mu] [member of] [[??].sup.s]. Since the left-hand side of this inequality is given by dim [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] by the Frobenius reciprocity, it follows that the restriction [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is admissible.

5. Generalization of Cartan decomposition to non-symmetric [G.sub.C]/[K.sup.s.sub.C]. Let us explain our construction of a slice S satisfying (V.1) for non-tube type G/K from the group theoretic viewpoint, which is an essential part of the proof for the implication (i) [??] (iii) of Theorem 1. More precisely, we find a subset B giving a decomposition

(7) [G.sub.C] = [G.sub.u]B[K.sup.s.sub.C],

which implies that S := B[K.sup.s.sub.C]/[K.sup.s.sub.C] meets every [G.sub.u]-orbit in [G.sub.C]/[K.sup.s.sub.C] (see (11) below).

The key ingredients are as follows: One is that we have a Cartan decomposition for any (not necessary Riemannian) symmetric space G/H of a reductive Lie group G, namely, there exists an abelian subgroup A of G such that G = KAH owing to Flensted Jensen [5]. In the case where H coincides with a maximal compact subgroup K of G, the decomposition G = KAK is nothing but the classical Cartan decomposition (see (4) in Section 4.3). The other is to apply the herringbone stitch method [14] for our setting via the symmetric subgroup [K.sub.C] (see Fig. 1).

Let us retain the notation as in Section 4.3. According to our strategy, we first focus on the symmetric space [G.sub.C]/[K.sub.C]. Then, the Cartan decomposition for symmetric pairs [5] gives the following decomposition

(8) [G.sub.C] = [G.sub.u]A[K.sub.C],

where A := exp [a.sub.0] and [a.sub.0] is a maximal abelian subspace in [p.sub.0].

Next, we treat the one-dimensional complex manifold [K.sub.C]/[K.sup.s.sub.C]. As G/K is of non-tube type, we have the decomposition (5) for some X [member of] [m.sub.0] satisfying X [not member of] [[??].sup.s.sub.0] (see Lemma 5). Then, the complexification [??] = [[??].sub.0] [[cross product].sub.R] C is decomposed as follows:

(9) [??] = [[??].sup.s] + RX + [square root of -1]RX

where [[??].sup.s] = [[??].sup.s.sub.0] [[cross product].sub.R] C. We set [Z.sub.T] = exp RX and [Z.sub.R] = exp [square root of -1]RX. Then, the decomposition (9) gives rise to a global decomposition as follows:

(10) [K.sub.C] = [K.sup.s.sub.C][Z.sub.T][Z.sub.R] = [Z.sub.T][Z.sub.R][K.sup.s.sub.C].

We are ready to apply the herringbone stitch method to our setting. As X [member of] [m.sub.0], three Lie groups [Z.sub.T], [Z.sub.R], A commute with one another. Since [Z.sub.T] is a subgroup of [G.sub.u], we get

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

Therefore, (7) holds if we set

(11) B := A[Z.sub.R] = [Z.sub.R]A.

The point is that our B is still abelian even though [G.sub.C]/[K.sup.s.sub.C] is not symmetric, from which S is a submanifold in D. In this sense, (7) is a generalization of the Cartan decomposition known for semisimple symmetric spaces [5] to some non-symmetric spherical homogeneous spaces. Consequently, we have proved:

Theorem 7. For a non-tube type Hermitian symmetric space G/K, one can find an abelian subgroup B of [G.sub.C] with dim B = rank G/K + 1 such that the following group decomposition holds:

[G.sub.C] = [G.sub.u]B[K.sup.s.sub.C].

In particular, B is given by (11).

6. Conclusion. Finally, we summarize the relationship between the multiplicity-freeness of representation and the complex geometry in our setting. If [G.sub.C]/[K.sup.s.sub.C] is spherical, or equivalently, G/K is of non-tube type, then we have a generalized Cartan decomposition [G.sub.C] = [G.sub.u]B[K.sup.s.sub.C]. Then, our slice for the [G.sub.u]-action on [G.sub.C]/[K.sup.s.sub.C] is given by S := B[K.sup.s.sub.C]/[K.sup.s.sub.C]. Applying the propagation theory of multiplicity-freeness, we obtain another proof for O([G.sub.C]/[K.sup.s.sub.C]) to be multiplicity-free as a representation of [G.sub.C].

doi: 10.3792/pjaa.91.70

References

[1] M. Duflo and J. A. Vargas, Branching laws for square integrable representations, Proc. Japan Acad. Ser. A Math. Sci. 86 (2010), no. 3, 49-54.

[2] J. Faraut, S. Kaneyuki, A. Koranyi, Q.-K. Lu and G. Roos, Analysis and geometry on complex homogeneous domains, Progress in Mathematics, 185, Birkhauser, Boston, MA, 2000.

[3] J. Faraut and A. Koranyi, Analysis on symmetric cones, Oxford Mathematical Monographs, Oxford Univ. Press, New York, 1994.

[4] J. Faraut and E. G. F. Thomas, Invariant Hilbert spaces of holomorphic functions, J. Lie Theory 9 (1999), no. 2, 383-402.

[5] M. Flensted-Jensen, Spherical functions of a real semisimple Lie group. A method of reduction to the complex case, J. Funct. Anal. 30 (1978), no. 1, 106-146.

[6] M. Kashiwara and M. Vergne, K-types and singular spectrum, in Noncommutative harmonic analysis (Proc. Third Colloq., Marseille-Luminy, 1978), 177-200, Lecture Notes in Math., 728, Springer, Berlin, 1979.

[7] S. Kobayashi, Irreducibility of certain unitary representations, J. Math. Soc. Japan 20 (1968), 638-642.

[8] T. Kobayashi, Discrete decomposability of the restriction of [A.sub.q]([lambda]) with respect to reductive subgroups and its applications, Invent. Math. 117 (1994), no. 2, 181-205.

[9] T. Kobayashi, Introduction to harmonic analysis on spherical homogeneous spaces, in Proceedings of 3rd Summer School on Number Theory "Homogeneous Spaces and Automorphism Forms" (Rikkyo University, 1995) and (Yamagata-mura in Nagano, 1995), 22-41 (ed. F. Sato), 1995. (in Japanese).

[10] T. Kobayashi, Discrete decomposability of the restriction of [A.sub.q]([lambda]) with respect to reductive subgroups. II. Micro-local analysis and asymptotic K-support, Ann. of Math. (2) 147 (1998), no. 3, 709-729.

[11] T. Kobayashi, Geometry of multiplicity-free representations of GL(n), visible actions on flag varieties, and triunity, Acta Appl. Math. 81 (2004), no. 1-3, 129-146.

[12] T. Kobayashi, Multiplicity-free representations and visible actions on complex manifolds, Publ. Res. Inst. Math. Sci. 41 (2005), no. 3, 497-549.

[13] T. Kobayashi, Restrictions of unitary representations of real reductive groups, in Lie theory, Progr. Math., 229, Birkhauser, Boston, MA, 2005, pp. 139-207.

[14] T. Kobayashi, A generalized Cartan decomposition for the double coset space (U([n.sub.1]) x U([n.sub.2]) x U([n.sub.3]))\U(n)/(U(p) x U(q)), J. Math. Soc. Japan 59 (2007), no. 3, 669-691.

[15] T. Kobayashi, Visible actions on symmetric spaces, Transform. Groups 12 (2007), no. 4, 671-694.

[16] T. Kobayashi and T. Oshima, Finite multiplicity theorems for induction and restriction, Adv. Math. 248 (2013), 921-944.

[17] T. Kobayashi, Propagation of multiplicity-freeness property for holomorphic vector bundles, in Lie groups: structure, actions, and representations, Progr. Math., 306, Birkhauser/Springer, New York, 2013, pp. 113-140.

[18] M. Kramer, Spharische Untergruppen in kompakten zusammenhangenden Liegruppen, Compositio Math. 38 (1979), no. 2, 129-153.

[19] A. Sasaki, A characterization of non-tube type Hermitian symmetric spaces by visible actions, Geom. Dedicata 145 (2010), 151-158.

[20] W. Schmid, Die Randwerte holomorpher Funktionen auf hermitesch symmetrischen Raaumen, Invent. Math. 9 (1969/1970), 61-80.

By Atsumu Sasaki

Department of Mathematics, Faculty of Science, Tokai University, 4-1-1 Kitakaname, Hiratsuka-shi, Kanagawa 259-1292, Japan

(Communicated by Masaki Kashiwara, M.J.A., April 13, 2015)

2010 Mathematics Subject Classification. Primary 22E46; Secondary 32M15, 14M17.

Printer friendly Cite/link Email Feedback | |

Author: | Sasaki, Atsumu |
---|---|

Publication: | Japan Academy Proceedings Series A: Mathematical Sciences |

Article Type: | Report |

Geographic Code: | 1USA |

Date: | May 1, 2015 |

Words: | 4461 |

Previous Article: | Middle convolution and non-Schlesinger deformations. |

Next Article: | Some problems of hypergeometric integrals associated with hypersphere arrangement. |

Topics: |