A characterization of bounded symmetric domains by curvature
A characterization of bounded symmetric domains by curvature
复制标题
通过曲率表征有界对称域
DOI:
10.1090/s0002-9947-1983-0688960-x
复制
发表时间:
1983
影响因子:
1.3
通讯作者:
Dotti Miatello
中科院分区:
文献类型:
--
作者:
BY Curvature;Dotti Miatello
This paper will prove that a bounded homogeneous domain is symmetric if and only if, in the Bergman metric, all sectional curvatures are nonpositive. Introduction. It is well known that a bounded symmetric domain has nonpositive sectional curvature in the Bergman metric. This paper is devoted to the converse, namely, that a bounded homogeneous domain (equivalently, homogeneous Siegel domain) which has nonpositive sectional curvature in the Bergman metric must be a symmetric domain. The proof uses the techniques of normal j-algebras [13], as well as results of Vinberg [16, 17], the first author [1-3], and, indirectly, Dorfmeister [5, 6]. The body of the paper is in two sections. In the first, we derive certain relations between the dimensions of the root spaces in an irreducible normal j-algebra, which, in the presence of the additional curvature assumption, show the equality of the dimensions of all root spaces corresponding to certain roots for which no multiple is a root and the equality of the dimensions of all root spaces corresponding to roots which are half of another root. In the second section, we show that the first dimension result implies that the cone of the corresponding Siegel domain is self-dual while the two dimension results together imply that the domain is quasi-symmetric. The proof is then finished because [3] a quasi-symmetric domain with nonpositive sectional curvature in the Bergman metric is known to be symmetric. We remark that some of the subsidiary results mentioned above are not essentially new. However, for consistency of presentation, it is necessary to have them in the language of normal jalgebras and for that reason they are included. There is also overlap here with work of Zelow (Lundquist) [21]. 1. Throughout this paper, (s, j) will denote a normal j-algebra with admissible form w). This means that s is a finite dimensional real split solvable Lie algebra with almost complex structure j: s -) s such that [X, Y] + j[jX, Y] + j[X, jY] = [jX, jY] and w is a linear form on s such that the bilinear form = o][jX, Y] is symmetric, positive-definite, and j-invariant (note that the assumption that s is real split was incorrectly omitted in [2]). Let n = [s, s] and let a be the orthogonal complement of n in s. By the basic structure theorem of Pyatetskii-Shapiro [13], a is a commutative Received by the editors July 13, 1981. 1980 Mathematics Subiect Classification. Primary 32MI0, 53C30. ' 1983 American Mathematical Society ? 0002-9947/82/000-0759/$09.00