Corner view on the crown domain (Q1000343)

This paper is about the crown domain \(\Xi\) which is the canonical complexification of a Riemannian symmetric space \(X\) of noncompact type. The author is interested in the nature of the boundary of \(\Xi\) and eventually in good compactifications of \(\Xi\). Let \(G\) be the connected component of the isometry group of \(X\). Then \(X=G/K\) is a maximal compact subgroup for \(K < G\). Let \(\partial \Xi\) be the topological boundary of \(\Xi\) in \(X_\mathbb C\). The boundary is a very complicated object and there is little hope to obtain an explicit description. However, \(\partial \Xi\) features some structure; for instance, in it one finds the distinguished boundary \(\partial_d\Xi\subset\partial \Xi\), introduced in [\textit{S. Gindikin} and \textit{B. Krötz}, Trans. Am. Math. Soc. 354, No.~8, 3299--3327 (2002; Zbl 0999.22019)]. The distinguished boundary is some sort of Shilov boundary of \(\Xi\) in the sense that it is the smallest closed subset in \(\partial \Xi\) on which bounded plurisubharmonic functions on \(cl(\Xi)\) take on their maximum. It is known that \(\partial_d\Xi\) is a finite (and explicit) union of \(G\)-orbits, say \(\partial_d\Xi = \mathcal O_1\coprod\dots\coprod\mathcal O_s\), where each \(\mathcal O_j\) is identified with a homogeneous space: \(G/H_j\). If \(G/H_j\) is a symmetric space, then it is a non-compactly causal symmetric space. Moreover, every non-compactly causal symmetric space \(Y = G/H\) appears in the distinguished boundary of the corresponding crown domain for \(X = G/K\). One aim of this paper is to better understand this result. To be more concise: what is the reason that precisely non-compactly causal symmetric spaces appear in the boundary? Answering this question eventually reveals the structure of \(\partial\Xi\). Assume the Lie algebra of \(G\) to be simple and write \(\mathfrak q\) for the tangent space of \(Y\) at the standard base point \(y_0 = H\in Y\). Note that \(\mathfrak q\) is a linear \(H\)-module. Now, non-compactly causal means that \(\mathfrak q\) admits a non-empty open \(H\)-invariant convex cone, say \(C\), which is hyperbolic and does not contain any affine lines. The theme of this paper is to view \(\Xi\) from the corner point \(y_0\in Y\). For \(C\subset\mathfrak q\) the minimal cone, the author forms in the tangent bundle \(TY= G\times_H\mathfrak q\) the cone-subbundle \(\mathcal C=G\times_HC\) and with that its boundary cone-bundle \(\partial\mathcal C=G\times_H\partial C\). In this context it is reasonable to ask: Is there a \(G\)-equivariant, genetically injective, proper continuous surjection \(p :\partial\mathcal C\to \partial\Xi_H\)? In other words, does there exist an equivariant ``resolution'' of the boundary in terms of the geometrically simple boundary cone-bundle \(\partial\mathcal C\). The author gives an affirmative answer to this question if \(X\) is a Hermitian tube domain. In this simplified situation the crown domain is \(\Xi = X\times\overline X\) with \(\overline X\) denoting \(X\) but endowed with the opposite complex structure (i.e., if \(X\) is already complex, then the crown is the complex double). Moreover, \(\partial_d\Xi = Y\) is connected, i.e., \(\Xi =\Xi_H\).