scientific article; zbMATH DE number 1076118

From MaRDI portal
Publication:4359560





Let \(G_\mathbb{C}\) be a simply connected complex Lie group and \(G\subseteq G_\mathbb{C}\) a real form which is simple and Hermitian. Then the Lie algebra \({\mathfrak g}\) of \(G\) contains a \(G\)-invariant closed convex proper cone \(C\) and the set \(\Gamma(C) =G\exp (iC)\) is a closed subsemigroup of \(G_\mathbb{C}\) with interior \(\Gamma(C)^0 =G\exp (iC^0)\), where \(C^0\) is the interior of \(C\) in \({\mathfrak g}\). The Hardy space \(H^2(\Gamma(C))\) associated to such a semigroup consists of all holomorphic functions on \(\Gamma(C)^0\) for which the supremum \[ \sup_{\gamma\in \Gamma(C)^0} \int_G\bigl|f(\gamma g) \bigr|^2dg \] is finite. This supremum then defines a Hilbert space norm on \(H^2(\Gamma(C))\) and it turns out that \(H^2 (\Gamma(C))\) admits a reproducing kernel which is called the Cauchy-Szegö kernel. Suppose that \(G^\flat\) is another Hermitian group, so that the corresponding Riemannian symmetric space \(G^\flat/K^\flat\) is a complex manifold which can be realized as a bounded symmetric domain or -- via a Cayley transform -- as a generalized Siegel upper half plane. In case the domain is of tube type, the Siegel realization can be chosen to be of the form \(T_\Omega= V+i \Omega\), where \(V\) is a Euclidean Jordan algebra and \(\Omega \subseteq V\) the associated cone of invertible squares. In those cases one has a classical Hardy space \(H^2(T_\Omega)\) consisting of holomorphic functions on \(T_\Omega\) for which the supremum \[ \sup_{y\in \Omega} \int_V\bigl|f(x+iy) \bigr|^2dy \] is finite. Again this supremum then defines a Hilbert space norm and the resulting Hilbert space \(H^2(T_\Omega)\) admits a reproducing kernel which is again called the Cauchy-Szegö kernel. In contrast to the case of \(H^2(\Gamma (C))\) this Cauchy-Szegö kernel is well known. The authors suggest a program to study the Hardy spaces \(H^2(\Gamma(C))\) via embeddings into classical Hardy space of the type \(H^2(T_\Omega)\). The essential tool here is a Cayley transform that maps \(\Gamma(C)^0\) onto a dense open subset of \(T_\Omega\). They treat the following cases of \((G,G^\flat)\) in some detail: \[ \bigl(U(p,q), SU(p+q,p+q)\bigr),\;\bigl(Sp(r,\mathbb{R}), Sp(2r,\mathbb{R}) \bigr),\;\bigl(SO^*(2l), SO^*(4l)\bigr). \] As a result they find that the classical Hardy spaces can be embedded into the Hardy spaces of the double covering semigroup of \(\Gamma(C)\). In the case \((Sp(r,\mathbb{R}), Sp(2r,\mathbb{R}))\) they characterize the image completely. In all three cases \(G^\flat\) acts unitarily on the classical Hardy space and \(G\times G\) can be embedded into \(G^\flat\) such that the intertwining operator coming from the Cayley transform maps the action of \(G\times G\) to the left-right regular representation. Then the proofs rest on a decomposition of the spaces \(H^2(\Gamma (C))\) under this regular representation. It should be noted that the paper contains a useful explicit construction of the double covering semigroups mentioned.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4359560)