Polarisations réelles et représentations associées d'un groupe de Lie résolube. (Real polarizations and associated representations of a solvable Lie group.) (Q1061246)

The basic concept in the representation theory of solvable Lie groups is the notion of polarisations. For a survey up to 1972 see \textit{J. A. Wolf} and \textit{C. C. Moore} [Proc. Symp. Pure Math. 126, 229-243 (1973; Zbl 0302.43014)]. Let \({\mathfrak g}\) be a real solvable Lie algebra, \({\mathfrak g}^*\) its dual space and \({\mathfrak g}_{{\mathbb{C}}}\) its complexification. Any \(f\in {\mathfrak g}^*\) defines naturally a symplectic form \(B_ f\) on \({\mathfrak g}_{{\mathbb{C}}}\). A subalgebra \({\mathfrak h}\) in \({\mathfrak g}_{{\mathbb{C}}}\) is called a polarisation in f if \({\mathfrak h}\) is a maximal isotropic subspace with respect to \(B_ f\) (i.e. f induces a homomorphism on \({\mathfrak h})\) and \({\mathfrak h}+\bar {\mathfrak h}\) is a subalgebra of \({\mathfrak g}_{{\mathbb{C}}}\). \({\mathfrak h}\) is called real if \({\mathfrak h}=\bar {\mathfrak h}\) (hence \({\mathfrak h}+\bar {\mathfrak h}={\mathfrak h})\) and positive, if \(im B_ f(X,\bar X)=im f([X,\bar X])\geq 0\) for all \(X\in {\mathfrak g}_{{\mathbb{C}}}\). We denote by M(f,\({\mathfrak g})\), resp. \(P^+(f,{\mathfrak g})\) the set of all real, resp. positive polarisations of \({\mathfrak g}.\) Now let G be a Lie group with algebra \({\mathfrak g}\). Then G acts on \({\mathfrak g}\) and \({\mathfrak g}^*\) and the stabilizer \(G_ f\) of f is a closed, but not necessarily connected subgroup. If \({\mathfrak g}_ f\) is the stabilizer of f in G then the connected component \(G^ 0_ f\) of \(G_ f\) in G is the analytic subgroup of G corresponding to \({\mathfrak g}_ f\) and is contained in the analytic subgroup \(D^ 0\), corresponding to the subalgebra \({\mathfrak d}={\mathfrak h}\cap {\mathfrak g}\) of \({\mathfrak g}\). As \(f([{\mathfrak d},{\mathfrak d}])=0\) there exists a unique unitary character \(\psi\) on \(D^ 0\) with \(d\psi =im f\), however, in general it is not possible to ''extend'' \(\psi\) to a representation of \(G_ f\). Thus the author makes the following basic assumptions: There exists a polarisation \({\mathfrak h}\in M(f,{\mathfrak g})\cup P^+(f,{\mathfrak g})\) which is invariant under \(G_ f\) and satisfies Pukanszky's condition. Moreover there exists a unitary representation \(\eta_ f\) of \(G_ f\) with \(d\eta_ f\) a multiple of the restriction of im f on \({\mathfrak g}_ f\). Then \(\eta_ f\) extends to a unitary representation \(\chi_ f\) of the closed subgroup \(D=G_ fD^ 0.\) The main result is the following: For a real polarisation \({\mathfrak h}\) and under the above assumptions the induced representation \(T=T(f,\eta_ f,{\mathfrak h},G)=ind^ G_ D \chi_ f\) is independent of \({\mathfrak h}\) and the intertwining algebra \((=commu\tan t)\) of T is isomorphic with those of \(\eta_ f\). For connected (hence solvable) G and positive, strongly admissible polarisation \({\mathfrak h}'\in P^+(f,{\mathfrak g})\) one can also form the holomorphically induced representation \(\rho (f,\eta_ f,{\mathfrak h}',G)=\rho.\) Theorem 1 states, that also \(\rho\) and T are equivalent.