A Wolff theorem for finite rank bounded symmetric domains (Q2401811)

The author extends some of her earlier finite dimensional results to the setting of a finite rank JB\(^*\)-triple \({\mathbf Z}\) with Bergman operators \(B(w,z)\) \((z,w \in {\mathbf B})\). Let \(f : {\mathbf B} \to {\mathbf B}\) be a compact holomorphic fixed-point free map of the open unit ball \({\mathbf B}\) of \({\mathbf Z}\). According to the classical Earl-Hamilton theorem, for each \(0\leq \alpha <1\), the map \(\alpha f\) admits a unique fixed point \(z_\alpha (\in \alpha \overline{\mathbf B})\) and we may choose an accumulation point \(\zeta :=\lim_k z_{\alpha_k} \in \partial {\mathbf B}\) with some sequence \(\alpha_k \uparrow 1\) of the net \(\big( z_\alpha \big)\). The main results can be summarized as follows: {\parindent=6mm \begin{itemize}\item[(1)] The limit operator \(R:= \lim_k (1-\| z_k\|^2) B(z_k,z_k)^{-1}\) is well-defined and bounded. \item[(2)] The sets \({\mathbf E}_\lambda := \{ w\in {\mathbf B}: \| B(w,w)^{-1} B(w,\zeta) R \|< \lambda\}\) \((\lambda>0)\) are non-empty convex \(f\)-invariant domains in \({\mathbf B}\) with \(\zeta\in \partial {\mathbf E}_\lambda\). \item[(3)] We can write \({\mathbf E}_\lambda = c_\lambda + T_\lambda({\mathbf B})\), where \(c_\lambda \in{\mathbf B}\) and \(T_\lambda\in\text{GL}({\mathbf Z})\). \item[(4)] \(c_\lambda\) and \(T_\lambda\) in (3) can be expressed in terms of the middle point tripotent \(e\) of the boundary component of \(\zeta\) and its atomic Peirce decomposition \(e=e_1+\cdots+e_r\) with suitable scalars \(0\leq\gamma_1\leq\cdots\leq \gamma_r\leq 1\) as \[ c_\lambda = \sum_{k=1}^r \gamma_k (\lambda+\gamma_k)^{-1} e_k,\;T_\lambda = B(v_\lambda,v_\lambda), \] where \(v_\lambda = \sum_{k=1}^r [ 1 - (1 + \gamma_k/\lambda)^{-1/2}]^{1/2}\). \item[(5)] We have \(e = c_\lambda - T_\lambda(e) \in\partial{\mathbf E}_\lambda\) for all \(\lambda>0\). \end{itemize}}