Some aspects in the theory of representations of discrete groups. II (Q1813895): Difference between revisions

Here we consider mainly equivalence relations among irreducible unitary representations (IURs) of an infinite wreath product group, constructed in the first part [Jap. J. Math., New Ser. 16, 197-268 (1990; Zbl 0729.22007)] of these notes. We keep to the notations of part I. 1. Commutativity of two kinds of inducing processes. Let \(T\) be a group and \(S\) its subgroup. Consider wreath product groups \({\mathfrak S}_A(S)\) and \({\mathfrak S}_A(T)\). Then we have two kinds of inducing representations: the usual one an the WP-inducing. We give a certain commutativity of these inducing processes. Start with a datum \(R = \{A, \rho_s, \chi,\;a = (a_\alpha)_{\alpha \in A}\}\) for an elementary representation of \(\rho(R)\) of \({\mathfrak S}_A(S)\). On the one hand, put \(\widetilde {\rho}_T = \text{Ind}^T_S \rho_S\), and let \(\widetilde {a}_\alpha = \text{Ind}^T_S a_\alpha \in V(\widetilde {\rho}_T)\) be the induced vector of \(a_\alpha \in V(\rho_S)\). Then \(\widetilde {a} = (\widetilde {a}_\alpha)_{\alpha \in A}\) is a reference vector for \((\widetilde {V}_\alpha)_{\alpha \in A}\) with \(\widetilde {V}_\alpha = V(\widetilde {\rho}_T)\), and denote it as \(\widetilde{a} = \text{Ind}^T_S a\). Thus we get a datum \(\widetilde {R} = \{A, \widetilde {\rho}_T, \chi, \widetilde{a}\}\) for \({\mathfrak S}_A(T)\) and correspondingly an elementary representation \(\rho (\widetilde {R})\) of \({\mathfrak S}_A(T)\). On the other hand, we have the induced representation \(\text{Ind} (\rho(R); {\mathfrak S}_A (S) \uparrow {\mathfrak S}_A(T))\). Theorem 1. Let \(R\) be a datum for an elementary representation of \({\mathfrak S}_A(S)\). Then the two representations \(\rho (\widetilde {R})\) and \(\text{Ind} (\rho(R); {\mathfrak S}_A(S) \uparrow {\mathfrak S}_A(T))\) of \({\mathfrak S}_A (T)\) are canonically equivalent to each other. A similar assertion holds for standard representations for \({\mathfrak S}_A (S)\) and \({\mathfrak S}_A (T)\). -- 2. Equivalence relations among standard representations. Take two induced representations \(\rho (Q_i) = \text{Ind} (\pi(Q_i); H(Q_i) \uparrow {\mathfrak S}_A(T))\), \(i = 1,2\), of \({\mathfrak S}_A (T)\), called standard, and let the corresponding data be \[ Q_1 = \{(A_\gamma, \rho^\gamma_{T_{1\gamma}}, \chi_{1\gamma})_{\gamma \in \Gamma},\;(a_1(\gamma))_{\gamma \in \Gamma},\;(b_{1\gamma})_{\gamma \in \Gamma}\}, \] \[ Q_2 = \{(B_\delta, \rho^\delta_{T_{2\delta}}, \chi_{2\delta})_{\delta \in \Delta},\;(a_2(\delta))_{\delta \in \Delta},\;(b_{2\delta})_{\delta \in \Delta}\}, \] where, in particular, \((A_\gamma)_{\gamma\in \Gamma}\) and \((B_\delta)_{\delta \in \Delta}\) are partitions of \(A\), and \(T_{1\gamma}\) and \(T_{2\delta}\) are subgroups of \(T\). For an element \(\zeta\) of \({\mathfrak S}_A\), we call an adjustment of \(Q_2\) by \(\zeta\) the datum \[ ^\zeta Q_2 = \{(\zeta (B_\delta), \rho^\delta_{T_{2\delta}}, \chi_\delta)_{\delta \in \Delta}, (a_2(\delta))_{\delta\in \Delta},\;(b_{2\delta})_{\delta \in \Delta}\}. \] Then \(\rho(Q_2)\) is equivalent to \(\rho(^\zeta Q_2)\) in a trivial fashion. -- Theorem 2. Assume that two data \(Q_1\) and \(Q_2\) satisfy the condition \((Q1)\), i.e., \(|\Gamma_f|\leq 1\), \(|\Delta_f|\leq 1\), and that both \(\rho(Q_1)\) and \(\rho(Q_2)\) are irreducible. Then they are mutually equivalent if and only if the following conditions hold: (EQU1) Replacing \(Q_2\) by its adjustment by an element in \({\mathfrak S}_A\) if necessary, we have a 1-1 correspondence \(\kappa\) of \(\Gamma\) onto \(\Delta\) such that \(A_\gamma = B_{\kappa (\gamma)}\) for \(\gamma \in \Gamma\). Further \(\chi_\gamma = \chi_{\kappa(\gamma)}\) for \(\gamma \in \Gamma\), and \(\text{Ind}^T_{T_{1\gamma}} \rho^\gamma_{T_{1\gamma}} \cong \text{Ind}^T_{T_{2\delta}} \rho^\delta_{T_{2\delta}}\) for \(\gamma \in \Gamma_f\) and \(\delta= \kappa(\gamma)\). -- (EQU2) For \(\gamma \in \Gamma_\infty = \Gamma \setminus \Gamma_f\), replace \(\delta = \kappa(\gamma)\) by \(\gamma\), and put \(T_{0\gamma} = T_{1\gamma} \cap T_{2\gamma}\). Then, for every \(\gamma \in \Gamma_\infty\), there exist an IUR \(\rho^\gamma_{T_{0\gamma}}\) of \(T_{0\gamma}\) and a reference vector \(a_0(\gamma) = (a_{0\alpha})_{\alpha\in A_\gamma}\), \(a_{0\alpha}\in V(\rho^\gamma_{T_{0\gamma}})\), \(|a_{0\alpha}|= 1\), such that for \(j = 1, 2\), \(\rho^\gamma_{T_{j\gamma}} \cong \text{Ind} (\rho^\gamma_{T_{0\gamma}}; T_{0\gamma} \uparrow T_{j\gamma})\) and \(a_j(\gamma)\) is Moore-equivalent to the induced vector \(\text{Ind}(a_0 (\gamma); T_{0\gamma} \uparrow T_{j\gamma})\) in the extended sense. -- (EQU3) For \(\gamma \in \Gamma_\infty\), put \(\chi_{0\gamma} = \chi_{1\gamma}\) \((= \chi_{2\gamma})\) and \(Q_{j\gamma} = \{A_\gamma, \rho^\gamma_{T_{j\gamma}}, \chi_{j\gamma}, a_j(\gamma)\}\), \(0 \leq j \leq 2\), and consider IURs \(\Pi(Q_{j\gamma})\) of \(H_{j\gamma} = {\mathfrak S}_{A_\gamma} (T_{j\gamma})\). Then there exists a unit vector \(b_{0\gamma} \in V(\Pi(Q_{0\gamma}))\) for every \(\gamma \in \Gamma_\infty\) such that \((b_{j\gamma})_{\gamma \in\Gamma_\infty}\), \(j = 1, 2\), are respectively Moore-equivalent in the extended sense to \((\widetilde {b}_{j\gamma})_{\gamma\in \Gamma_\infty}\) with \(\widetilde {b}_{j\gamma} = \text{Ind}(b_{0\gamma}; H_{0\gamma} \uparrow H_{j\gamma})\), with respect to the representations \(\Pi(Q_{j\gamma})\) and \(\text{Ind}(\Pi(Q_{0\gamma})\); \(H_{0\gamma} \uparrow H_{j\gamma})\). Here note that, under the condition (EQU2), the IUR \(\Pi(Q_{j\gamma})\) is equivalent to the induced one \(\text{Ind} (\Pi(Q_{0\gamma})\); \(H_{0\gamma} \uparrow H_{j\gamma})\) for \(j = 1, 2\), by Theorem 1.