Conjugates of equivariant holomorphic maps of symmetric domains (Q1175129)

Let \(G\) [resp. \(G'\)] be a semisimple algebraic group over \(\mathbb{Q}\) of Hermitian type with associated symmetric domain \(D\) [resp. \(D'\)]. An equivariant pair \((\rho,\tau)\) consists of a homomorphism \(\rho:G \to G'\) and an equivariant holomorphic map \(\tau:D \to D'\). Let \(X=\Gamma\backslash D\) and \(X'=\Gamma' \backslash D'\) be arithmetic varieties, where \(\Gamma\) [resp. \(\Gamma'\)] is an arithmetic subgroup of \(G\) [resp. \(G'\)]. Suppose \(\rho(\Gamma)\subset \Gamma'\). Then \(\tau\) induces a map of algebraic varieties \(\varphi:X \to X'\). Let \(\sigma \in \Aut(\mathbb{C})\). By a theorem of \textit{D. Kazhdan} [Lie groups Represent., Proc. Summer Sch. Bolyai Janos math. Soc., Budapest, 1971, 151-217 (1975; Zbl 0308.14007) and Isr. J. Math. 44, No. 2, 139-159 (1983; Zbl 0543.14030)], \(X^ \sigma\) and \(X'{}^ \sigma\) are arithmetic varieties, so \(D^ \sigma\) and \(D'{}^\sigma\) can be defined in a natural way. Lifting \(\varphi^ \sigma:X^ \sigma \to X'{}^\sigma\) to the universal covering spaces gives the conjugate of \(\tau\) by \(\sigma\), viz. \(\tau^ \sigma:D^ \sigma \to D'{}^ \sigma\). The main theorem of this paper is that there exists a homomorphism \(\rho^ \sigma\) such that \((\rho^ \sigma,\tau^ \sigma)\) is an equivariant pair. For a detailed discussion of equivariant holomorphic maps see \textit{I. Satake}'s `Algebraic structures of symmetric domains', Princeton Univ. Press, Princeton, NJ (1980; Zbl 0483.32017). Note, however, that Satake always assumes that the equivariant homomorphisms commute with the Cartan involutions; this is not assumed in this paper. When \(G'\) is a symplectic group, \(X'\) is a moduli space for Abelian varieties; pulling back the universal family of Abelian varieties over \(X'\) to \(X\) via \(\varphi\) gives a family of Abelian varieties over \(X\), called a Kuga fiber variety. As an application of his main theorem, the author shows that any conjugate of a Kuga fiber variety is again a Kuga fiber variety. The assumption on page 140 that \(X\) is compact is unnecessary; it is not used in the proof.