Correspondence of modular forms to cycles associated to O(p,q) (Q1076713)

The study of certain liftings of automorphic forms and their relation to the dual forms of certain totally geodesic cycles in locally symmetric spaces has been the main theme of a series of papers of S. Kudla and J. Millson, resp. Y. L. Tong and the author. On one side the theory of the Weil representation and theta functions for dual reductive pairs is used to construct liftings of automorphic forms. On the other hand, these have a geometric interpretation in terms of harmonic forms arising as the Poincaré duals of totally geodesic cycles for arithmetic quotients attached to arithmetic subgroups of the given groups. The paper under review deals with the case of the dual reductive pairs (O(p,q), Sp(2r,\({\mathbb{R}}))\) with \(r\leq p/2\). As in \textit{Y. L. Tong} and the author [Duke Math. J. 52, 649-688 (1985; Zbl 0582.10018)], the method is to pair directly a theta function and a harmonic cusp form to obtain a cusp form with period integrals as Fourier coefficients. This is in the spirit of Shintani's construction of modular forms [cf. \textit{T. Shintani}, Nagoya Math. J. 58, 83-126 (1975; Zbl 0316.10016)]. The proof of the main result involves various intermediate results which will not be summarized here but are of independent methodological interest, also for the general case. The technical tool for computation is the representation theory of O(p). For \(q=1\) and a compact arithmetic quotient the result coincides with the one given by \textit{S. S. Kudla} and \textit{J. J. Millson} [Compos. Math. 45, 207-271 (1982; Zbl 0495.10016)]. There is the unified approach to the cases O(p,q), U(p,q) or Sp(p,q) by the same authors [The theta correspondence and harmonic forms. I, Math. Ann. 274, 353-378 (1986; see the preceding review); resp. II (preprint)].