On modular forms of characteristic \(p>0\) (Q1590265)

The classical theory of elliptic modular forms has been reformulated in terms of automorphic forms and representations by Langlands, Deligne, and others. Modular forms thereby correspond bijectively to automorphic forms, i.e., newvectors in certain representation spaces. The situation becomes less transparent in replacing elliptic modular forms through Drinfeld modular forms, certain rigid analytic functions living in positive characteristic \(p\). Whereas the right substitute of automorphic forms is obvious (Jacquet-Langlands in their monumental work developed their theory for both global number and function fields), there is a loss of information in the correspondence between Drinfeld modular forms and their automorphic forms. The situation had been clarified in [\textit{E.-U. Gekeler} and the author, J. Reine Angew. Math. 476, 27-93 (1996; Zbl 0848.11029)], in so far as Drinfeld modular forms of weight two were concerned. These, or more precisely, their ``logarithmic integrals'' (which are characteristic-zero-valued), may entirely be described through automorphic data. It is the aim of the article under consideration to remedy this defect, and to establish a similar correspondence for modular forms of arbitrary weight. The author's result Thm. 3.7 proposes such a correspondence between modular forms (of characteristic \(p\)) and some sort of automorphic forms (of characteristic 0). We feel unable to give here more details about these important results, since both the definitions and the precise statements are too technical.