Mean-periodicity and zeta functions (Q1931237)

The conjectural analytic properties of Hasse-Weil \(L\)-functions of curves over number fields are a consequence of the widely believed expectation that such functions arise from the world of automorphic representations. This paper deals with zeta functions of arithmetic schemes, which are quotients of Hasse-Weil \(L\)-functions up to finitely many bad factors. Such quotients do not inherit automorphic properties and this work attempts to replace the lost automorphicity with a mean-periodicity condition. On the other hand, it is stated as an expectation that mean-periodicity can be proved independently of automorphicity. This claim was motivated by the appearance of mean-periodicity in the first author's works on two-dimensional adelic analysis [Doc. Math., J. DMV Extra Vol., 261--284 (2003; Zbl 1130.11335); Mosc. Math. J. 8, No. 2, 273--317 (2008; Zbl 1158.14023); J. K-Theory 5, No. 3, 437--557 (2010; Zbl 1225.14019)], though the document in question does not require knowledge of this as a prerequisite. Further evidence is discussed at intervals throughout the paper. The need for approaches to meromorphic continuation and functional equation which do not depend on automorphicity was mentioned by \textit{R. P. Langlands} [in: Representation theory and automorphic forms. Providence, RI: American Mathematical Society. Proc. Symp. Pure Math. 61, 457--471 (1997; Zbl 0901.11032)]. As remarked in the introduction, mean-periodicity is not widely used in number theory and thus deserves some further discussion -- much care is taken to introduce three formulations of the notion, each of which is valid in only certain classes of functional spaces, with large overlaps between these classes. The first states that the set of translates of a function is not dense in the functional space, the second is existence of a non-trivial homogeneous convolution equation, and the third is the existence of a form of generalized Fourier expansion. The main results of the paper state that, up to technical conditions, the meromorphic continuation and functional equation of zeta functions of arithmetic schemes is equivalent to the mean-periodicity of a related function, referred to in this paper as the ``boundary function''. The boundary function is given as a transform of the product of finitely many Riemann zeta functions and the zeta function of the arithmetic scheme. It is so called on account of its interpretation as an integral over a semi-global adelic boundary in dimension two -- this idea is not needed for understanding of the paper. Aside from meromorphic continuation and functional equation this paper also considers zeros and poles of zeta functions. In particular, it is studied a ``single sign property'', which is again motivated by the ideas originating in the first author's adelic analysis. The paper concludes with interesting examples of mean-periodic functions arising from Dedekind zeta functions, cuspidal automorphic forms and Eisenstein series.