The method of Hadamard and de la Vallée-Poussin (According to Pierre Deligne) (Q2266746)

In ''La conjecture de Weil. II'' [Publ. Math., Inst. Hautes Étud. Sci. 52, 137-252 (1980; Zbl 0456.14014)], \textit{P. Deligne} proves a fairly general fact regarding the non-vanishing of a large class of L-series on their line of convergence. Roughly speaking, he works with a locally compact group G which is an extension of a group \(\Gamma\) (either \({\mathbb{Z}}\) or \({\mathbb{R}})\) by a compact group, and an infinite collection \(\Sigma\) of conjugacy classes of G. A real quasi-character of \(\Gamma\) gives the notion of a norm of an element of \(\Sigma\), plus a family of one dimensional representations \(\omega_ s\) of G, indexed by the complex numbers. The author then looks at the Riemann surface of isomorphism classes of irreducible ''quasi-unitary'' representations of G (those of the form \(\omega_ s\) times a unitary representation). This space is the disjoint union of complex planes and possibly tubes, and comes equipped with a natural notion of ''real part''. He assumes that the ''Euler product'' corresponding to the norms of the elements of \(\Sigma\) converges for \(Re(s)>1\), which ensures the absolute convergence of the L- series associated to every quasi-unitary representation with real part strictly greater than 1. (The L-series is function on the space of representations, and is a natural infinite product of determinants taken over \(\Sigma\).) The theorem is that if the L-series continues meromorphically to \(Re(\tau)\geq 1\) and is holomorphic there except for possibly a simple pole at the original quasi-character, then the L series doesn't vanish on the line of representations with real part equal to 1, except for possibly one exceptional one dimensional order two representation. - Deligne's proof of this result is quite elegant (and goes back to an idea of Mertens), and essentially is a result about integer valued functions on representations of a compact group (here: orders of poles of the L-functions). It is extremely worthwhile in this context to look at the book by \textit{J.-P. Serre}, ''Abelian \(\ell\)-adic representations and elliptic curves'' (1968; Zbl 0186.257) to appreciate the origins and significance of some of Deligne's ideas, and to obtain an elementary calculus view of the Tate-Sato conjecture. The author's paper, to my mind, doesn't help very much in understanding the above. In fact, I soon forgot about it and went to Deligne and Serre. However, it does have the virtue of publicizing Deligne's theorem, which, as the author says, should be brought to the attention of number theorists.