A cohomological interpretation of Archimedean zeta integrals for \(\mathrm{GL}_3\times\mathrm{GL}_2\) (Q2065743)

The study of special values of \(L\)-functions connects arithmetic, analysis and geometry, in terms of the nonvanishing properties, subconvexity bounds, algebraicity properties and many other fascinating features. A fruitful approach is to interpret these special values as certain cohomology classes, and along this line we have the period conjecture of \textit{P. Deligne} [Proc. Symp. Pure Math. 33, 313--346 (1979; Zbl 0449.10022)] for Hasse-Weil \(L\)-functions and the generalization of \textit{L. Clozel} [Perspect. Math. 10, 77--159 (1990; Zbl 0705.11029)] to automorphic \(L\)-functions for general linear groups. The paper under review focuses on the algebraicity problem for Rankin-Selberg \(L\)-functions for \(\mathrm{GL}_3 \times \mathrm{GL}_2\). \textit{A. Raghuram} [Forum Math. 28, No. 3, 457--489 (2016; Zbl 1417.11082)] showed that, up to an explicit factor of algebraic numbers, the special value of a Rankin-Selberg \(L\)-function for \(\mathrm{GL}_{n+1} \times \mathrm{GL}_n\) agrees with the product of its Whittaker period (as introduced by \textit{A. Raghuram} and \textit{F. Shahidi} [Int. Math. Res. Not. 2008, Article ID rnn077, 23 p. (2008; Zbl 1170.11009)]) and the inverse of certain archimedean zeta integrals. In the paper under review, the authors considered the special case of \(n = 2\), and applied the generalized modular symbol method adopted by \textit{J. Mahnkopf} [J. Inst. Math. Jussieu 4, No. 4, 553--637 (2005; Zbl 1086.14019)]. Based upon the studies of \textit{M. Hirano} et al. [Proc. Japan Acad., Ser. A 92, No. 2, 27--32 (2016; Zbl 1417.11085)] and with an effective form of the Eichler-Shimura map for \(\mathrm{GL}_3\), the authors showed that these archimedean zeta integrals coincide with multiples of the desired \(\Gamma\)-factors. More precisely, let \(\pi^{(2)}\) and \(\pi^{(3)}\) be cohomological irreducible cuspidal automorphic representations of \(\mathrm{GL}_2(\mathbb{Q}_{\mathbb{A}})\) and \(\mathrm{GL}_3(\mathbb{Q}_{\mathbb{A}})\) respectively, then the main result of this paper is that, under an appropriate sign condition, we have \[ I^{\pm}(m, \pi^{(3)}, \pi^{(2)}) = 2^{-2} (-1)^{\delta} \sqrt{-1}^{\frac{l_3}{2}-m-1} \binom{\frac{l_3}{2} - 1}{l_3 - 2 - m} \binom{\frac{l_3}{2} - 1}{m - l_2 + 1} \frac{L(m + \frac{1}{2}, \pi^{(3)} \times \pi^{(2)})}{\Omega_{\pi^{(3)}} \ \Omega^{\pm}_{\pi^{(2)}}} \] where \(I^{\pm}(m, \pi^{(3)}, \pi^{(2)})\) is the cup products of certain cohomological classes associated to \(\pi^{(3)}\) and \(\pi^{(2)}\), and \(\Omega_{\pi^{(3)}}\) and \(\Omega^{\pm}_{\pi^{(2)}}\) are their Whittaker periods. Furthermore, the ratio \(L(m + 1 / 2, \pi^{(3)} \times \pi^{(2)}) / \Omega_{\pi^{(3)}} \Omega^{\pm}_{\pi^{(2)}}\) is algebraic. The authors also pointed out that, besides its role in the algebraicity problem of special values, the explicit study of these archimedean zeta integrals is important for the theory of \(p\)-adic Rankin Selberg \(L\)-functions as introduced by \textit{F. Januszewski} [``Non-abelian $p$-adic Rankin-Selberg $L$-functions and non-vanishing of central $L$-values, Preprint, \url{arXiv:1708.02616}].