arXiv · 2611.22687
Spectral recurrences and maximal analyticity: cosmological correlators at one and two loops
Abstract
We develop a method for computing de Sitter boundary correlators based on the maximal analyticity of the two-point function and differential--difference equations satisfied by Schwinger--Keldysh diagrams. The vertex integrations are analytically continued through the relevant de Sitter tuboids onto purely imaginary Lobachevskij cycles. Taking the Mehler--Fock transform of the invariant bubble kernels on these Riemannian cycles converts their differential--difference equations into algebraic recurrences in the spectral parameter. The mixed-branch kernel obeys a homogeneous recurrence, whereas the chronological and antichronological kernels satisfy sourced recurrences whose local terms are determined by dimensionally continued tadpoles. We solve these recurrences, constructing explicit particular solutions in terms of generalised hypergeometric functions and identifying the analytic conditions selecting their physical completion. Attachment to the external boundary points then yields an intrinsic spectral representation and a complete solution of the boundary in--in diagrams considered here. Our method is fully coordinate-independent. Specialising to a Poincaré chart readily gives expressions that can be compared directly with those obtained by the standard coordinate-dependent approach. Composing two or more bubbles yields a multi-loop chain whose spectral multiplier is determined by the one-loop building blocks and the sum over intermediate Schwinger--Keldysh branches.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Bruce A. Bassett, Sergio L. Cacciatori, Ugo Moschella. 2026-11-17. Spectral recurrences and maximal analyticity: cosmological correlators at one and two loops. https://arxiv.science/abs/2611.22687
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite arXiv Science by Indiciae: BibTeX · RIS.