The paper is about mathematical objects called "Hida families," which are ways of organizing certain number-theoretic data into continuous p-adic families, where p is a prime number. Specifically, the paper deals with cohomology classes attached to spaces called locally symmetric spaces, which arise naturally in the study of automorphic forms. A central goal in this area is to understand how algebraic and geometric structures in these families behave coherently, and a key tool for doing this is a bilinear pairing, which is essentially a way of multiplying two objects together to get a number (or something like one) that captures important structural information.
The authors construct such a bilinear pairing for p-adic families of cohomology classes at a specific level structure near the prime p. A core technical challenge they overcome is that a naive pairing would only work between objects of complementary types, but the objects of greatest interest, called "anti-ordinary" classes, sit on the same side and cannot be paired directly with each other without extra work. The authors resolve this by finding a symmetry of the underlying geometric spaces that swaps the ordinary and anti-ordinary parts of the cohomology. This symmetry generalizes a classical construction called the Atkin-Lehner involution, well known from the theory of modular forms, to a much broader setting.
They then verify that their construction is compatible with specialization, meaning that when you pin down specific values of the p-adic parameter, the family pairing they build matches the natural duality pairings that already exist on the individual cohomology spaces at those specialized points. As a consistency check and grounding for the work, their construction recovers a pairing previously built by Ohta in the classical case of modular curves, situating the new results as a genuine generalization of that earlier framework to higher-dimensional and more general automorphic settings.