Spaces of Infinite Measure and Pointwise Convergence of the Bilinear Hilbert and Ergodic Averages Defined by L-Isometries

We generalize the respective “double recurrence” results of Bourgain and of the second author, which established for pairs of L∞ functions on a finite measure space the a.e. convergence of the discrete bilinear ergodic averages and of the discrete bilinear Hilbert averages defined by invertible measure-preserving point transformations. Our generalizations are set in the context of arbitrary sigma-finite measure spaces and take the form of a.e. convergence of such discrete averages, as well as of their continuous variable counterparts, when these averages are defined by Lebesgue space isometries and act on Lp1 × Lp2 (1 < p1, p2 < ∞, p −1 1 + p 2 < 3/2). In the setting of an arbitrary measure space, this yields the a.e. convergence of these discrete bilinear averages when they act on Lp1 × Lp2 and are defined by an invertible measure-preserving point transformation.