A New Way of Looking at Distributional Estimates; Applications for the Bilinear Hilbert Transform

Distributional estimates for the Carleson operator acting on characteristic functions of measurable sets of finite measure were obtained by Hunt [12]. In this article we describe a simple method that yields such estimates for general operators acting on one or more functions. As an application we discuss how distributional estimates are obtained for the linear and bilinear Hilbert transform. These distributional estimates show that the square root of the bilinear Hilbert transform is exponentially integrable over compact sets. They also provide restricted type endpoint results on products of Lebesgue spaces where one exponent is 1 or the sum of the reciprocal of the exponents is 3/2. The proof of the distributional estimates for the bilinear Hilbert transform rely on an improved energy estimate for characteristic functions with respect to sets of tiles from which appropriate exceptional subsets have been removed.