Inflation predicts B-mode polarization with correlations that span superhorizon scales at recombination. In contrast, the correlations set up by causal sources, such as phase transitions or defects, necessarily vanish on superhorizon scales. Motivated by BICEP2's B-mode detection, we consider the prospects for measuring the inflationary superhorizon signature in future observations. We explain that the finite resolution of an experiment and the filtering of the raw data induces a transfer of spurious subhorizon power to superhorizon scales, and describe ways to correct for it. We also provide a detailed treatment of possible sources of noise in the measurement. Finally, we present forecasts for the detectability of the signal with future CMB polarization experiments.