Measurements with different techniques point to a strong affinity of hydroxide ions for interfaces between water and hydrophobes, but some spectroscopic experiments do not detect excess hydroxide at the interface, while others do. Hydroxide ions are unusual in that they reduce the relative permittivity of an electrolyte solution more than other monovalent, monatomic ions. This implies that they suppress the collective dipole-moment fluctuations of nearby waters. We show that the absence of these fluctuations leads to a Hamaker-like force on the hydroxide ion that attracts it to regions where dipole-moment fluctuations are smaller than in bulk water, in other words, to regions of low relative permittivity. We show also that there is no contradiction between the picture of the basic, negatively charged interface and spectroscopic measurements. This is, in part, because the hydroxides are mostly below the outermost molecular layers. By combining a simple model for this fluctuation force with a modified Poisson-Boltzmann equation, we reproduce the dependence of the zeta-potential on pH, including the low isoelectric point, the approximate magnitude of the experimental surface charge density, and the Jones-Ray data for the dependence of surface tension on electrolyte concentration. We discuss also the apparent contradiction between molecular-dynamics simulations that deny and experiments that support a basic, negatively charged interface.