Simulations of the moon-forming impact suggest that most of the lunar material derives from the impactor rather than the Earth. Measurements of lunar samples, however, reveal an oxygen isotope composition that is indistinguishable from terrestrial samples, and clearly distinct from meteorites coming from Mars and Vesta. Here we explore the possibility that the silicate Earth and impactor were compositionally distinct with respect to oxygen isotopes, and that the terrestrial magma ocean and lunar-forming material underwent turbulent mixing and equilibration in the energetic aftermath of the giant impact. This mixing may arise in the molten disk epoch between the impact and lunar accretion, lasting perhaps 10 2-10 3 yr. The implications of this idea for the geochemistry of the Moon, the origin of water on Earth, and constraints on the giant impact are discussed.