Gene duplications have been recognized as an important source of evolutionary innovation and adaptation since at least Haldane, and their varying fates may partly explain the vast disparity in observed genome sizes. The expected fates of most gene duplications involve primarily non-adaptive substitutions leading to either non-functionalization of one duplicate copy or subfunctionalization, neither of which yields novel function. A significant evolutionary problem is thus elucidating the mechanisms of adaptive evolutionary change leading to evolutionary novelty. Currently, the most widely recognized adaptive process involving gene duplication is neo-functionalization (NEO-F), in which one copy undergoes directional selection to perform a novel function after duplication. An alternative, but understudied, adaptive fate that has been proposed is escape from adaptive conflict (EAC), in which a single-copy gene is selected to perform a novel function while maintaining its ancestral function. This gene is constrained from improving either novel or ancestral function because of detrimental pleiotropic effects on the other function. After duplication, one copy is free to improve novel function, whereas the other is selected to improve ancestral function. Here we first present two criteria that can be used to distinguish NEO-F from EAC. Using both tests for positive selection and assays of enzyme function, we then demonstrate that adaptive evolutionary change in a duplicated gene of the anthocyanin biosynthetic pathway in morning glories (Ipomoea) is best interpreted as EAC. Finally, we argue that this phenomenon likely occurs more often than has been previously believed and may thus represent an important mechanism in generating evolutionary novelty.