This paper argues for the existence of two different mechanisms in the grammar that can give rise to resumption in syntactic movement: some resumption patterns result from prominence-related pronunciation requirements, while others result from cliticization. Empirical motivation for this split comes from extraction patterns in Atchan (Kwa, Côte d’Ivoire), where PP subextraction and subject extraction behave differently: all PP subextraction requires resumption, subject extraction requires resumption only when a pronoun is extracted. I argue that the first pattern is best captured via a pronunciation requirement (and cannot be captured by cliticization accounts), while the reverse holds for the subject extraction pattern. I thus argue that these two resumptive mechanisms are not incompatible, and that they have different empirical coverage.