In this paper I highlight the central role of an oft-neglected source language modelin the presence of discontinuous negation, also known as NEG2 (e.g. yo no sé cuáles no ‘I don’t know which one it is’), in the Spanish of Chocó, Colombia. Namely,I point to the presence of analogous negation patterns in Ewe, a widely-spokenvariety of Gbe across modern-day Ghana, Togo, and Benin. Data from theTrans-Atlantic Slave Trade Database as well as a 1759 census of the enslaved African(-descendant) population in Chocó mining camps strongly suggest that Gbe languages were spoken by one-third of enslaved Africans taken to the Pacific lowlandregion of Colombia via the Caribbean port of Cartagena in the 17th and 18th centuries. Traditionally, non-canonical patterns of negation in Ibero-Romance contactvarieties involving discontinuous (i.e. pre-verbal and utterance-final) and strictlyutterance-final negator morphemes have been attributed to Bantu languages suchas Kikongo, some varieties of which feature similar structures. However, for theyears in which enslaved Africans were being trafficked directly from West Africa toChocó via Cartagena (1650–1800), archival sources indicate that those from Bantu-speaking regions of West Central Africa comprised close to one-sixth of the totalAfrican-born population, and thus just around half as many as the Gbe-speakinggroup. Together these findings lead to the conclusion that speakers of Gbe (especially Ewe) played a foundational role in the development of non-canonical negation patterns in Spanish varieties spoken in Chocó, Colombia.