De Luna makes a case for treating single words as historical evidence, arguing that the intellectual life of enslaved Africans can be reached through the sounds and etymologies of the terms they used. Since the colonial archive preserves so little of what captive people thought, she borrows the comparative linguistics that Africanist historians developed for oral societies.

Four episodes from eighteenth-century Saint-Domingue carry the argument: the contested name of the healer and poisoner Makandal; the Kongo vocabulary the planter Baudry des Lozières compiled from newly arrived captives, read for what they chose to teach him about hearing and being heard; the term mundélé, built from a word for cloth, by which West-Central Africans classified Europeans; and the overlap in sound of loi, lwa and roi.

Long histories of multilingualism and convergence among related African languages, she argues, gave enslaved people terms pliant enough to debate in across language boundaries.