Ethnographies of Latin American and Caribbean gangs tend to emphasize the violent practices of these social groups. My own research in Kingston, Jamaica, on criminal leaders known as 'dons', analyzed their entanglement with the Jamaican state and studied how their relations with inner-city residents resemble citizenship relations. While dons are undoubtedly directly associated with Jamaica's high homicide levels, they also provide social welfare and public order, and my work has focused principally on this relatively non-violent side of donmanship. In this article, I reflect on the factors underlying this particular representational emphasis, which has sometimes led to accusations of my 'romanticizing the dons'. Assessing these various factors, the article considers whether it is ethically justifiable and epistemologically sound to produce ethnographies of violent actors without dwelling in detail on their acts of violence. Given the skew in studies of criminal organizations, and of inner-city life, toward extensive description of and analytical attention to violence, I propose that producing ethnographic narratives that decenter it is in fact justifiable. In addition, studying criminal organizations through less common frameworks can open new avenues of exploration that may be obscured if we continue to satisfy popular and academic expectations of the centrality of violence.