In 1998 I conducted a series of interviews with Zimbabweans who recounted, often using English, their memories of Albert Sumbo-Ncube. From these I have selected and transcribed five interviews with ZIPRA ex-combatants in which they tell the story, as they remember and elaborate on their memories, of Sumbo's escape from Rhodesian police custody at the Victoria Falls in 1977 during the Zimbabwean liberation struggle. The interviews represent Sumbo as a hero and reveal the folk hero creation process at work. This hero figure was created by people who needed an effective figure of oppositional propaganda and who did not have access to the technology and resources of the Rhodesian government. Their narratives were communicated orally and they fused material found in the Rhodesian government-controlled newspapers with an amaNdebele oral tradition. I shall draw on Hobsbawm's (1972) notions of the social bandit and Robens's (1989) study of the folk hero creation amongst post-slavery African-Americans in order to understand the ZIPRA guerrillas' hero creation. The Sumbo folk hero creation served to promote an ideal self for the Zimbabwean guerrillas and their recruits. Sumbo's daring and his ability successfully to defy authority evoked admiration amongst the guerrillas in the 1970s, and in 1998 revives for them the idealism of the struggle. In Zimbabwe the 'hero' has become a contested category, because of the government's will to control the historical representation of the liberation struggle by promoting an official history with official categories of heroes. Working with Barber's notion of popular African arts (1987 and 1997), I argue that a folk hero can be redefined as a 'popular hero' when created by a proletariat and expressed by means of a popular art form. The interviewees use a specific form, the oral historical narrative, to preserve and transmit the Sumbo hero figure. I argue that though this oral historical narrative is less fixed in form and occasion than praise poetry, songs and genealogies, it nevertheless possesses identifiable and recurrent characteristics and I have established a number of criteria for identifying oral historical narrative as a genre. In order to avoid taking a generalised and essentialising approach to the notion of 'African culture', I have drawn on theory that is as specific as possible to the understanding of oral historical narratives within the context of siNdebele speakers in Zimbabwe. I have drawn on research published by Hofmeyr (1993) and Scheub (1975) because they focused on Nguni-speaking societies. Their research is further supported by my own research conducted in the rural area of Tsholotsho in Zimbabwe. The analysis of the oral historical narrative genre used by the interviewees demonstrates that significant formal and performance skills occur in this type of narrative which takes place within apparently informal conversations. / Thesis (M.A.)-University of Natal, Durban, 2000.