The prehistorical, implications of Guthrie's comparative Bantu Part I: Problems of internal relationship

Abstract

After a brief survey of the organization of Guthrie's Comparative Bantu, consideration is devoted to the putative prehistorical interrelationships of the Bantu languages, in the light of Guthrie's and subsequent scholarship. The intended ‘introductory’ nature of Guthrie's work is stressed, his comprehensive survey covering only 28 Test Languages (among over 350). Recent criticisms of and divergences from Guthrie's prehistorical conclusions are considered, together with the inadequacy of the ‘genealogical tree’ approach. The apparent contradiction of Guthrie by Henrici and Heine in particular is shown to be less substantial than at first appears: maps based on the work of all three confirm the remoter relationship of NW. Bantu to the remainder, and also a basic East-West division (but with Luba and Bemba occupying an intermediate position). It seems probable that some NW. languages are to be traced directly to a proto-language in the Cameroun area, rather than to Guthrie's Katangan nucleus, but this means only that Guthrie cast his net too wide, not that his whole thesis is false. It is recommended that the use of the term ‘Proto-Bantu’ be clarified by use of the term ‘Proto-Bantu 1’ for an original NW. nucleus, ‘Proto-Bantu 2’ for a secondary nucleus S. of the Congo forest, and ‘Proto-Bantu 3’ for a tertiary (and overlapping) nucleus to the east of 2. An amended version of Guthrie's sequence of Bantu linguistic expansion is proposed, although attention is drawn to the hazards of necessarily basing this on the evidence of modern Bantu languages only. Henrici's statistical treatment of Guthrie's data assists in distinguishing temporal from spatial distance among the Test Languages.