Abstract

‘Methodological approaches to the ‘classification’ of African languages: ‘good science vs. ‘bad science’

There are ca. 2,000 distinct languages in sub-Saharan Africa, one third of the world’s total (Ethnologue, 2013), and multilingualism is extensive. Ex-colonial languages are still in use, in addition to important pidgins/creoles based on European and major African languages, and lingua francas, e.g., Nigerian Pidgin (< English), Fanagalo (< Zulu). The ancestors of modern humans started to migrate out of Africa 100,000+ years ago, so language as a unique human innovation must have been used there longer than anywhere else. The massive time-depth involved correlates with extensive linguistic diversity, and has meant that the phylogenetic classification of the languages has been complex and controversial. The standard reference point remains Greenberg’s (1963) four families, a model which replaced earlier hypotheses sometimes based on irrelevant non-linguistic criteria (e.g., association with the pernicious “Hamitic Hypothesis”). When comparing languages, it is methodologically important to distinguish chance look-alikes, borrowings, and cognate items. Most African languages remain undescribed/ underdescribed, with many endangered or already extinct, under pressure from larger expansionist lingua francas, e.g., Hausa (west), and Swahili (east/central). Typical features are: noun-class systems in Niger-Congo (esp. Bantu), e.g., ki-Swahili; unit phonemes /mb/, /nd/, and /kp/, /gb/; lexical/grammatical tone is (near) universal; “clicks” in Khoisan (unique); “emphatic” consonants in Afroasiatic, e.g., ejectives /t’/ in Amharic, /ts/ in Hausa; ideophones, e.g., (Hausa) buguzumzum emphasizes a fat ungainly person; “serial verbs”, e.g., (Yoruba) mo gba omo naa gbo [I take child the hear] = ‘I believed the child’.