¹ Roland, Oliver, ‘The Problem of the Bantu Expansion’, Journal of African History, VII (1966) 363–76. I wish to acknowledge here my indebtedness to Professor Oliver for comments and suggestions regarding an earlier version of this paper.Google Scholar

² Oliver, , op. cit. 367.Google Scholar

³ Oliver, , op. cit. 366.Google Scholar

⁴ Examples from The Languages of Africa, 2nd edition, Mouton, The Hague (1966), include 9–13, discussion of resemblances of the noun class system of the Adamawa Eastern languages with the remainder of Niger-Congo; 31, comparisons of the noun class system of Bantu with the rest of Niger-Congo; 35, grammatical innovations of Bantu; 46–8, Afroasiatic grammatical features found in the Chadic languages; 68–71, grammatical resemblances of Hottentot to the rest of Khoisan; 73–5, grammatical resemblances of Hatsa to the rest of Khoisan; 86–9, common grammatical features of the Nilotic and ‘Nilo-Hamitic’ languages; 509–57, grammatical features of the Eastern Sudanic languages found in the remainder of the Chari-Nile group; 530–3, grammatical evidence for the unity of Nilo-Saharan; 549–52, grammatical resemblances between the Kordofanian and the Niger-Congo families.

⁵ Greenberg, , op. cit. 31.Google Scholar

⁶ Greenberg, , op. cit. 35.Google Scholar

⁷ Greenberg, , op. cit. 53–24.Google Scholar

⁸ Tucker, A. N. and Bryan, M. A., The Non-Bantu Languages of North-Eastern Africa (Oxford University Press, London, New York and Capetown, 1956),Google Scholar

⁹ Oliver, , op. cit. 366.Google Scholar

¹⁰ Alexander, Sperber, A Historical Grammar of Biblical Hebrew (Brill, Leiden, 1966), 54.Google Scholar

¹¹ The phrase ‘indisputably Bantu languages’ is employed here because there is at least one group of languages, the Jarawa, which I consider to be Bantu and which is in the central part of Nigeria, north and west of the area which I posit as the ultimate point of Bantu origin.

¹² Guthrie, M., The Classification of the Bantu Languages (Oxford Press, London and New York, 1948). There are, of course, some differences of detail regarding these zones in the more recent work of Guthrie.Google Scholar

¹³ Guthrie, M., ‘Bantu origins: a tentative new hypothesis’, Journal of African Languages, 1, 9–23 (1962), 35.Google Scholar

¹⁴ Evidence concerning the lexical archaism of Icelandic is presented by Bergsland, and Vogt, in their article ‘On the validity of glottochronology’, Current Anthropology, III: (1962), 335–29.Google Scholar

¹⁵ For a general exposition of the doctrine of the Neolinguistica see Bertoni, G., Breviario di neolinguistica (Modena, 1928).Google Scholar

¹⁶ Journalof African History III, 2 (1962), 186–7.Google Scholar

¹⁷ Guthrie, , op. cit. p. 11 n.Google Scholar

¹⁸ It should be noted that I nowhere speculated regarding the route taken by the Bantu in reaching their present location.

¹⁹ Posnansky, M., ‘Bantu Genesis: Archaeological Reflexions’, J. Afr. Hist., IX, x (1968), II.Google Scholar

²⁰ Oliver, , op. Cit. 362.Google Scholar

²¹ Guthrie, , op. cit. 9–10.Google Scholar

²² Guthrie, , op. cit. 9–10.Google Scholar

²³ I wish to express my thanks to Eric Hamp for making a copy of this excursus available to me. I am also indebted to Professor Hamp for a critical reading of the final version and particularly for his remarks on my Indo-European examples cited later in this paper.

²⁴ Greenberg, , op. cit. 33–7.Google Scholar

²⁵ A possible source for the spread of this m- form in Bantu is the following. The object pronoun of Class I is mu− not only in 13antu but elsewhere (e.g. Fula mo). It could have spread analogically from the object pronoun to the noun and adjective and then to other classes. It is perhaps unnecessary to remark that the force of the argument against the conspiracy theory is not dependent on the validity of this specific theory to account for the Bantu situation.

²⁶ Codrington, R. H., The Melanesian Languages (Oxford, 1885), 10.Google Scholar

²⁷ Codrington, , op. cit. 11.Google Scholar

²⁸ In Table I, I have avoided diacritics wherever possible, and have in other ways simplified the transcription. The only phonetic explanation required is that 2 stands for the glottal stop and c for a voiced pharyngeal. Grammatical elements such as prefixes and suffixes not belonging to the root are enclosed in parentheses.

²⁹ Guthrie, , op. cit. 21, n. 3.Google Scholar

³⁰ Winston, F. D. D., ‘The Nominal Class System of Lok’, African Language Studies 111 (1962), 49–70.Google Scholar

³¹ Noun class prefixes are enclosed in parentheses. Proto-Bantu forms are cited, where possible, from Guthrie's Comparative Bantu, Part II. Where Guthrie gives multiple forms, I have cited the most common first and included some of the others in parentheses. (The grave accent indicates low tone, the acute high tone, and the circumflex falling tone.)

³² ‘An outline of the Dakarkari noun-class system’ in La classification nominale dans les angues négro-africaines (Paris, 1967), 237–62.Google Scholar

³³ Rowlands, E. C., ‘Notes on some class languages of Northern Nigeria’, African Language Studies, III, 71–81.Google Scholar

³⁴ Guthrie, , op. cit. 18.Google Scholar

³⁵ Guthrie, , op. cit. 18.Google Scholar

³⁶ Fodor, I., The problems in the Classification of the African languages: Methodological ond theoretical conclusions concerning the classification system of Joseph H. Greenberg (Budapest, 1966).Google Scholar

³⁷ This alternation still survives in Bantu itself in Kirundi and a few other languages, but the palatalization explanation becomes confirmed, and strengthened by the evidence of other Benue-Congo languages. I presented these results in a talk at the West African language Congress at Ibadan in 1964. They were never published, but in 1965, Stappers published a paper which contains essentially identical data and conclusions. (See Stappers, L., ‘Het Hoofdtelwoord in de Bantoe-Talen’, Africana Linguistica, 11, 377–99, Tervuren, 1965).Google Scholar