¹ In a lecture delivered at the Ecole Nationale des Langues Orientales Vivantes, Easter 1960: see Guthrie, M.., ‘Problèmes de génétique linguistique: la question du Bantu Commun’, Travaux de l'Institut de Linguistique, iv (Université de Paris, 1959), 83–92Google Scholar (actually published in 1961). This was followed by a paper directed at historians rather than linguists, presented at the Third Conference of African History and Archaeology, School of Oriental and African Studies, 1961: see Guthrie, M.., ‘Some developments in the prehistory of the Bantu languages’, J. Afr. Hist. III, 2 (1962), 273–82.CrossRefGoogle Scholar

² Oliver, R.., ‘The problem of the Bantu expansion’, Journal of the Royal Society of Arts, CXIV (1966), 852–69Google Scholar; reprinted J. Afr. Hist., vii, 3 (1966), 361–76.Google Scholar

³ Published by Gregg International Publishers, Farnborough (Hants), 1967–71, price £50

⁴ See Gray, R.., ‘A report on the Third Conference on African History and Archaeology’, J. Afr. Hist., III, 2 (1962), 175–91, esp. 185–7.CrossRefGoogle Scholar

⁵ Guthrie, M., ‘Contributions from Comparative Bantu to the study of African prehistory’, in Dalby, D., ed., Language and History in Africa (London, 1970), 20–49.Google Scholar

⁶ ‘Papers on Comparative Bantu’, African Language Studies, XIV (SOAS, 1973)Google Scholar, including: Meussen, A. E., ‘Test cases for method’, 6–18Google Scholar; Bennett, Patrick R., ‘Identification, classification and Bantu linguistics’, 19–25Google Scholar; Mann, M.., ‘Sound-correspondences and sound-shifts’, 26–35Google Scholar; Carter, H.., ‘Tonal data’, 36–52Google Scholar; Slavíková, M.. and Bryan, M. A., ‘The case of two Swahili dialects’, 53–81Google Scholar; Henrici, A.., ‘Numerical classification of Bantu languages’, 82–104.Google Scholar

⁷ Italicized references to Guthrie's Part I will be to his chapter/paragraph numbers.

⁸ Guthrie employs the term ‘topology’ to cover the study of the geographical dispersion of Common Bantu items.

⁹ See Guthrie, M.., The Classification of the Bantu Languages (London, 1948), 20–30Google Scholar; Dalby, D.., ‘Reflections on the classification of African languages, with special reference to the work of Sigismund Wilhelm Koelle and Malcolm Guthrie’, African Language Studies, xi (SOAS, 1970), 147–71, esp. 152–4.Google Scholar

¹⁰ For the actual number of items considered, see Part I, 61.41 and footnote.

¹¹ Cf. 83.13: ‘The most that can be safely inferred from the available statistical evidence is that PB-X could have been spoken somewhere in the southern sub-equatorial region between the upper Lualaba in the east and the upper Kwilu in the west’; and 83.14: ‘It would therefore seem feasible to interpret these indications [of topographical terms in Common Bantu] as implying a location to the north-west of the Katanga, say between the 20th and 25th meridians west and between the 5th and 7th parallels south, which fits well within the limits suggested for the central nucleus …’ (my italics, to emphasize the tentative nature of Guthrie's inference).

¹² As expressed in its modest sub-title: ‘Introduction to the comparative linguistics and prehistory of the Bantu languages’.

¹³ African Language Studies, xiv.

¹⁴ i.e. Guthrie's ‘two-stage method’, by which his establishment of a reconstructed Common Bantu corpus (based on the synchronic comparison of languages) was kept rigidly separate from his subsequent establishment of Proto-Bantu ‘source-items’, designed to explain the likely historical origin of his (modern) Common Bantu items: see Guthrie, M.., ‘A two-stage method of comparative Bantu study’, African Language Studies, III (SOAS, 1962), 1–24.Google Scholar

¹⁵ African Language Studies, xiv, 17–18.Google Scholar

¹⁶ Cf. Cope, A. T., ‘A consolidated classification of the Bantu languages’, African Studies, xxx (1971), 213–36, esp. 215CrossRefGoogle Scholar; Dalby, , ‘Reflections’, 159–60 and note 30.Google Scholar

¹⁷ Greenberg, J. H., ‘Linguistic evidence regarding Bantu origins’, J. Afr Hist., xiii, 2 (1972), 189–216, esp. 195.CrossRefGoogle Scholar

¹⁸ See Henrici, , African Language Studies, xiv, 88Google Scholar; also Bennett, , op. cit., 24.Google Scholar

¹⁹ See note 9 above.

²⁰ Bennett, op. cit.

²¹ See also Mann, , African Language Studies, xiv.Google Scholar

²² Guthrie, M.., ‘Some uses of arithmetical computation in comparative Bantu studies’, Transactions of the Philological Society for 1964 (1965), 108–28.Google Scholar

²³ Heine, B.., ‘Zur genetischen Gliederung der Bantu-Sprachen’, Afrika und Übersee, LVI (1972–1973), 164–85.Google Scholar

²⁴ Henrici, , op. cit., 82Google Scholar; Heine, , op. cit., 175 (my translation).Google Scholar

²⁵ Guthrie, , Classification, 20.Google Scholar

²⁶ Henrici, , op. cit. Fig. 5.Google Scholar

²⁷ Henrici, , op. cit. 87–8Google Scholar (including discussion of botanical/zoological ‘phenetic’ and ‘phyletic’ classifications, the former according to resemblances between observed characteristics and the latter according to a suggested evolutionary relationship).

²⁸ In terms of the number of Common Bantu items shared between the members of each pair.

²⁹ Map 2 below is based on Henrici, op. cit. Fig. 6(c); additional statistical information on the extension and amalgamation of clusters at higher levels of ‘Linguistic Distance’ (as defined by Henrici) was kindly provided by Michael Mann (pers. comm.). Mr Mann has pointed out a minor error in the drawing of Henrici's Fig. 6(c), in that RII should cluster with L33, not with K14.

³⁰ On the basis of Eastern unity versus a relative Western disunity, however, rather than between two equal unities.

³¹ Greenberg, , ‘Linguistic evidence’;C. Ehret, ‘Bantu origins and history: critique and interpretation’, Transafrican Journal of History, II 1 (1972), 1–9.Google Scholar

³² Ehret, , op. cit. 1.Google Scholar

³³ Guthrie, M.., ‘Bantu origins: a tentative new hypothesis’, Journal of African Languages, I, 1 (1962), 9–21Google Scholar, esp. 20 (although this use of the term ‘Pre-Bantu’ is in the context of an ill-informed statement about relationships between Bantu and certain West African languages).

³⁴ This use of ‘Proto-Bantu’ would seem preferable to any further use of ‘Pre-Bantu’, the latter being potentially ambiguous (in the alternative sense of non-Bantu speaking people who may have preceded the Bantu in any particular area of Africa).

³⁵ Greenberg, , ‘Linguistic evidence’, p. 195Google Scholar, referring to Oliver, , ‘Bantu expansion’Google Scholar; see also Oliver, R.., ‘The beginnings of Bantu history’, Perspectives nouvelles sur le passé de l'Afrique Noire et de Madagascar (Mélanges offerts à Hubert Deschamps, Paris, 1974), 159–69.Google Scholar

³⁶ Guthrie, , ‘Bantu origins’, 15.Google Scholar

³⁷ Oliver, , ‘Beginnings of Bantu history’, 168.Google Scholar

³⁸ There is also the possibility that some Bantu languages, stemming directly from Proto-Bantu I, were carried along the northern margins of the forest into what is now N. E. Zaire (as would be consonant with the sharp linguistic divide between Bantu languages of this area and Bantu languages further east). In this respect, Heine's still very tentative results (‘Zur genetischen Gliederung’, 172 and 182) are of interest in respect of Baali D21, Kumu 023, Amba D22, Bira D32, Nyali D33 and Mbuttu: these languages are largely among Guthrie's ‘sub-Bantu’ languages, in which the typically Bantu grammatical system is present only in a fragmentary form (see Guthrie, , ‘Classification’, 19 and 40–2Google Scholar). Unfortunately, the absence of adequate documentation for these languages did not allow the selection of any one of them as a Test Language for Comparative Bantu.

³⁹ As suggested by a computerized hierarchy of overlapping clusters, based on nearest neighbour classification (described by Henrici, , African Language Studies xiv, 100Google Scholar): data kindly provided by Michael Mann.

⁴⁰ See further discussion in the second part of this article (and cf. note 45 below).

⁴¹ Cf. Mann, W. M., ‘Internal relationships of the Bantu languages: prospects for topological research’Google Scholar, in Dalby, , Language and History, 133–45Google Scholar, esp. Diagram 1 (showing the substantially larger proportion of predominantly ‘Eastern’ reflexes in L33 and L31 than in any other Western Test Languages, although substantially less than in any Eastern Test Language, and also the substantially larger proportion of predominantly ‘Western’ reflexes in L33 and L31 than in any Eastern Test Language).

⁴² See further discussion in the second part of this article, including the role of the acquisition of iron in the triggering of the Bantu-speaking ‘explosion’.

⁴³ Henrici, , African Language Studies, xiv, Fig. 2.Google Scholar

⁴⁴ Henrici, , op. cit. 89.Google Scholar

⁴⁵ Where there is such a clear divergence, as with C71 and L31, geographic proximity is likely to have reduced the effect of (historical) linguistic distance, so that the gap recorded by the computer between these two languages is probably of even greater historical significance than at first appears.

⁴⁶ Or the largest average number in the case of clusters of closely related Test Languages (D62/EII/E13/E15, E51/E55 and S33/S41/S42).

⁴⁷ It is important to recognize the synchronic status of the languages shown on this diagram, notwithstanding the overlay of the hypothesized Proto-Bantu nuclei (which might be logically criticized for this reason). There is no justification for describing one modern language as ‘older’ than another, or—on the basis of the present diagram—for suggesting that Herero R31, for example, might be ‘derived’ from Luba-Katanga L33.