² The name ‘Erythraic’ is thus used here to embrace the following Larger Units: Semitic, Cushitic, Ancient Egyptian, Berber, Chadic, and any languages with comparable characteristics which cannot be allocated to one of these Larger Units. For the arguments underlying the choice of this name, see Tucker, and Bryan, , Linguistic analyses: the non-Bantu languages of northeastern Africa, OUP, 1966, 2.Google Scholar

³ Even here results can be disappointing: Dr. B. W. Andrzejewski tells me that he has made a detailed study of all available published vocabulary material in the various Cushitic languages and can find only the following correspondence values:

Somali and Galla—about 20%

Somali-Galla and Afar-Saho-Sidamo—under 3%

Agau-Bedauye and Kaffa-Ometo—under 1%.

⁴ Taken front Aspinion, Robert, Apprenons le berbère: initiation aux dialectes chleuhs, Rabat, 1953.Google Scholar

⁵ Taken from Mitchell, T. F., ‘Particle-noun complexes in a Berber dialect (Zuara)’, BSOAS, XV, 2, 1953, 376. These two dialects have been selected as giving the most complete forms.Google Scholar

⁶ Regarded. by F. NV. Parsons as a ‘4th person’—indefinite.

⁷ In the absence of ‘starred’ forms, the elements are set out as actually heard, hence the apparent similarity of many of them.

⁸ The n of the 1st person plural is often of a different kind from that of the 2nd and 3rd person plural, and is often difficult to distinguish from element I above.

⁹ It is rare, though not impossible, for a single language to display both t and k here.

¹⁰ Many languages, however, distinguish gender in the 3rd person singular only.

¹¹ Whereas in many languages this element may occur in both genders in the plural 3rd person, in the singular 3rd person it is confined to the masculine.

¹² The term ‘prefix paradigm’ is used here to indicate a form of conjugation in which prefixes occur (whether supplemented by suffixes or not); the term ‘suffix paradigm’ indicates a form of conjugation in which only suffixes (and no prefixes) occur.

¹³ Tucker, and Bryan, , The non-Bantu languages of north-eastern Africa, OUP, 1956;Google ScholarTucker, and Bryan, , Linguistic analyses: the non-Bantu languages of north-eastern Africa, OUP, 1966.Google Scholar

In the later volume, the ‘Larger Unit Cushitic’ (32) of our earlier volume has been subdivided into:

(a) ‘Orthodox’ Cushitic languages, i.e. those which show the main features accepted by most authorities as Cushitic: Bedauye, Agau (including Bilin, Awiya), Saho-Afar, Sidamo, Galla, Somali;

(b) partially Cushitic languages, i.e. those in which some Cushitic features occur, but which differ in many important respects from the ‘orthodox’ languages: Janjero, Ometo, Gimira, Kaffa;

(c) languages with little or no claim to be Cushitie: Konso-Geleba.

¹⁴ ‘The Mogogodo, a forgotten Cushitic people’, JAL, II, 1, 1963, 29–43.Google Scholar Greenberg's linguistic material is derived from Hobley's, C. W.Ethnology of A-kamba and other East African. tribes, Cambridge, 1910.Google Scholar

¹⁵ As rearranged by Greenberg in his Mogogodo article.

¹⁶ Greenberg's names follow consecutively in his own lists, but are spaced out here to show where they tally with our own subdivisions.

¹⁷ Allocation tentative, owing to absence of linguistic information at the time, and hence subject to correction.

¹⁸ See n. 16, above.

¹⁹ Manuals di Sidamo, Milano, 1940, 286–7, in which he compares the satem-centum division in Indo-European.Google Scholar

²⁰ Greenberg, op. cit., 42–3, allocates this language to his ‘Eastern Cushitic’ Group, witl possible affiliation with Konso-Geleba.

²¹ Made possible by a study leave grant from the School of Oriental and African Studies. 1965–6.

²² Final vowel ‘semi-mute’, see below.

²³ See Dammann, E., ‘Einige Notizen über die Sprache der Sanye (Kenya)’, ZES, XXXV. 3–4, 1950, 227–34—Greenberg's probable source.Google Scholar

²⁴ These four languages were erroneously given Bantu ‘stippling’ on the map accompanying the Linguistic survey of the northern Bantu borderland, I and IV.

²⁵ So as not to prejudge the issue, the term should be accepted in a geographical rather than a philological sense.

²⁶ Greenberg's Khoisan allocation is:

IV Khoisan

IV.A.1 South African Khoisan

IV.A.1 Northern South African Khoisan (Bushman)

2 Central South African Khoisan (Hottentot, Naron)

3 Southern South African Khoisan (Bushman)

IV.B Sandawe

IV.C Hatsa

In the ‘Handbook’ (supplement by E. 0. J. Westphal in Tucker and Bryan, The non-Bantu languages of north-eastern Africa) the allocation is: 37 Sandawe-Hottentot, 38 Bushman-Hadza.

²⁷ So as not to prejudge the issue, no attempt is made to attach numbers to the elements here or in the ‘Fringe’ languages.

²⁸ Probably demonstrative.

²⁹ Taken from Reinisch and Cerulli; note that the latter gives only tā and nē in the 1st and 2nd person singular.

³⁰ Abstracted from the paradigm kongo shobonhon etc. ‘I etc. am good’.

³¹ Kindly contributed by Professor H. W. Fairman.

³² Information on Konso is lacking.

³³ Many of the paradigms shown in this section have, for the sake of brevity, been abstracted from the larger grammatical contexts in which they normally occur.

³⁴ Largely disappeared, however, as an individual entity.

³⁵ Verb classes with conjugated auxiliary suffixes are, however, to be found in some non-Erythraic languages, such as Tubu.

³⁶ Whereas Class I verbs are relatively common in Bedauye and Saho-Afar, there are only five of them in Somali and Awiya.

³⁷ Note incapsulation of -n- and absence of t- in the singular before a triliteral stem.

³⁸ Italicized material taken from Moreno, op. cit.

³⁹ Note that, in all tenses, tone values can vary with grammatical context.

⁴⁰ Verbs of this sort can be regarded as derivative verbs in that in many cases the implication is ‘to do something for oneself’.

⁴¹ Information obtained under field conditions, and in need of checking, especially for tense.

⁴² Although a click word, it follows the regular ‘orthodox’ verbal pattern.

⁴³ The -ite etc. of the 2nd person plural in Ometo and Kaffa is actually more reminiscent of the -ata, -atai suffixes of the 2nd person plural in the Paranilotic languages, Lotuho, Karimojong, Teso.

⁴⁴ Only the longest and the shortest forms of the suffixes are given here. Note that there is here also a type of conjugation which involves merely the prefixing of the personal pronoun to the uninflected verb stem.

⁴⁵ There is, in addition, a totally different type of interlocking pattern. (Note that these remarks refer to Geleba only; information on Konso is lacking.)

⁴⁶ Kindly contributed by Professor H. W. Fairman.

⁴⁷ Typical context for these tenses: in answer to questions of the type: ‘Who will go?’ etc.

⁴⁸ There is some uncertainty as to whether these two forms distinguish (or distinguished) gender. Most speakers apparently use them interchangeably now.

⁴⁹ A parallel may be drawn in the Bantu field, where all authorities hold as equally valid for membership (a) a certain basic vocabulary which may be related to a vocabulary of proto-forms by inflexible rules of sound change, and (b) the presence of a series of paired noun classes, characterized by prefixes and a system of concordial relationship in the grammar with these prefixes. Any language which lacks this concord system is not deemed to be Bantu (Guthrie applies the term ‘Sub-Bantu’ to some of these languages). Conversely a language (such as Temne) which has a concord system (and on those grounds merits close investigation) but which lacks a sufficient body of basic vocabulary, is likewise not regarded as Bantu—though authorities from Westermann onwards have not doubted certain affinities with Bantu.

It should further be stressed that it has so far been impossible to establish proto-forms for any of the ‘Fringe languages discussed in this article.

⁵⁰ Dr. B. W. Andrzejewski, on the basis of further material supplied by the author, suggests that Waat^a is virtually a Galla dialect, Boni and Aweera have some traits in common with Somali as well as with Galla, but that Dahalo shows considerable vocabulary divergence from the other two, apart from its relatively small vocabulary of click words. These latter were compared with Hadza and Sandawe equivalents; Dr. J. C. Woodburn found no correspondence in Hadza, but Dr. E. K. ten Raa found some interesting correspondences in Sandawe.

⁵¹ The author has no Konso material, so can say nothing about that language.

⁵² But see p. 672, n. 43.

⁵³ Unfortunately the authors of Linguistic analyses had only second-hand Tepes (and no Ik^a) material to work with, and hence could make no deductions at the time.

⁵⁴ Compare the pattern in Geleba.

⁵⁵ The tentative allocation of Hadza to a ‘Bushman-Hadza’ Larger Unit in the ‘Handbook’ (p. 162) is no longer considered valid by us, though the Larger Unit ‘Sandawe-Hottentot’ is still recognized. (Note that the earlier claims of Hottentot to ‘Hamitic’ membership are now strongly contested.)

The following works have also been consulted:

Cerulli, E., Studi etiopici, IV. II linguaggio dei Oiangerd ed alcune lingue sidama dell'Omo, Roma, 1938.Google Scholar

Cerulli, E., Studi etiopici, IV. La lingua caffina, Roma, 1951 (quoted in italics).Google Scholar

Copland, B., ‘A note on the origin of the Mbugu with a text’, ZES, XXIV, 4, 1934, 241–5.Google Scholar

Green, E. C., ‘The Wambugu of Usambara’, TNR, 61, 1963, 175–89.Google Scholar

Meinhof, C., ‘Linguistische Studien in Ostafrika. X. Mbugu’, MSOS, IX, 3, 1906, 294–323.Google Scholar

Moreno, M. M., Introduzione alia lingua ometo, Milano, 1938 (quoted in italics).Google Scholar

Moreno, M. M., Manuals di Sidamo, Milano, 1940.Google Scholar

Palmer, F. R., ‘The verb in Bilin’, BSOAS, XIX, 1, 1957, 131–59.CrossRefGoogle Scholar

Palmer, F. R., ‘The verb classes of Agau (Awiya)’, MIO, VII, 2, 1959, 270–97.Google Scholar

Reinisch, L., Die Kafasprache in Nordost-Afrika, Wien, 1888 (quoted in italics).Google Scholar

Reinisch, L., Die Saho-Sprache, Wien, 1889–1890.Google Scholar

Reinisch, L., Die Bedauyesprache in Nordost-Afrika, Wien, 1893.Google Scholar

Shackleton, E. R., The Merille or Gelubba (cyclostyled), 1932 (quoted in italics).Google Scholar

Whiteley, W. H., A short description of item categories in Iraqw, Kampala, 1958.Google Scholar