¹ Fleming, Harold C., ‘Baiso and Rendille: Somali outliers’, Rassegna di Studi Etiopici, xx, 1964, 35–96.Google Scholar

² Neumann, Oscar, ‘From the Somali coast through southern Ethiopia to the Sudan’, Geographical Journal, xx, 4, 1902, 373–401.CrossRefGoogle Scholar

³ Maass, Edgar, ‘Eine Reise durch das abessinische Seengebiet’, Mitteilungen der Geographischen Gesellschaft in Hamburg, XLVII, 1941, 365–91Google Scholar; Kuls, Wolfgang, ‘Beiträge zur kulturgeographie der südäthiopisehen Seenregion’, Frankfurter Geographische Hefte (Frankfurt am Main), 32, 1958.Google Scholar

⁴ Lewis, Herbert S., ‘The origins of the Galla and Somali’, Journal of African History, VII, 1, 1966, 27–46.CrossRefGoogle Scholar

⁵ For example, Sasse, Hans-Jürgen, ‘The extension of Macro-Somali’, a paper read at the Colloque International sur lea Langues Couchitiques et les Peuples qui les Parlent, Paris, 09 1975.Google Scholar

⁶ Heine, Bernd, ‘Notes on the Rendille language (Kenya)’, Afrika und Übersee, LIX, 3, 1976, 176–223Google Scholar; ‘Bermerkungen zur Elmolo-Sprache’, Afrika und Übersee, LIX, 4, 1976, 278–99Google Scholar; ‘Bermerkungen zur Boni-Sprache (Kenya)’, Afrika und Übersee, LX, 4, 1977, 242–95Google Scholar; ‘The Sam languages: a history of Rendille, Boni and Somali’, Afroasiatic Linguistics, forthcoming.

⁷ I wish to acknowledge here my gratitude to the School of and Oriental African Studies for enabling me to undertake field study in Ethiopia throughout the summer of 1976, during which period the material presented in this paper was collected.

⁸ The names of the following mainland Bayso villages were recorded: Sayno, Madoo, Worbe, Ṭiranṭira, Baafanno, Ḳéme, inḳile, Jigeessa, and Alge. Alge, where I worked, is the southernmost Bayso village.

⁹ An identical pattern of resettlement on the part of the Omotic Gaṣamba people is currently in progress. Virtually all the former inhabitants of the large south island on Lake Abaya are also now to be found along the western shore of the lake.

¹⁰ During July work on Bayso had been undertaken with an older man, Ato Buto Zako. This informant knew no Amharic and a Wolayta interpreter was employed. In spite of the fact that this arrangement proved too unwieldy and was abandoned after a few days, the data recorded showed remarkable consistency with that collected later from Ato Yaltatte. It is probably safe to say that Bayso contains no significant dialect variations.

¹¹ The principles of the International Phonetic Association, London, 1949.Google Scholar

¹² As, for example, in Moreno, M., Grammatica teorico-practica della lingua Galla, Milano, 1939, 24.Google Scholar

¹³ In the present paper items taken directly from Fleming's article are printed in italics and preceded by the initials H.F.

¹⁴ A series of glottalized sonorants are recorded by Cerulli for Sidamo; see Cerulli, Enrico, Studi etiopici, II, Roma, 1938, 41.Google Scholar

¹⁵ Concerning length Fleming cautiously observes: ‘Although vowel length and consonant doubling are normally significant in Cushitic languages, the recorded Baiso data shows this not to be the case here. In a number of cases doubling appears to be irrelevant…. However, no conclusions should be drawn until further data, including tape recordings, are available because of the strong presumption that length is probably significant’, ‘Baiso and Rendille’, 39–40Google Scholar. Except in contrastive verb forms consonant length is difficult to hear in Bayso and in many words I agree with Fleming that there is considerable fluctuation in terms of actual duration. By contrast vowel length seems to have a far greater functional load in Bayso, and is much easier to distinguish.

¹⁶ In addition to the pitch change I have an impression of greater intensity.

¹⁷ Fleming's spelling Baiso has been followed by most subsequent writers.

¹⁸ A detailed treatment of the problems of homophonous radical extension morphs is given in Hayward, R. J., Categories of the predicator in ‘Afar, with especial reference to the grammar of radical extensions, unpublished Ph.D. thesis, University of London, 1976.Google Scholar

¹⁹ The term ‘complementary pair’ is used for a transitive: intransitive pair of extendedstem verbs for which no simple-stem form is found.

²⁰ The term ‘neuter-passive’ was adopted for Cushitic by Andrzejewski, B. W., ‘Some observations on hybrid verbs in Somali’, ALS, X, 1969, 46–89Google Scholar. It signifies a passive which is strictly agentless.

²¹ Radical extensions with m and t occur either with the Cii- stem alternant of C.III verbs (see § 2.2.1.1.3) or with a special extended-stem form, Caam- and Caat- respectively.

²² Treatment of these functionally disparate categories as constituting some sort of semiautonomous unit in the morphology enables one (a) to recognize the internal cohesion of this group of markers with respect to variation in location within the verb phrase; (b) to accommodate the discontinuity of the person and number elements in 2 p. forms; and (c) to identify (synchronically) the n elements in 2 p., 3 p., and 1 p. forms, in spite of the differences in sequential location. From the structural point of view these elements of the PA seem to occur as a ‘bundle’, whether suffixed to the stem of the main verb or the auxiliary.

²³ The term concord set refers to any sub-set of the total paradigm in which all members of the sub-set are similarly inflected for such categories as aspect, tense, mood, etc., but which differ from each other in the forms of the subject markers.

²⁴ Modal concord sets, such as proximate future and dubitative, etc., are distinguished from mood concord sets, such as jussive and imperative, on formal grounds. Mood forms (a) mark negation by means of the particle aró; (b) never contain tense/modal auxiliaries; and (c) consist of an abbreviated concord set—there are two forms in the imperative, and three in the jussive only.

²⁵ The same alternations are witnessed in at least one other r-final verb, aar- ‘resemble’, but they are not found in the bulk of r-final verbs, such as: ḳaar- ‘flay’, barar- ‘become replete’, irir- ‘sew’, etc.

²⁶ It is noted that two verb roots which have obviously been taken from Amharic both appear in C.II, e.g. ḳamasee ‘he tasted’, aanee ‘he loaded’.

²⁷ As the labels assigned to them suggest the distinctive functions of many of the concord sets presented here are understood only very imperfectly.

²⁸ Examination of the future I and II forms of C.IV indicate that this formation is based on the 1 s. stem form. Thus the structure is kora-ro, kora-tto etc. The IN-infinitive likewise appears to be based on the 1 s. stem; see § 2.2.2.1.1.

²⁹ Unfortunately the negative jussive of a C.III verb was not recorded.

³⁰ Similar observations have been advanced about a whole group of verbs in ‘Afar among which the locative-existential verb -en is a prominent member; see Hayward, R. J., ‘The stative conjugation in 'Afar’, AION, NS, XXVIII, 1, 1978, § IV.Google Scholar

³¹ It is reasonable to suggest that the Bayso copula is cognate with a prefix verb form hypothesized for Proto-'Afar-Saho, *-ey, in which the particular vowel quality e is regarded as unpredictable and, therefore, as part of the lexeme; see Hayward, R. J., ‘Participles in ‘Afar—evidence for the restructuring of verb suffixes’, paper read at the Colloque International sur les Laugues Couchitiques et les Peuples qui les Parlent, Paris, 09 1975, § II.Google Scholar

³² In other Ethiopian languages transitive compound particle verbs formed with a verb meaning (in independent usage) ‘do/make’ are also found. Such a typo does not seem to occur regularly for Bayso; hemee-yel- ‘spend the evening’ (cf. hemeen ‘night/evening’) was recorded, however. Two compounds with bees- (transitivized counterpart of bee- ‘go out’) were also found, i.e. mége bees- ‘propose a name for’, cf. mége ‘name’: butto bees ‘make a hole’, cf. butto ‘hole’.

³³ cf. Amharic bäkkaň, bäḳḳah, etc.

³⁴ The form laara is the C.III compound imperfect of lee- ∼ lii-; see § 2.2.1.2.6.

³⁵ From a consideration of one or two examples noted, it would appear that an identical subordinate clause structure may express an adverbial subordinate clause of manner, e.g.

waa kaani gógo kaate

according as our God ordained

The functional range of gógo appears very like that of Amharic ǝndä or Oromo akka.

³⁶ The occurrence of waat- ‘lack’ as an auxiliary in negative conditional clauses is strikingly like what one finds in ‘Afar, which employs way- ‘lack’, an obvious (but simplestem) cognate of Bayso waat-. Compare

³⁷ This and the following two examples are taken from a text in which the speaker is describing the changes in the pattern of Bayso life brought about by movements to the western shore of Lake Abaya.

³⁸ Part II of this article will appear in BSOAS, XLII, 1, 1979.Google Scholar