^{page 131 note 1} Also known as Bilen. I prefer the form Bilin, as a transliteration of the name used by the Bilin themselves—bəlin; the form Bilen is derived from the Tigre and Tigrinya form of the name—bəlen.

^{page 131 note 2} cf. Tucker, A.N. and Bryan, M.A., The non-Bantu languages of north-eastern Africa (Handbook of African Languages, Part in), London, O.U.P., 1956.Google Scholar

^{page 131 note 3} cf. for this term, Firth, J.R., ‘Technique of semantics’, TPS, 1935, 62.Google Scholar

^{page 131 note 4} cf. Firth, J.R., ‘Sounds and prosodies’, TPS, 1948, 151,Google Scholar and the works quoted by Sprigg, R.K., ‘The tonal system of Tibetan’, BSOAS, XVII, 1, 1955, 134,Google Scholar n. 2.

^{page 132 note 1} For the relation between grammar and phonology cf. my ‘The “broken plurala” of Tigrinya’, BSOAS, XVII, 3, 1955, 548–9.Google Scholar

^{page 132 note 2} In my enumeration each successive digit denotes a subsection of the previous section. Thus § 11, § 12, § 13, and § 14 are subsections of § 1; § 211, § 212, and § 213 are subsections of § 21, which is a subsection of § 2.

^{page 132 note 3} I take ‘theme’ from Leslau's, W.Documents tigrigna, Paris, 1941,Google Scholar in preference to ‘derived form’ or ‘aspect’, since I use ‘form’ and ‘aspect’ in a different sense. It cannot be too strongly stressed that this name, and all other names of categories, as well as all the translations, are employed solely for the identification of the categories or forms; they do not form an integral part of the linguistic analysis, which avoids all reference to notional criteria.

^{page 132 note 4} For the technical terms system and structure cf. Allen, W.S., ‘Retroflexion in Sanskrit’, BSOAS, XVI, 3, 1954, 556,CrossRefGoogle Scholar n. 2. Important technical terms, where first used, are printed in italics.

^{page 132 note 5} I use a transliteration of the Ethiopic script used by my informant, except that I write ə for the ‘sixth form’ only where there is a vowel; where there is no vowel, the consonant alone is written. The symbols have roughly the values of those of I.P.A., except that t is used for [t‘], C for [], q for [k'], j for[], y for [j], for [ħ] and kw, gw, qw, Xw, and ŋw for the labiovelars. The prominent syllable is indicated by an acute accent on the vowel symbol.

^{page 133 note 1} For instance, only one tense is here recognized for gɐbɐx^w ‘who refused’ (§ 3221), but these forms agree in case (as well as number-cum-gender and person) with the noun. There is, therefore, a different paradigm (of seven members) for each case; cf. the Appendix. On the other hand the number of tenses could be stated as 16 only, if all paradigms with similar endings are classed together as one tense. The number 40 is a convenient compromise between the two extremes.

^{page 134 note 1} I prefer to use C and V, and use ‘consonant’ and ‘vowel’ as phonetic terms. For much of the theoretical approach cf. my ‘“Openness ” in Tigre’, BSOAS, XVIII, 3, 1956, 561–77.Google Scholar

^{page 137 note 1} Except in the case of the final syllable of a sentence-final word, where a high falling pitch is the exponent either of a prominent syllable or of the absence of prominence in the word.

^{page 138 note 1} The identification of any Bilin verb is most conveniently made by the use of this form, the ‘infinitive’, for which see the end of § 3.

^{page 141 note 1} For the development of this see § 32112.

^{page 142 note 1} Verbs with stem C require a different statement from all others, and are treated in § 4.

^{page 144 note 1} For the translations see the list at the end of the section.

^{page 145 note 1} For gɐbo and jəbé see above.

^{page 145 note 2} The translational difference between the paired tenses is one of past (Aspect A) and present or future (Aspect B). But such criteria are not used to establish Aspect, and it should be noted that gɐbrá (§ 3221) ‘he will refuse’ is of Aspect A.

^{page 147 note 1} For the alternation of half close central and nil in two syllables see § 214. For the vowels of the y prosodic paradigm see below, this section.

^{page 148 note 1} My informant preferred jəbigɐt to jəbígɐt, but jəbígɐx^w to jəbigɐx^w, though he accepted both. Similarly he preferred jəbígi, jəbígid, etc., to jəbigí, jəbigíd, etc., but he accepted only jəbɐgí, etc., and not *jəbɐgi, etc.

^{page 149 note 1} The forms subsumed under this have concord with the ‘antecedent’, in case as well as number-cum-gender. For details see the Appendix.

^{page 149 note 2} An alternative form of the Aspect A tense is gɐbosɐna etc. This would be treated as the unit ௉ (the y prosodic paradigm jəbésɐna may be treated as similar to jəbét, witht ŋ, or jəbé, with ௉).

^{page 149 note 3} 3 Some of these have ‘negative’ translations, but negative is defined formally in terms of the infix –3g–.

^{page 150 note 1} This is also a feature of certain verb stems, cf. § 432. But note that it is not always a feature of the relation between Aspect A and Aspect B, as shown by the paradigms of gɐbɐx^w (A) and gɐbɐx^w (B).

^{page 151 note 1} i.e., so that the V element of the junction syllable is either 3 or the unit i. The exponents are the same—half close central or nil in z and close front in y. Their exponents differ, however, for the verbs treated in § 42, the exponent of a being half close central and of i, close front (z prosodic only); but this does not assist in the decision, since forms of both the type fəti (3) and fiti (i) are attested for this class of verbs.

^{page 151 note 2} Some integration of these paradigms with the others may be achieved if special statements are made about the junction of the alveolar and dental ‘liquids’ r, n, and 1.

^{page 152 note 1} The forms subsumed under this have concord in number-cum-gender with the ‘antecedent’. For details see the Appendix.