[8] This does not preclude the aspiration of consonants, or the occasional local change of a palatal into a guttural.
[9] As already mentioned, a somewhat similar concord is also present as regards the suffixes of the Fula and the Kiama (Tem) languages in Western Africa, and as regards the prefixes of the Timne language of Sierra Leone; it exists likewise in Hottentot and less markedly in many Aryan, Semitic and Hamitic tongues.
[10] An apparent but not a real exception to this rule is in the second person plural of the imperative mood, where an abbreviated form of the pronoun is affixed to the verb. Other phases of the verb may be occasionally emphasized by the repetition of the governing pronoun at the end.
[11] The full hypothetical forms of the prefixes as joined with definite articles—Ñgumu, Mbaba, Ñgimi, Ñgama and so on—are added in brackets. Forms very like these are met with still in the Mt. Elgon languages (Group No. 3) and in Subiya group (No. 32).
[12] This is prominently met with in East Africa, and also in the various Bechuana dialects of Central South Africa, where it takes the form of ñ at the end of words.
[13] Or perhaps ñga-ba-ntu (afterwards ña-ba-, aba-); the form ñgabantu is actually met with in Zulu-Kaffir: also ñgumuntu.
[14] Likewise ba- may have meant "two" (Bantu root Bali = two); a dual first and then a plural.
[15] Wa- in Luganda. In Lusoga (north coast of Victoria Nyanza) Wa- becomes Γa (Gha).
[16] Mi is possibly a softening of ñgi, ñi; ñgi becomes in some dialects nji, ndi, ni or mbi; there is in some of the coast Cameroon languages, and in the north-eastern Congo, a word mbi, mba for "I," "me," which seems to be borrowed from the Sudanian Mundu tongues. The possessive pronoun for the first person is devired from two forms, -ami and -añgi (-am, -añgu, -anji, -ambi, &c.).
[17] An exception to this rule is the verbal particle li or di, which means "to be."