Associative a and ka in Niger-Congo

Abstract

A grammar of almost any Bantu language will attest a morpheme a in a variety of constructions involving association, usually of a preceding noun with a following noun or pronoun. In many Bantu languages a second morpheme, with the shape ka or something similar, is used in some similar constructions. These morphemes have frequently been labelled 'possessive', with reference to one of their more obvious uses, or genitive, with the implication that more than literal possession may be involved; the variety of constructions in which they are found makes the more neutral term 'associative' preferable. In the history of efforts to establish and define the relationship of the Bantu languages with most of the languages of West Africa, it has not been widely noted that striking parallels can be found in a number of unique morphological details of such a type that borrowing or other influence is most improbable, thus strongly reinforcing the arguments for recognizing a genetically related NigerCongo language family. It is the purpose of this article to point out some reflexes of the Bantu associative a and ka in some widely divergent non-Bantu languages of West Africa, in most of the same uses and with significantly parallel restrictions. A comprehensive treatment is impossible at present for lack of sufficient evidence; the evidence that is available, however, will establish that these morphemes can be reconstructed for proto-Niger-Congo. It is hoped that this will also stimulate further investigation of comparable phenomena in other NigerCongo languages. The treatment accorded here to these morphemes in the Bantu languages themselves will seem sketchy and elementary to a Bantuist. He is sufficiently familiar with the material, and needs little reminder of the varieties and ramifications of these constructions in Bantu. Others, however, will find a brief sampling of the uses of the associative a and ka in a few Bantu languages necessary andit is hoped-adequate as a background for the non-Bantu evidence. The more restricted and subtle evidence of these morphemes in non-Bantu Niger-Congo languages, on the other hand, will require more detailed treatment irrespective of one's language background.