The nanosyntax of Nguni noun class prefixes and concords
Tarald Taraldsen
March 2009

This paper argues for a fine-grained syntactic decomposition of Bantu noun class markers.

The Nguni languages (Xhosa, Zulu, Ndebele, Swati) exhibit systematic correspondences between the shape of an agreement marker (subject concords, object concords, adjectival concords) and a subpart of the full nominal class prefix of the corresponding noun class. For example, the class 2 adjectival concord is 'ba' and the full class 2 nominal prefix is 'a-ba', and the class 4 adjectival concord is 'mi' and the full class 4 nominal prefix is 'i-mi'. In general, the adjectival concords correspond to the part of the full nominal prefix following the initial vowel. However, subject concords and object concords seem to pick up on different parts of the nominal prefix in different noun classes, e.g. the class 2 subject concord is again 'ba', like the adjectival concord, while the class 4 subject concord is 'i'.

In this paper, I show that we can make sense of this only by assuming that a single morpheme can lexicalize a "stretch" of terminals rather than just a single terminal. Once we realize this, however, we can algorithmically use the distribution of morphemes over noun class prefixes, concords and demonstratives/determiners to determine the exact morphosyntactic structures underlying the various forms.

Format: [ pdf ]
Reference: lingbuzz/000876
(please use that when you cite this article)
keywords: bantu nominal morphosyntax, the morphosyntax/lexicon interface, morphology, syntax
Downloaded:2008 times


[ edit this article | back to article list ]