An argument for ezafe constructions and construct state in Zulu

Taylor Jones

doi:10.3765/plsa.v3i1.4353

Abstract

Zulu (Niger-Congo, South Africa) exhibits a complicated interplay between morphological and phonological processes that, combined with an inherited, traditionalist approach to syntactic categories in the Nguni languages, obscures an overt phoneme that is argued to be common to all complex DPs. Here it is claimed that the traditional categories of adjective, relative, compound noun, possessive, and demonstrative can all be unified under a DP approach that takes this overt phoneme to be a functional marker that is analogous to those appearing in Persian ezafe constructions and construct state constructions in Afro-Asiatic languages. This approach reduces a variety of seemingly different Zulu-specific phenomena to a single, cross-linguistically established phenomenon.

Highlights

This paper argues that a wide range of DP phenomena in Zulu can be conceptually reduced to the same phenomenon; one which is not Nguni or even Bantu specific
It is shown that complex DPs in Zulu all exhibit the same “linking” vowel
It seems clear that there is some evidence for construct state in Zulu, and that the noun phrase in 31, ‘the enemy’s destruction of the city’ exhibits the functional /a/ we expect to find in complex nominals if the above analysis is correct

Summary

Introduction

This paper argues that a wide range of DP phenomena in Zulu can be conceptually reduced to the same phenomenon; one which is not Nguni or even Bantu specific (cf. Halpert 2012 on Zulu, and Baker 2003, Baker 2008, Diercks 2012 after discussion in Halpert 2012). While Zulu noun class augments (or pre-prefixes) are generally taken to be determiners due to facts about their patterning in certain environments (e.g., their phonological absence in some kinds of negation, their absence in vocatives, etc., see de Dreu 2008, Halpert 2012 inter alia), here I remain agnostic as to the nature of the class augment, and instead pursue a different avenue of investigation, which takes the “linking” /a/ to be a functional head, potentially an overt realization of D This reduces the required architecture to analyze Zulu nominals, and brings a wide range of phenomena under the same conceptual umbrella. Segments in parentheses are generally taken to be part of the prefix because they are the subject agreement marker, but do not surface on the noun)

Plural aba imi ama izi iziN iziN

Type of DP adjective relative compound noun possessive demonstrative

Inja enkulu Izinja ezinkulu

Indlu imnyama

Conclusion