The Prosody and Syntax of Zulu Relative Clauses

Abstract

Jokweni suggests this data can be accounted for by placing a prosodic phrase break both at the right and at the left edges of a clause: that is, a prosodic phrase is roughly coextensive with a CP. The prosody of restrictive relative clauses and related structures, as they have an embedded CP, could provide crucial evidence that Nguni prosodic phrases are consistently coextensive with CP, however Jokweni (1995) does not examine them. In this paper we present a study of Durban Zulu, another Nguni language spoken in South Africa. We show first that Zulu has basically the same prosodic phrasing as Xhosa for data comparable to that in (1): prosodic phrases are roughly coextensive with a maximal clause. However, the prosodic phrasing of restrictive relative clauses shows that it is not accurate to propose that, formally, prosodic phrases are consistently coextensive with CPs in Zulu. While such a phrasing algorithm correctly predicts that there should be a prosodic phrase boundary at the right edge of the relative clause, it wrongly predicts that there should either be a prosodic phrase break separating the head noun from the rest of the relative clause (in a traditional analysis) or one separating the whole relative clause including the internal head from the rest of the sentence, in a Kaynian analysis. We argue in section 2 that restrictive relative clauses provide evidence that prosodic phrasing is most consistently conditioned only by the right edge of CP in Zulu. Further evidence for the primacy of right edges of CP in conditioning prosodic phrasing comes from the fact that, in Zulu and in Xhosa, there is no prosodic phrase break preceding a sentential complement, as shown in (1b). In sections 3 and 4, we examine the phrasing of structures related to restrictive relative clauses, namely, non-restrictive relatives and clefts. These constructions appear to be problematic for the analysis, as we find a consistent prosodic phrase