Locative object marking and the argument-adjunct distinction

Abstract

This paper examines the status of locative phrases in Bantu and the argument-adjunct distinction. We look at verbal locative agreement and at other morphosyntactic patterns related to locative phrases in different Bantu languages including Kiswahili, Sambaa, Haya, and several Nguni languages. We propose that object marking cannot be taken to be evidence for the objecthood of the corresponding NP in Bantu. We focus on three domains where verbal agreement marking of locative complements differs from the corresponding marking of non-locative complements: (i) object-marking paradigms without locative markers, (ii) contexts in which locative object markers may be used, but where other, non-locative object markers are disallowed, and (iii) locative complements marked by post-verbal locative clitics rather than, or in addition to, locative object markers. We then show how variation in locative object marking is related more generally to corresponding variation in locative marking in Bantu on one hand, and on the other hand to several processes of grammaticalisation for which locative markers serve as a starting point. We propose that, because the system of morphosyntactic coding of locatives in Bantu is in flux, locative object marking may be related to more or less object-like grammatical functions, depending on the particular state within the wider process of transition of locative marking, and hence presents a mixed picture.