The syntax of number and modification: An investigation of the Kipsigis DP
Maria Kouneli
May 2019

This dissertation is an in-depth investigation of DP structure in Kipsigis, an extremely understudied Southern Nilotic language spoken in Kenya. All data come from original fieldwork, and the theoretical focus is on the following three broad areas: number morphology, the syntax of adjectives, and the syntax of determiner spreading. Kipsigis has a tripartite system of number marking: some nouns are morphologically unmarked in the singular and form their plural with a plural suffix, some nouns are unmarked in the plural and form their singular with a singulative suffix, while a third class of nouns always have a singulative suffix in the singular, and a plural suffix in the plural. I argue that this pattern is due to the existence of a noun classification system based on number features on little n, which interact with number features on Num in a way that generates the three types of number marking. My analysis corroborates the existence of number-based noun classification, which has been argued before for Tanoan languages (Harbour 2007). I also show that the Kipsigis morphological number classes are orthogonal to the mass/count distinction in the language, contra Grimm’s (2012; 2018) claims for Nilo-Saharan. Finally, I argue that the term ‘singulative’ is misleading, with at least two types of singulatives cross-linguistically: true allomorphs of singular (e.g., in Kipsigis) or morphemes with a classifier-like function in the syntax (e.g., in Ojibwe). Adjectives in Kipsigis can only modify a noun as predicates inside a relative clause, despite overwhelming evidence that they constitute a morhosyntactic category distinct from (stative) verbs in the language. The Kipsigis data, thus, provide strong support for Baker’s (2003a) claims that adjectives are a universal lexical category and that direct nominal modification is not their flagship property. They also support a separationist view of adjectival syntax, along the lines of Cinque’s (2010) – among many others – direct vs. indirect modification distinction, and provide clear evidence for the analysis of the latter type of adjectives in terms of a relative clause structure. Kipsigis is also added to a list of languages that completely lack direct modification adjectives, and I briefly discuss why the syntax of direct modification might be unavailable in some languages. In Kipsigis each adjective or full relative clause that modifies the noun must be preceded by one of four determiners; three of these determiners are demonstratives, making Kipsigis a rare example of a language with demonstrative spreading (as opposed to the common spreading of definite articles; cf. Alexiadou 2014). I argue that the complex distribution of these determiners is best captured by an analysis of relative clauses as D’s with a clausal complement (Kayne 1994 among others), with my analysis of determiner spreading being close to Alexiadou & Wilder’s (1998) analysis of determiner spreading in Greek. I discuss the implications of the Kipsigis data for the typology of determiner spreading (also called definiteness agreement in the literature), and I conclude that a relative clause structure is key to the understanding of the phenomenon, even in languages that are often analyzed in terms of agreement/concord (e.g., many Semitic languages). Finally, I show that demonstratives in Kipsigis are best analyzed as D heads, and conclude that demonstratives can be either heads or phrasal depending on the language.
Format: [ pdf ]
Reference: lingbuzz/004599
(please use that when you cite this article)
Published in: Ph.D. dissertation, NYU
keywords: determiner spreading; adjectives; kipsigis; nilotic; number morphology, morphology, syntax
Downloaded:552 times


[ edit this article | back to article list ]