Nisgha syntax and the ergativity hypothesis

Nisgha has been classified by at least three different linguists as syntactically ergative (Rigsby, Rood, and Tarpent). This is motivated by the fact that in certain constructions the agent of a transitive verb patterns differently than the patient of the transitive or the single argument of an intransitive. A new definition of syntactic ergativity has been proposed recently by Alec Marantz (1981) and is called the Ergativity Hypothesis. The definition essentially says, given the grammatical functions [NP,VP] and [NP,S], we will have the following associations of grammatical functions and thematic roles: Syntactically Ergative Syntactically Accusative Language Language Agent-[NP,VP] Agent-[NP,S] Patient-[NP,S] Patient-[NP,VP] Morphological ergativity is considered a different phenomenon. It is essentially defined as follows (following the traditional definition): Morphologically ergative languages mark the subject of a transitive verb, and leave the object of a transitive and the single argument of an intransitive unmarked. Given the above definition of syntactic ergativity and syntactic accusativity, I believe Nisgha can be shown to be syntactically accusative and morphologically ergative. Levin (1983) proposes a number of diagnostics for syntactic ergativity, none of which come out "positive" for Nishga. Some don't apply, and some show syntactic accusativity. As well as outlining the Ergativity Hypothesis in Chapter 1, I also present some of the foundational principles of the Government-Binding framework (Chomsky 1981) which will be relevant to the subsequent analysis of Nisgha. In particular, I discuss the theta-criter ion, the Projection Principle, the Case Filter, Government, Proper Government and the Empty Category Principle. Chapter II In Chapter 2 it is argued that Nisgha shows syntactic accusativity because of the position of subcategorized prepositional phrases and also because of the distribution of reflexive anaphors. The position of prepositional phrases suggests movement. I can find no factor to force extraposition of the prepositional phrase, but there does appear to be a factor forcing verb movement, namely the need to provide a means for assigning Case to [NP,S]. If we assume a D-structure SVO and verb movement to clause-initial position we get correct predictions concerning both the position of prepositional phrases and that of reflexive anaphors. We also explain the weak crossover facts in Nisgha in this way. One of the conclusions of this discussion of verb movement is that Nisgha categories only govern rightward. A further argument for rightward government (based on the structure of genitives) is also presented here. Also in Chapter 2 I discuss Marie-Lucie Tarpent's analysis of Nisgha, which is the most recent and comprehensive analysis of the language. Tarpent (1982) argues that Nisgha is syntactically ergative. Her arguments for ergativity are reviewed in the light of the Ergativity Hypothesis and are shown to be non-arguments. Chapter III This chapter contains a short discussion of the learnability issue and how it might be used as a criterion for evaluating competing analyses, such as those discussed above.Arts, Faculty ofLinguistics, Department ofGraduat