Investigating differential case marking in Sümi, a language of Nagaland, using language documentation and experimental methods

One goal in linguistics is to model how speakers use natural language to convey different kinds of information. In theories of grammar, two kinds of information: “who is doing what (and to whom)”, the technical term for which is case or case role; and pragmatic information about “what is important”, have been assumed to be expressed by different means within a language. However, linguists have recently discovered that in numerous languages spoken in Australia, New Guinea, and South Asia, there are noun suffixes or enclitics that appear to simultaneously provide both case and pragmatic information. The existence of such systems suggests that our current theories of grammar need to be modified, though it is unclear how as we still know little about how these grammatical systems work. In this project, I looked at Sumi, a Tibeto-Burman language of North-east India, which has such a system of case marking. In this system, speakers do not consistently mark the subject of a transitive or intransitive sentence with an enclitic that conveys case information, but their choice depends on additional semantic and pragmatic factors. This was the first study of a Tibeto-Burman language to use a combination of new quantitative corpus methods with traditional linguistic fieldwork methods, including the recording, transcription, and tagging of spoken language, to identify semantic and pragmatic factors that are relevant to speakers’ choice of noun enclitic. In this study, some factors found to be relevant were: whether the sentence had a direct object or not; the animacy of the subject; and whether it was the first mention of a subject in connected speech or not. This was also the first study of a language with such a case system to include a perception study that investigated if intonation was used by native listeners to disambiguate whether a noun suffix was conveying either case or pragmatic information. This study showed that listeners were not using differences in intonation, but rather relied on the type of sentence the suffix occurred in to determine its meaning