Search

The Ninde language (Oceanic, Vanuatu) has been placed in a Malekula Interior subgroup(Tryon 1976) and a Western Malekula Linkage (Lynch 2016), but oral tradition identifies two parent languages (Letpen 2018). These three proposals have different consequences for which languages should be included in a comparative study, but the two monophyletic approaches treat Ninde as a divergent branch within its subgroup or linkage. Based on new data and oral history collected as part of a community-led documentation project, this dissertation provides the first diachronic study of Ninde as a mixed language. Computer-assisted applications of the comparative method were used to assess these competing proposals. A random forest model was fit to lexical data from northern Malekula, offering an alternative to the analysis by Lynch and Brotchie (2010) that linguolabial consonants represent an areally diffused trait. Diachronic sound change models, implemented algorithmically as sequences of exceptionless innovations, aided in identifying two sources of words that together vaccount for most of Ninde's lexicon. These novel tools do not replace the comparative method as the standard by which phylogenetic claims are assessed, but instead, complement them. Ninde is not a divergent member of any subgroup or linkage; rather, it inherits lexicon and grammar from two parent languages. One of these is also an ancestor of Naati and Nahavaq, which are spoken in the South West Bay region where Ninde is spoken today. Descendants of the other include Avava and Neverver, which are spoken in North Central Malekula. These findings highlight the importance and precision of local traditional knowledge, which merely corroborated by the innovative computational tools used in this dissertation.

Spoken data from language-contact situations is extremely varied. This heterogeneity makes it difficult to make comparisons across corpora, and to use corpus linguistic tools on the data. Standardized systems have been proposed for other kinds of linguistic annotation, such as phonetic transcription, e.g., IPA (International Phonetic Association 1999); orthographic transcription, e.g., GAT (Selting 1998; Selting et al. 2009; Schmidt et al. 2015); and POS-tagging, e.g., STTS (Schiller et al. 1999; Westpfahl et al. 2017), but there is no standardized system for normalization or language-tagging. I propose such a system, i.e., I construct and implement guidelines for the systematic normalization and language-tagging of (German-English) data. My normalization and language-tagging system is based on the annotation systems of ten existing German variety corpora. As a case-study for the implementation of my proposed normalization and language-tagging guidelines, I also construct a corpus of Texas German using transcriptions of conversational interviews from the Texas German Dialect Project (Boas et al. 2010).