FIELDWORK, REVITALIZATION, AND DIGITAL TOOLS
My work is built upon data collected through fieldwork with speakers of understudied and endangered languages. Since 2015, I've conducted fieldwork on Nukuoro, an endangered Polynesian Outlier language spoken in Micronesia and the US, which has about 1000 speakers. You can find my Nukuoro field materials in the California Language Archive at UC Berkeley.
I am involved in community-led efforts to document and revitalize the Nukuoro language, which includes building digital language tools like the Nukuoro Talking Dictionary, our YouTube channel, educational materials, a Nukuoro phrasebook. A Nukuoro creation story told by Johnny Rudolph is also available in both Nukuoro and English on YouTube.
Many languages distinguish ergative subjects syntactically, and it has been claimed that all such languages have morphological ergative marking. In Nukuoro, however, there is an ergative extraction restriction but no overt ergative marking. While this may call for a dissociation of syntactic and morphological ergativity, I question whether this is a stable pattern or reflective of an unmarked abstract ergative Case.
THE LIMITS OF AGREE
I investigate the syntactic mechanism of Agree in a variety of empirical domains, including cross-linguistic participant number "agreement" with internal arguments and ditransitive person restrictions in Caquinte, which are reminiscent of the Person Case Constraint. In these cases, what appear to be Agree-based phenomena can actually be attributed to distinct mechanisms in the semantic or morphological component, respectively. By understanding what Agree does not do, our theory can better understand what Agree is responsible for and the different strategies languages use to accomplish similar cognitive or economic goals.
Verbal elements in Nukuoro and other Polynesian languages are often nominalized. I am exploring the structure of these clausal or verbal nominalizations, particularly subordinated verbal elements and genitive relative clauses, which are all large enough to host tense and aspect. Interestingly, genitive relative clauses mark alienability despite there being no semantic possession relationship; this alienability marking appears to track either agentivity or ergativity.