This week on the ELAR blog, SOAS graduate Rebekah Hayes tells us about her experience using an ELAR collection for her MA dissertation. Rebekah completed her MA in Language Documentation and Description at SOAS in 2019, and is now a Research Fellow on the True Echos research project at the British Library.

Why did you choose to work with an ELAR collection?

I decided to work with an ELAR collection for three reasons, two of which were time and financial restrictions. Although I did my MA part-time over two years, I was working throughout and it would have been tricky to both take time out for fieldwork and to fund fieldwork myself. Thirdly, I started working at ELAR in April 2018 so by the time I began to plan my dissertation, I was well acquainted with the collections and wanted to make use of them in my work!

I’ve always been interested in syntax and Austronesian languages so I already had clear areas of interest on which to base my search. I used ELAR’s website to look for collections within the geographic region that I was interested in and then explored the materials that were in relevant collections.

I decided to use Dr Timothy Brickell’s deposit ‘Tonsawang: a collaborative multimedia project documenting an endangered language of North Sulawesi‘. The collection materials are nearly all open-access and they include audio and video recordings, as well as ELAN files.

I initially contacted Dr Brickell in December 2018 as I wanted to ask permission to use his collection for my dissertation work. Although the collection was mainly open access, I wanted to ensure that my research wouldn’t overlap with his own research/publishing plans. Dr Brickell was very helpful in answering my questions over the following months and letting me know about updates to the collection materials and his own analysis.

A benefit to working with the collection was the inclusion of annotated transcriptions in the form of ELAN files. Transcription is very time-consuming – especially with a dissertation deadline! – so it was helpful to have this work already completed for the 11 texts I used In my sample.

A challenge was using corpus-based texts to conduct morphosyntactic research instead of using elicited data; the dataset could provide only positive evidence and couldn’t show what was ungrammatical. Due to the nature of the texts and my small sample size, I also wasn’t able to carry out different subjecthood tests, for example. However, I was still able to carry out my research and – as I noted in my dissertation – using secondary data will become more necessary for studying languages that are no longer spoken.

As part of my research, I wanted to find where Tonsawang fits within the typology of Western Austronesian languages. These languages have symmetrical voice alternations but can be subdivided into Philippine-type or Indonesian-type based on different morphosyntactic criteria. I expected Tonsawang to be Philippine-type like other languages in Northern Sulawesi (Himmelmann 2005; Blust 2013).

In fact, I found that although some features of the language suggest Tonsawang is Philippine-type, others like case-marking and constituent order indicate that the language is Indonesian-type. I suggested this could be due to structural change because of language endangerment and because Tonsawang could be a transitioning from a Philippine-type language to a more Indonesian-type language (Hemmings 2015, 2016).

I would firstly recommend having clear research interests before looking for collections that you might want to use. It certainly helps for narrowing down your search, especially considering the large number of collections that ELAR holds!