Title

<nicholas.thieberger@paradisec.org.au>

University of Melbourne

Documentation in practice, developing a linked media corpus for
citation of archival media

In this talk I will outline the methodology employed in preparing a linked text and media corpus of South Efate, an Oceanic language of central Vanuatu. The corpus was presented as part of my PhD dissertation and allowed me to provide sources for most of the example sentences and all of the texts in the grammar. The ability to move by mouseclick to any point in some 18 hours of field recordings has meant that the data becomes available to both the analyst and then to speakers of the language in the future. Although analogue tapes can also provide access to audio data, the ease of use of digital material means that we can easily check on claims made, for example, in a thesis, by reference to the recordings. These linked transcripts and media files are then good archival forms that are built in the course of normal fieldwork methods.  

The software designed to build this media corpus is called Audiamus. I will be demonstrating the use of Audiamus as part of this talk.