Testing mutual intelligibility between closely related languages in an oral society

This paper describes a new methodology for testing intelligibility across closely related languages and dialects in a traditional oral society in Vanuatu. There are many reasons why it could be useful to establish how well speakers of related varieties can understand one another: such knowledge is relevant to language planning and policy making, and it can shed light on the dynamics of language contact. However, conventional approaches to intelligibility testing, such as 'recorded text testing' (Hickerton et al. 1952; Pierce 1952; Voegelin & Harris 1951), are time consuming to score, and difficult to implement consistently. In Europe, fast and efficient intelligibility testing has been successfully carried out across closely related varieties (cf. Vanhove 2014; Gooskens forthcoming; Schüppert & Gooskens 2011a, 2011b, inter alia). However, these methods assume that test subjects are literate and computer-savvy. The methodology discussed in the present paper adapts European methods to conventional 'fieldwork' conditions. In Vanuatu we piloted a picture task and a translation task. Although some words had to be removed from the final analysis, the experiment was successful overall and we anticipate that this method can be fruitfully applied in other oral language communities.