Tracking the dynamics of kinship and social category terms with AustKin II

The first AustKin project (AustKin I) collected a large database of kinship terms from Aboriginal languages all over Australia, endeavouring to maintain standards of spelling, kin formulae and group identities, without losing the details of original sources used. An online geospatial interface has been used to map distributions of forms of terms and their polysemies or equations. The patterns of the latter provide identification of kinship systems as defined in ethnology. The project proposed and tested hypotheses about the evolution of such systems in Australia based on knowledge of the common polysemies and related changes. The next stage, AustKin II, builds on hypotheses from the current authors and others, testing these further by adding two more components to the database: the marriage rules and the social categories used by each group. Of the latter, section and subsection systems are unique to Australia. The aim is to gauge how these different systems fit together and propose how they evolved over time and how they influenced each other. 1.The AustKin project 1.1The design of the AustKin database. The AustKin database documents words in the domain of kinship terminologies for 316 Australian languages or dialects (which could be grouped into about 200 languages, depending on criteria used). The 3i6 languages/dialects have an average of approximately two different wordlists each. from different ethnographic or historical sources for each language or dialect, with a total of over 22 000 words that belong to the domain of kinship. Designing a database and an interface to such a database has revealed itself to be a complex matter since the number and diversity of variables that need to be taken into account are considerable. In summary, the following had to be taken into account: A – Systemic variables 1) Kinship terminologies are not just words, but also relationships; and in particular they are related among each other. 2) A kinship terminology constitutes a system; but not all kinship terminologies belong to the same type of system. 3) Kinship terminologies change and they need to be placed against their chronological and historical background. B – Sporadic variables 1) Kinship terminologies are recorded by humans, and often by non-linguists; they include errors. 2) Kinship terminologies are seldom complete, and need to be completed when possible through other sources. 3) Original Informants may not always have been local speakers. Arriving at, or at least proposing, potential solutions to the B-type variables was not as difficult as it may appear. The solution chosen was to keep each kin term in its original form as it appears in the original source, while working with rewritten words (for instance those in an orthography standard for the whole database) linked to the original source, so as to be able to retrace every step of transformation and analysis