Abstract
This study seeks to evaluate the degree to which classical kinship models are able to capture variation in kinship terminology. Following a description of the classical typologies of kinship, the authors use statistical approaches and a new database, Kinbank, to assess the internal coherence of these typologies; they find that these typologies are not necessarily accurate predictors of kinship terms across cultures. In addition, this analysis showed that the use of one sort of kinship typology for one generation may only weakly indicate its use in an adjacent generation. The authors set out to try and identify new typologies using statistical modeling based on the single-generation kinship terms of 306 languages. They successfully identify 9 clusters of kinship terms for just one generation alone, indicating that kinship terminology is much more diverse than previously thought. They conclude kinship typology and variation in kinship terminology needs to be investigated more thoroughly.