Disentangling geography from genealogy

One of the major challenges for current typological-comparative linguistics is to find methods to disentangle which of these reasons apply in any given situation of typological similarity. Traditionally, there has been great interest in developing methods to separate historical factors from a-historical factors through various kinds of sampling. The basic idea behind such methods is that it is possible to control for the influence of historical factors by carefully selecting languages across known genealogical and areal groupings. The remaining question then is how to distinguish universals from chance. Both questions – how to remove historical influences from the data, and how to subsequently distinguish universals of human language from chance effects – are much-debated questions in the field of linguistic typology (cf. Cysouw 2005 for a survey), and much more could be said about these topics. Yet, the current paper will not concentrate on such a-historical characteristics of human language, but focus on historical factors that result in typological similarity. When investigating historical factors leading to typological similarity, it might seem as if there are likewise two questions to be dealt with. First, how to factor out a-historical factors, and, second, how to distinguish similarities caused by genealogical descent from similarities caused by borrowing. However, on closer inspection the first question turns out to be easily answerable.