Separating layers of information

Abstract Linguistic areas, or Sprachbünde, can be described very broadly as geographical areas where a group of languages have become similar to each other because of prolonged contact between the speakers of the different languages. Numerous linguistic areas have been proposed over the years and all over the globe, and the study of linguistic areas has become an integral part of the more general study of contact linguistics and contact-induced change. Nevertheless, linguistic areas are notoriously hard to define in a consistent and meaningful way, and some scholars have even suggested abandoning the notion altogether. In this paper, I argue against abandoning the notion of linguistic area as such, but I am in favor of changing the approach or the procedure we follow in detecting or confirming linguistic areas. I would like to introduce a conceptual change, moving away from the idea of a linguistic area as a well-circumscribed geographical area towards an idea of contact zones as windows on the past. The procedure I propose is based on what may be called the anatomy of linguistic areas, consisting of a geographical, cultural-historical, communicative, and (structural-)linguistic layer. An approach to a linguistic area can start in any of these layers, but crucially refers to the others as well. One research set-up that is made possible by this approach is to set up a hypothesis-test procedure whereby one or more layers predict the others.