?
Geospatial effects on phonological complexity in the world’s languages
Linguistic complexity has generally been seen as influenced by ecological, demographic, and sociolinguistic factors and has been approached by seeking correlations of increased complexity along one linguistic dimension with one or another extralinguistic factor. Here we use a multidimensional definition of phonological complexity and analyze its global patterning quantitatively across predefined continents or sets of continents. We use linear and nonlinear regression models to estimate the gradient of geographical distributions of various complexity measures. We found significantly lower phonological complexity levels in South America and Australasia, suggesting isolation by distance (from centers of demographic and economic influence in the Old World). An unexpected finding is that our models show a broad transitional zone between the different levels of complexity in Africa and Europe. This Euro-African Transition Zone (ETZ) has its core between the Equator and the African coastline of the Mediterranean Sea. It has not been regarded as a language area before and is not a convergence area, but it has a robust geospatial complexity profile and we propose it as an area. We also find reasons to regard phonological complexity as a coherent property in itself, with its several dimensions behaving as a whole, and not as adaptive but as geospatially patterned. This paper is an exercise in using synchronic analysis of geolinguistic patterns of complexity to draw inferences about long-term change and evolutionary trends in language. On a higher level, the paper employs methods of geospatial analysis novel to this particular problem with the goal of raising new questions and laying groundwork for wider research into the distribution of linguistic complexity.