Ergativity and linguistic geography

Abstract

Abstract Nothing in our present understanding of accusative and ergative structures and their functions would predict major discrepancies in the frequencies of alignments in the various continents, but such asymmetries exist: Australia and Eurasia have far more ergative languages than expected, the New World has far more stative‐active languages, and Africa has far fewer non‐accusative languages. In terms of its association with other grammatical features, its areal distribution, and its consistency in genetic groupings, ergativity proves to be a recessive phenomenon, while stative‐active is not recessive but simply low in frequency. This makes the high frequency of ergativity in two continents even more unexpected. The continental asymmetries cannot be explained as reflecting the evolution of language or founder effects in the peopling of continents. Most of them have resulted accidentally from structural and genetic skewings created by spread zones in continental interiors. The alignment asymmetries suggest that the spread zones have been in existence for many, many millennia.