PhylProGramm

Morphosyntactic properties have long been considered a great challenge for the reconstruction of language history. On the one hand, they can establish deep genealogical relationships because they remain stable for millennia and evolve slowly in time, serving as markers for families. On the other hand, they may emerge abruptly as a result of borrowing from one language to the other or language restructuring under contact. At the same time, grammars develop common properties independently of history and geography, reflecting universal trends. Recent findings have shown that morphosyntactic variability significantly correlates with genetic variability, supporting the potential of morphosyntax to unravel both genetic and cultural evolutionary history. Yet, the first attempts to reconstruct phylogenetic trees based on morphosyntactic data have been limited in terms of precision and resolution, and there are major disagreements on the genealogical stability and diffusibility of morphosyntactic features. 

The project “Phylogenies probing Grammar. Exploring morphosyntax at different scales of language change” (PhylProGramm) aims to break new ground in linguistics by pioneering a new way to evaluate morphosyntax in terms of the historical information it contains, using phylogenetic inference methods as a tool. Building upon preliminary results which have shown that trees constructed from morphosyntactic data reflect historical and geographical relationships, we will devise novel computational methods that, in combination with rigorous linguistic evaluation, will allow us to determine which features are conservative, prone to transfers/borrowings or reflect universal patterns. The investigation will lead us to formulate and explain the profile of specific units, their interactions, their organization into networks and/or hierarchies associated with different scales of variation, and their representation in grammar, providing a unique window into the dynamics and typology of language change and variation.

The project will bring together an interdisciplinary team of linguists, evolutionary and computational biologists and computer scientists who will identify the morphosyntactic features that reflect genealogical relationships, as opposed to areal and typological ones. It will employ established and novel biology inspired/ quantitative methods to track the role of features in trees. The results will be evaluated and interpreted via an in-depth linguistic investigation, at three different scales of variation and change: Meso-variation (variation across different genera and languages of the same family, e.g. the properties that characterize Romance, as opposed to Greek, Albanian, Armenian), micro-variation (variation across genealogically related dialects, e.g. Greek dialects spoken in Asia Minor, Italo-Greek, Tsakonian Greek, and languages belonging to the same genus, e.g. Romance, Germanic, Slavic), and macro-variation (variation across families in different areas of the world).