The Evolution of Noun Phrases in Indo-Iranian: empirical foundations and theoretical modeling'

The project aims at tracing the evolution of noun phrase structures in the Indo-Iranian branch of Indo-European in order to study the mechanisms, triggers, and causes of type transitions in close detail. We choose this branch because it offers a uniquely rich dataset: we have access to several layers in the history of the branch, going back over 3000 years, and can therefore trace changes with textual evidence; NPs have heavily diversified in the branch, incorporating many innovative structures (e.g. ezafe in the West, numeral classifiers in the East) and changing back and forth between options (e.g. in recursive noun compounding); and the branch has been in contact with highly diverse languages (e.g. Semitic and Tibeto-Burman) and equally diverse sociolinguistic conditions.

We will track the evolution of several key aspects of NP structure and probe the evidence for  contact effects, possibly universal principles and chance processes, using both qualitative comparative methods and quantitative modeling. These aspects include, but are not limited to, the evolution of the ezafe construction, numeral classifiers, definiteness marking, adnominal clausal and adpositional modifiers; the various attritions of case and agreement morphology and the consequences thereof for NP expandability and differential case marking; and overall changes in configurationality (tightly integrated vs. disintegrated NPs).

In order to assess contact effects, we include a rich sample of languages that were or are in contact with Indo-Iranian; for assessing universal effects, we will draw on on existing databases enriched by further data that we will collect.

Expected results include a series of articles devoted to specific aspects on NP evolution as well as a synthesis study. In all cases we will combine qualitative and quantitative methods with the goal of establishing the probabilities with which specific structural traits change under varying conditions. In addition, we will use our empirical results to update computational models of trait evolution and assess the extent to which these updates improve their ability to reconstruct known ancestral states (e.g. in Vedic or Avestan).

Funding

Swiss National Science Foundation (SNSF), 1.9.2017-31.8.2019