Quantifying syntactic influence: Word order, possession and definiteness in Old Church Slavonic and Greek
Hanne Martine Eckhoff · 2018
Canonical Old Church Slavonic (OCS) is the earliest attestation of Slavonic and our only evidence for many Common Slavonic features. Since virtually all of the texts are translations from Greek, separating Greek syntactic features from native Slavonic ones is not straightforward. This article sets out to examine how we can distinguish between native syntax and syntactic influence using statistical modelling (classification trees) and parallel data in two case studies, both extreme cases. The first case study deals with OCS word order, which is known to follow the Greek to a very great extent. The second one concerns adnominal possessive constructions, where OCS displays great independence and translates Greek adnominal genitives into at least five different constructions. In the first case study, we see that Greek word order outranks all other predictors in the statistical model. In the second case study we see that Greek morphosyntactic predictors are outranked by a number of semantic and pragmatic factors. The final classification model also provides a better understanding of the relative weight of these factors than the literature has previously provided. The differences between the results in the two case studies are encouraging: Similar models can be used to weigh language-external against language-internal factors in the less straightforward cases in between those two poles, if we have sufficiently large and sophisticated data sets.