A multifactorial corpus analysis of grammatical synonymy
In the present study, both monofactorial and multifactorial corpus methods are applied to the alternative use of the analytic adpositional construction and the synthetic case construction in present-day written Estonian. 600 examples from the fiction sub-corpora of written Estonian were annotated for 16 different (seven morphosyntactic and nine semantic) factors. In order to determine which factors are most influential in determining the choice between the two constructions, a logistic regression model was built to fit the data. The analysis confirmed the statistical influence of the following factors: lexical complexity, type of verb, word order, word class, and mobility. The results reported in this study align with previous research, which has shown that case affixes are used to express more abstract relations and adpositions more concrete ones.