Exploring probabilistic differences between genetically related languages

The study proposes an analytical pipeline for quantitative corpus-based research in the domain of contrastive linguistic analysis. Specifically, it is proposed that a combination of conditional inference tree models for data exploration combined with subsequent statistical modeling via generalized linear models can help reveal and describe interesting contrasts between the languages under investigation. The main goal of this paper is to provide an approachable introduction into the rationale underlying these techniques and to illustrate their application on the basis of a contrastive analysis of English and German relative clause constructions.