What triggers the Hungarian objective paradigm? A structural and feature-based account

I argue that neither purely structural (e.g. Bartos 1999) nor purely semantic approaches (e.g. Coppock 2013) to the distribution of Hungarian verb paradigms can account for the observed data. I propose, following Bartos (1999) that the structure of the direct object (DO) noun phrase is crucial, while certain semantic features of the DO also have to be taken into account. I claim that the role of [DEF] and the semantic effects observed by Coppock (2013) are mostly correct but that her analysis makes false predictions regarding possessive structures. Taking into account both structural and interpretative factors, I propose an approach covering a wider range of attested data.