András Bárány's website

Inverse agreement and global case splits

I wrote my dissertation as a member of the ReCoS project project which explored a new approach to parametric variation in syntax. You can download the dissertation here or buy (or borrow from your library) my OUP monograph based on it.

My dissertation and my related monograph deal with several related phenomena regarding the interaction of case and agreement across languages. The first part focuses on differential object agreement in Hungarian. In Hungarian, the finite verb always agrees with the subject of a sentence but sometimes it also agrees with the object. One puzzle in this case is the trigger of object agreement has been argued to be syntactic or semantic in previous literature; I argued for a hybrid approach: the syntactic structure and the semantic properties of the object are crucial.

Another puzzle concerns the distribution of object agreement with personal pronouns. I argued (building on work by Katalin É. Kiss) that Hungarian shows an “inverse agreement” system: the morphological realisation of agreement depends on the features of the subject and the object. Inverse patterns are those in which the object's features are more prominent than the subject's, for example with a first person object and a third person subject. In some languages, notably Native American languages, the verb spells out a so-called “inverse marker” in such configurations. I argued that in Hungarian, in contrast, the verb only shows subject agreement in such inverse contexts.

In later chapters, I put the discussion of Hungarian object agreement in a cross-linguistic context. I argue that the phenomenon of “global case splits” represents a dependent-marking counterpart to inverse agreement. In global case splits, case-marking of the subject or the object of a sentence depends on properties of the both arguments, not just the argument that is case-marked (or not). Rather than spelling out an inverse marker (or nothing, as in Hungarian) in inverse configurations, the verb assigns a special case-marker to the one of its arguments in global case splits.

Global case splits are also interesting because they suggest that case-marking can be a consequence of agreement rather than agreement (or not) being a consequence of case-marking. I also address how both of these directions of interaction can be modelled in grammar.

Finally, I sketch the idea underlying my current project about ditransitive constructions and discussed some ideas of parameter hierarchies related to case and agreement.