The Inverse Agreement Constraint in Uralic Languages

Abstract

The paper aims to answer the question why object–verb agreement is blocked in Hungarian, Tundra Nenets, Selkup, and Nganasan if the object is a first or second person pronoun. Based on Dalrymple & Nikolaeva (2011), it is argued that object–verb agreement serves (or served historically) to mark the secondary topic status of the object. The gaps in object-verb agreement can be derived from the Inverse Agreement Constraint, a formal, semantically unmotivated constraint observed by Comrie (1980) in Chukchee, Koryak and Kamchadal, forbidding object-verb agreement if the object is more ʻanimate’ than the subject: The paper claims that the Inverse Agreement Constraint is a constrainton information structure. What it requires is that a secondary topic be less topical than the primary topic. An object more topical than the primary topic can only figure as a focus. A version of the constraint can also explain why Hungarian first and second person objects have no accusative suffix, and why accusative marking is optional in the case of objects having a first or second person possessor.