DP-internal double agreement is not double Agree: Consequences of Agree-based case assignment within DP

The German possessor doubling construction is remarkable in that there is double agreement: the possessive pronoun agrees both with the possessor as well as the possessee in phi-features. We will argue that double agreement is not to be interpreted as resulting from two syntactic Agree operations: Positing two such operations involving the possessive pronoun leads to insurmountable technical problems; rather, we will introduce novel empirical evidence showing that the agreement between possessor and possessive pronoun is anaphoric rather than grammatical in nature. While this solves the double agreement problem, the dative case on the possessor still needs to be accounted for. We will propose that dative case does not come from the possessive pronoun or from a preposition as in the Predicate Inversion approach, but rather from N. This not only avoids an overgeneration problem resulting in case assignment to the wrong goal, but also opens up the possibility to unify case assignment in Germanic with possessor agreement in languages like Turkish.