ABSTRACT:
Spanish is split in two dialects with respect to agreement in Existential Constructions with haber. In Spanish Dialect I (SI) the verb has default agreement (3er person, singular), but in Spanish Dialect II (SII), the verb agrees with the internal nominal, despite the fact that the nominal is Accusative. Additionally, nominals marked for [person] are not allowed under haber, not even with list-readings. I propose that Spanish existential constructions with haber have a vP that accounts for the presence of ACC. The small v has only [number]. This explains the ban on nominals with [person] in these constructions, under the assumption that phi-incomplete heads cannot value the Case of phi-complete nominals. In the dialect where the nominal agrees with the verb (SII), T has an interpretable [person] feature; given that, T can probe v in this dialect, producing the effect of object-agreement. This is possible because T has only one uninterpretable feature, namely [number], as small v does. With the same mechanism, a higher T associated with a raising verb can also probe v, and then it will agree with the internal nominal too. In SI, small v (which is phi-incomplete) cannot value the ö-features of T, which is phi-complete; therefore, in SI T needs to resort to another strategy to value its phi-features, which will result in a default value.
