This paper discusses one empirical domain where a transitive vP, though being a full propositional construction, does not form a phase. Drawing data from the past transitive structures, aka ergative constructions, in Kurdish, I will demonstrate that the interface-oriented arguments underlying the determination of the first phase (v*P vs. vP) fall short of accounting for the emergent case/agreement pattern in these structures. Instead, à la Chomsky (2008), a view will be endorsed of the first phase in which a vP constitutes a strong phase only if its head is associated with a full complement of φ-features, regardless of whether an external θ-role is projected in [spec,vP]. Thus construed, the identification of a strong phase need not be procrastinated until the whole vP is formed, rather the status of a phase will be decided once a φ-complete/-defective v° enters into the derivation.
