In this thesis, I propose that the distribution of nominal phrases is constrained by the relative positions of phi-features that nominal phrases contain. More specifically, I propose a condition according to which two syntactic nodes bearing visible phi-features cannot be directly merged. This constraint results in the well-described generalization according to which Agreement is severely limited in the Nominal Domain and complements to nominals may not bear direct case-marking, unlike the complements of verbs and prepositions. Next, I suggest that the propagation of phi-features can be blocked by a formal feature – epsilon – that is borne on such morphemes as prepositions and linkers, as well as oblique case markers. A nominal phrase merged with an epsilon may combine with another nominal phrase when a bare nominal phrase may not. Next, I argue that the syntactic identity of nouns can be fully reduced to phi-features. The proposed approach is shown to extend to adjectives – another category whose syntactic behavior is largely determined by the presence of phi-features. To that effect, I show that the ban on direct marking of the complements of both adjectives and nouns can be understood as a constraint on the merger of two phi-bearing structures. I show how the proposed theory can account for several well known syntactic phenomena, including the ban on double Absolutives in Ergative languages and the lack of Structural Dative marking in the Nominal Domain. The empirical data is mainly drawn from Slavic, Romance, Germanic and Iranian languages of the Indo-European family and from Nakh-Dagestanian languages.