This study is concerned with phi-features, their nature and the role they play in the grammar of Dutch. The set of phi-features traditionally includes the categorical features plus/minus N and plus/minus V (defining the lexical categories Noun, Verb, Adjective and Preposition), the pronominal features of person, number and gender (involved in agreement relations), the binding features plus/minus anaphoric and plus/minus pronominal, the Case features (defining the distinct Cases such as Nominative and Accusative), and grammatical features such as plus/minus wh (yielding morphonologically marked subsets of pronouns). The aim of this study is to show that phi-features are among the major building blocks of the syntactic system, and that they are crucially and prominently involved in shaping both the form and the content of different levels of representation and the nature of their relationship. Thus the main body of the text serves to demonstrate in more or less detail that what appears to be a quite heterogeneous and rather desperate collection, has more structure, and is far more homogeneous than usually thought. As a matter of fact, the phi-features will turn out to be drawn from a small set, which is probably universally available. And instead of being distributed at random over the lexicon, these features, c.q. the categories they constitute, are subject to the ordinary principles of grammar.