Relative clauses serve to provide information about an entity introduced by a noun (phrase) in the matrix clause, the antecedent. They are obligatorily introduced by (a phrase containing) a relative element coreferential with this antecedent and contain a trace indicating the original position of the relative element. The overall structure of relative clauses is therefore as indicated in (82); see De Vries (2002: appendix III) for an overview of different syntactic analyses of relative clauses.

For instance, the relative pronoun dat in the examples in (81), repeated here as (83), is coreferential with its antecedent boek'book', and the trace ti indicates the original position of the pronoun in the relative clause.

      Relative clauses can be either restrictive or non-restrictive. Like other restrictive modifiers, restrictive relative clauses form an intonation unit with their antecedent, while reducing the referent set of the noun phrase by restricting the denotation of the noun. Given this, it is plausible to assume that, like other restrictive modifiers, restrictive relative clauses are part of the NP-domain: in (84), this is expressed by placing the relative clause within the NP brackets and by coindexing the relative element with some projection of the head noun smaller than NP (which may contain other restrictive modifiers).

Non-restrictive relative clauses are separated from the noun by an intonation break, represented in the examples in (85) by means of commas, and merely provide additional information about the referent set of the full noun phrase, which can be established independently of the relative clause. Given this, we assume that non-restrictive relative clauses are external to the NP, and that the relative element is coindexed with the NP; see Section 3.3.2.3.1 for details.

The abstract structure in (85) shows that we assume that the non-restrictive relative clause is still part of the DP-domain; the reasons for this are given in 3.1.2, sub II. Since many researchers have argued that besides the NP- and DP-projection, other intermediate projections can be distinguished within the noun phrase, such as NumP discussed in Section 1.1.2, sub IIA, example (6), it might be the case that the relative elements of non-restrictive relative clauses are coindexed with some higher projection in between NP and DP. The main issue here is, however, to indicate the scope of the two types of relative clauses: a non-restrictive relative clause provides additional information about the referent of the noun phrase, and since the reference of the noun phrase is determined by the determiner, the relative clause must have this determiner in its scope; a restrictive relative clause does not provide information about the referent of the noun phrase but restricts the denotation of the noun, and it is therefore plausible that it is only the NP that falls inside its scope
      Since this section is mainly concerned with postmodification of nominal heads, the discussion will focus on relative constructions of the type in (84) and (85) above and semi-free relatives like (86a); free relative constructions such as (86b), in which an overt nominal antecedent is lacking, will not be extensively discussed.