One reason for assuming that the phrase headed by N2 is not a DP is that it can never be preceded by a determiner (an article or a demonstrative/possessive pronoun). This is shown in (99a&b) for both count and non-count nouns. In contrast, the primed examples give the corresponding partitive constructions in which the projection of N2 acts as a full DP; cf. Section 19.4 for a discussion of this construction.

One might think that the unacceptability of the primeless examples is due to the fact that the determiners make the projection headed by N2 definite, but indefinite articles cannot be used either, as shown in (100a); compare this QBC with the partitive construction in (100b), in which the indefinite article must be expressed.

Another reason for assuming that N2 does not head a DP is that substituting a pronoun for the projection of N2 yields an unacceptable result. The unacceptability of (101a&b) is not conclusive, since we are dealing with definite pronouns. However, example (101c) shows that existential quantifiers are also excluded. For comparison, the primed examples give the corresponding partitive constructions; note that van het/zeof it/them is not possible, but this is due to the general rule that replaces the inanimate pronouns het/zeit/them with the R-pronoun er in this syntactic context.

Although the above examples support the idea that N2 does not head a DP, we cannot conclude that N2 is a bare noun. This is clear from the fact that it can take an argument, as shown for the relational noun vriendjesfriends in (102a). Moreover, N2 can be modified by all kinds of modifiers: an attributive adjective in (102b), an appositive phrase in (102c), an adjunct PP in (102d), and a restrictive relative clause in (102e).

For completeness’ sake, note that N2 can be modified by an adjective in the positive or comparative form, but not by an adjective in the superlative form. This may be due to the fact that noun phrases containing a superlative are definite: de/*een aardigste studentthe/a nicest student. Finally, note that a pseudo-superlative such as alleraardigste could be used, but these do not necessarily trigger a definite interpretation; cf. een alleraardigste studenta very nice student.

Cardinal numerals and quantifiers cannot precede N2. However, this has no bearing on the size of the projection of N2, since a plausible explanation for the impossibility of (104) can be found in the fact that they are in the scope of the N1 paar, which also has quantifying force; cf. *veel vijf studentenmany five students.

The claim that N2 heads a projection that is somewhat smaller than a DP can also be supported by evidence involving initial coordination, i.e. coordination by correlative coordinators like of ... of ...either ... or ... and zowel ... als ...both ... and ...; cf. C38.4.2. In the primeless examples in (105) the two conjuncts each include an article, so we can safely conclude that we are dealing with full DPs and we see that the result of initial coordination is fine; in the primed examples, on the other hand, we are dealing with the smaller phrases oude mannenold men and oude vrouwenold women, and the result of initial coordination is unacceptable.

When N2s head a phrase that is smaller than a complete noun phrase, we predict that initial coordination of phrases headed by such nouns is impossible. The primed examples in (106), with the correlative coordinator zowel ... als ..., show that this expectation turns out to be correct for quantifier nouns and collective nouns. Note that it is not coordination as such that causes unacceptability, as the primeless examples with the conjunctive coordinator enand are perfectly acceptable.

The primeless examples in (107) show that the projection headed by N2 can never be moved independently of N1; the noun phrase consisting of N1 and N2 cannot be split. The primed examples show that the same holds for numerals and quantifiers: standard Dutch does not allow this so-called split topicalization construction.

That the judgments on the primeless and the primed examples are related is clear from the fact that those dialects that allow the primeless examples also allow the split patterns in the primed examples; we refer the reader to Coppen (1991), Vos (1999), and Van Hoof (2006/2017) for a discussion of split topicalization of this kind.

That the phrase headed by N2 and the nominal projection following a numeral sometimes exhibit similar behavior is also clear from the fact already discussed in Section 17.1.1.3, sub IVA, that both can be replaced by quantitative er if N1 is a quantifier or measure noun, as shown by the acceptability of the examples in (108) in a conversation about penguins. This again shows that the projection of N2 is smaller than DP, given that DPs cannot be replaced in this way.