With respect to agreement with the finite verb, the partitive genitive construction behaves like the nominal part of the construction in isolation: it triggers singular agreement. This is illustrated in (23).

The presence of the expletive er'there' in example (23) from the previous subsection shows that both the quantificational pronoun iets and the partitive genitive construction can be nonspecific (cf. also the discussion of (21a)); in general, only nonspecific indefinite subjects license expletive er. If expletive er is not present, the partitive genitive construction is given a specific or a generic interpretation, just as is the case with other indefinite subjects. This is illustrated in (24): on its specific interpretation in (24a), which is marginal for some speakers, the partitive genitive construction refers to a certain exciting thing known to the speaker; on its generic reading in (24b), which is fully acceptable for all speakers, the sentence expresses that any exciting thing would be welcome.

Like other noun phrases, partitive genitive constructions can be modified by means of a relative clause. The examples in (25) show that relative clauses can also modify the quantificational pronoun (n)iets, that is, the partitive genitive adjective need not be present.

Note in passing that it is not the case that all clauses that follow the partitive genitive construction modify the noun phrase as a whole. Comparative dan-clauses, for example, follow the noun phrase as well, but since they are dependent on the comparative -er morpheme on the adjective, the adjective is obligatorily present. Something similar holds for intensifying phrases licensed by the modifier te.

The fact that the partitive genitive construction may occur in most regular NP-positions has already been illustrated for the subject and object positions in (23) and (21), respectively. In (27), this is illustrated for the prepositional object position: (27a) involves a prepositional complement of the verb and (27b) an adverbial adjunct.

Partitive genitive constructions are not readily used as indirect objects, which is related to the fact that indirect objects generally refer to +animate entities or institutions, whereas the partitive genitive construction generally refers to -animate entities; cf. the discussion of the examples in (57) to (62). Some relatively acceptable examples are given in (28), in which the partitive genitive construction refers to some sort of institution.

Given the similarity in syntactic behavior and distribution between indefinite noun phrases and the partitive genitive construction, it seems safe to conclude that the quantificational pronoun iets is the head of the complex construction. This of course coincides with the fact that the partitive genitive adjective cannot be used in isolation in the NP-positions in (21) to (27); the noun iets is obligatorily present in these examples (see the discussion of (4)). The fact that the partitive genitive construction is headed by the indefinite noun does not imply, however, that it is always possible to omit the adjective; omitting the adjective in (29a&b), for example, causes the example to be less acceptable under the intended reading. The fact that the two (c)-examples are fully acceptable shows, however, that something special is going on in the (a)- and (b)-examples.

The difference in acceptability between (29a'&b') and (29c') is probably related to the presence of the sentential negation in the first two examples. The examples in (30) show that the quantificational pronoun iets'something' normally cannot be preceded by the sentential negator niet'not'; instead, the negative pronoun niets'nothing' is used.

If we apply the same merging rule to (29a'), we get the result in (31a). However, this structure expresses constituent negation, and not sentential negation. In order to express sentential negation, the negative noun phrase must be moved into the position that is otherwise occupied by the negative marker; cf. the discussion of the examples in (67) in Section 2.3.1, sub IIB2. This can be done can by applying R-pronominalization and R-extraction. This results in the acceptable structure in (31a'), and for the same reason, (29b') surfaces as (31b).

The fact that the sentential negator is possible in (29a&b) may be related to the fact that R-pronominalization is not possible if the pronoun is part of a larger phrase: the examples in (32) are therefore ungrammatical, and this leaves us with (29a&b) as the only means to express the intended meanings. This may account for the contrast in acceptability between (29a'&b') and (29c').