This section concludes the discussion of nominal complementives with a look at of copular constructions such as (126), which have given rise to a debate about whether the neuter pronoun het/dat/dit or the noun phrase aardige jongens functions as subject of the construction. In our discussion below, we will use examples headed by the copula zijn 'to be' , but such constructions also occur with other copulas like worden 'become' or blijven 'to stay' .

Bos (1961), following De Groot (1949:153), argued that it is the noun phrase that functions as subject and the pronoun that functions as predicate. One reason is that the finite verb in the primed examples in (126) agrees in number with the plural noun phrase, not with the singular pronoun; the examples in (127) show that plural agreement is normally not possible if the pronoun het/dat/dit functions as subject of an adjectival or prepositional copular construction.

Another reason for assuming that the pronouns function as predicates is that if the nominal predicate is replaced by a pronoun, the latter will appear in its nominal form; this cannot, of course, be shown for the second person, plural pronoun given that the subject and object form are identical.

      The claim that the pronouns in (126) function as predicates of the copular constructions was challenged by Merckens (1961), who argued that these examples are actually ambiguous. This is illustrated by the examples in (129), the meanings of which will become clear from the English translations. The function of the left-dislocated constituents in these and the following examples is to force the intended subject/predicate reading on the pronoun dat.

The same ambiguity is apparent from embedded clauses like those in (130). given that subject pronouns are normally right-adjacent to the complementizer and predicates left-adjacent to the verbs in clause-final position; cf. Ik vind dat dat mooi is'I think that that is beautiful' versus *Ik vind dat mooi dat is. This means that dat functions as a subject in (130a) and as a predicate in (130b). The primed examples show that this conclusion is supported by the interpretation of these examples.

A similar contrast can be found in yes-no questions, where the subject pronoun normally appears right-adjacent to the verb in sentence initial position: Is dat mooi?'Is that beautiful?' versus *Is mooi dat?. This means that dat functions as a subject in (131a) and as a predicate in (131b). The primed examples show again that this conclusion is supported by the interpretation of these examples.

A piece of indirect evidence not mentioned by Merckens in favor of the claim that we are dealing with ambiguous structures is provided by the vinden-constructions in (132). Given that the complementive always follows its logical subject in the middle field of the clause, the fact that both orders are possible in (132) show that dat may function either as the subject or as the complementive of the construction; again this is supported by the interpretations of these examples.

A final piece of evidence involves pronominalization. Consider the discourse chunk in (133), in which participant B is backing up participant Aʼs claim that Jan and Piet are nice boys. In Bʼs reaction the noun phrase aardige jongens is not replaced by the referential personal pronoun zij, as would be expected if this noun phrase were the subject of the sentence, but by dat, as would normally be the case when we are dealing with a predicate. Observe that the copula is plural in this case despite the fact that normally the pronouns het and dat are both syntactically singular.

The discussion above has conclusively shown that Bos’ claim that the neuter pronouns in the copular constructions in (126) can only function as the predicate of the construction cannot be maintained; the structures are syntactically ambiguous in the sense that the pronoun can function either as the subject or as the predicate of the copular construction. In fact, there is even reason for assuming that, due to its sentence-initial position, the pronoun het'it' in example (126a) must be interpreted as the subject of the copular construction. The reason for this is that sentence-initial het normally functions as subject; the object pronoun het, for example, cannot be topicalized as will be clear from the grammaticality contrast between Hij zag het'He saw it' and *Het zag hij. The (b)-examples in (134) show that the ban on topicalization also holds for het if it functions as a complementive; whereas the primeless example is ungrammatical with, the corresponding primed example without topicalization is fully acceptable (although perhaps less preferred than its counterpart with dat, which is indicated here with a question mark.

      The remainder of this section will pinpoint some special properties of the copular construction under discussion. We have already seen in (126) and (127) that the predicate must be nominal; adjectival and prepositional complementives are excluded. The contrast between the primeless and primed examples in (135) show, however, that it is not the case that any nominal predicate can be used: whereas the predicate can either contain an article or be bare if the subject is a noun phrase or a regular pronoun, the pronouns het, dat, and dit require that an article be present.

The examples in (127) show again that although the pronouns het, dit and dat are syntactically singular, the examples in (136) can be used to refer to sets of entities.

Although the pronouns het, dit and dat are syntactically neuter, they can be used to refer to non-neuter antecedents. This is already clear from example (136a) but even more conspicuous in examples such as (137a) where the predicate does agree in gender with the antecedent of the pronoun dat. For completeness’ sake, (137b) provides an example in which the antecedent differs both in number and gender from the pronoun dat.

In the examples above, the antecedent of the neuter pronoun is referential in the sense that it denotes a (possibly singleton) set of entities. The examples in (138) show that the antecedent can also be generic, although it seems that the indefinite generic noun phrase in (138c) is somewhat marked.

The markedness of (138c) may be related to the fact that examples in which the antecedent of the neuter pronoun is quantified are also marked; whereas the generic example in (139a) is fully acceptable, the corresponding quantificational construction in (139b) is degraded.