The examples in (381) show that the two interrogative personal pronouns can function as both subject and object; this shows that they do not exhibit case distinctions.

When the pronoun is part of a PP, the behavior of the two interrogative pronouns diverges: while wie can easily occur as the complement of a preposition, wat behaves like the referential personal pronoun hetit in that it triggers R-pronominalization. This is most likely related to the semantic distinction with respect to the feature [±human].

The restriction that watwhat cannot occur as the complement of a PP does not apply to echo questions like (383a), where the question word is stressed. In echo-question contexts, wat can also be used as a request to repeat/clarify a previous utterance; example (383b) shows that in this case wat sometimes alternates with the form watte.

The primeless examples in (384) show that wie and wat can also function as nominal predicates in copular constructions. However, this is not possible in the vinden construction, as shown in the primed examples.

The unacceptability of (384b') is probably not due to the fact that wat cannot be used as a nominal predicate in the vinden-construction but to the fact that the predicate in this construction must express a subjective evaluation; this is suggested by the fact that example (385b) is markedly better than (384b').

What is rather special is that wat in the vinden-construction in (385b) is typically used to question a property expressed by adjectives, as in (386a). Note, however, that open questions would normally be phrased slightly differently: with the PP van hem instead of the object hem, as in (386b).

A final special use of wat is illustrated in (387). Here it is not questioning an argument or a predicate, but a measure phrase, which is obligatory but behaves in various ways like an adjunct, for which reason it is often considered a quasi-argument of the verb; cf. Section V2.4 for further discussion.

The examples in (388) show that the interrogative pronouns are formally third person and masculine (or neuter): this is clear from the form of the finite verb and from the fact that the third-person possessive pronoun zijnhis can take the interrogative pronoun as its antecedent. Note, however, that the use of feminine pronouns is promoted for politically correct writing/speech.

However, the use of wie does not express the speaker’s presupposition that the answer will involve a third-person constituent or a male person, as will be clear from the fact that a speaker can easily ask (rhetorical) questions like those in (389). The presupposition regarding a third-person constituent does hold in the case of wat, but this may be due to the fact that this pronoun refers to inanimate things, whereas the speaker and the addressee are typically human.

The pronouns wie and wat genuinely differ in number: the form of the finite verb in (390a) shows that the [+human] pronoun wie can be formally either singular or plural; the fact that the [-human] pronoun wat in (390b) is only compatible with the singular form of the finite verb shows that it is formally singular. This does not mean that wat cannot be used to question more than one thing: an answer to (390b) can easily involve a list of objects. The fact that the quantifier allemaal can be used in (390b) also shows that wat can be semantically plural; cf. Zij zijn allemaal ziekthey are all ill versus *Hij is allemaal ziekHe is all ill.

Example (391a) shows that using the quantifier allemaal leads to a marked result when the [+human] pronoun wie triggers singular agreement on the verb. However, it is not hard to find fully acceptable cases like this on the internet. For example, singular agreement is regularly found with the verb komento come in (391a'). Perhaps this is related to the fact that this verb can take a secondary predicate in the form of a PP, since copular constructions like (391b&b') are also clear exceptions to the general tendency to avoid singular agreement in the presence of allemaal.

For completeness’ sake, consider example (392a). This example is not a counterexample to our earlier claim that wat cannot trigger plural agreement on the finite verb, because it is part of a larger subject phrase wat voor een dingenwhat for a things, which happens to be split. That this larger phrase must trigger plural agreement is illustrated in (392) with the unsplit variant of the wat-voor phrase.

The fact that the examples in (390) and (391), in which the pronoun functions as the subject of the clause, contain the expletive erthere shows that the interrogative pronouns are indefinite: if the expletive is omitted, the result is unacceptable, unless some presuppositional constituent is present; cf. (381a) and (388a).

This also holds for wie in the partitive-like construction wie van julliewho of you in (394b).

The unacceptability of the examples without the expletive may be due to the fact that the pronouns wiewho and watwhat are typically not D-linked, i.e. not related to the domain of discourse, which is also referred to as non-specific in the literature on Dutch; cf. Section 21.1.2, sub I. Some speakers reportedly allow the examples examples in (393) and (394) without the expletive; cf. Wesseling (2018: 102-3) for data based on an acceptability judgment test. One possible explanation is that these speakers more easily accept a D-linked reading for wiewho and watwhat: insofar as the questions without er are intelligible, the speaker seems to prompt the addressee to select one or more entities from a preestablished set in the discourse. We will leave this issue for further research.

The examples in (395) show that the pronouns wie and wat can be modified by elements like dan ook or om het even. However, this results in the loss of their interrogative force: the meaning of these phrases comes close to that of English phrases with any. Perhaps this is not so surprising for wat, since we will see in Section 18.2.1.3 that this pronoun can also be used as a quantificational personal pronoun, but it is surprising for wie, which lacks this option.

Interrogative personal pronouns do not easily allow for other forms of modification. For example, the examples in (396) involving postmodification are marginal at best.