Table 15 shows that the form of the D-linked demonstrative modifiers depends on the nominal features of the head noun. The pair deze/die has the same distribution as the definite article dethe; these demonstratives are used in singular non-neuter and plural noun phrases. The pair dit/dat has the same distribution as the definite article het; these demonstratives are only used in singular neuter noun phrases. Both pairs can also be used with non-count nouns; since non-count nouns do not have a plural form, the choice in these cases depends only on the gender of the noun.

The choice between the non-D-linked demonstratives zo’n and zulke depends on number: zo’n is used with singular nouns (although, according to De Rooij 1989, some Dutch dialects also allow it with plural nouns), whereas zulke requires a plural noun. Non-count nouns always take zulk(e), the inflectional ending -e depending on the gender of the head noun in the same way as attributively used adjectives. Note that besides zulke wijn/zulk bier it is also possible to have zo’n wijn/bier; since it is also possible to have zulke wijnen/bieren, it seems plausible that the nouns in zo’n wijn/bier actually function as count nouns denoting types of wines/beers (cf. Section 15.2.2.1, sub III), although it must be admitted that the difference in meaning between the two singular cases is not easy to discern.

We have already mentioned that the Dutch tradition refers to zo’nsuch a and zulk(e)such as indefinite demonstratives, suggesting a similarity in meaning to the indefinite articles eena and Ø. Note, however, that DPs headed by these demonstratives do not refer in the same sense as a DP headed by an indefinite article: whereas an indefinite noun phrase like een vrouw introduces a new entity into domain D or refers to an entity unknown to the addressee, a noun phrase like zo’n vrouw rather evokes a sense of comparison, which is clear from the fact that it can be paraphrased as “a woman like that”. Another way of expressing this would be to say that a noun phrase like een vrouw refers to a token, whereas zo’n vrouw refers to a type. Since DPs headed by the demonstratives zo’n and zulke do not refer to expressions in the same sense as DPs headed by an indefinite article, using the term indefiniteness can be confusing. For this reason, we will not use this term here: instead, DPs headed by demonstratives like deze/die and dit/dat will be said to be D-linked (i.e. linked to the discourse), while DPs headed by demonstratives like zo’n and zulke will be said to be non-D-linked (not linked to the discourse).

Although the non-D-linked demonstrative modifiers zo’nsuch a and zulk(e)such are considered determiners in traditional grammar, it should be noted that zo’n is a contracted form of zo een. Therefore, it could also be argued that what we have here is the indefinite article eena, which is premodified by the adverb zoso.

Demonstrative pronouns are typically used deictically, i.e. they refer to a referent that is physically present in the situation in which the utterance is made. The proximate and distal demonstratives indicate different relative distances between the referent of the noun phrase and the speaker; the proximate ones indicate that the referent is close to the speaker, whereas the distal ones indicate that the referent is more distant from the speaker. The notion of distance can be interpreted literally and metaphorically; cf. Alexiadou (2007:100-1) and the references cited there. In examples such as (506), involving literal distance, the relevance of relative distance can be emphasized by modifying the noun phrase with the locational pro-forms hier and daar. The former is more readily used with the proximate demonstratives, while the latter is preferably used with the distal ones.

Note that the pro-forms are usually at the right edge of the noun phrase, but given that they are related to the reference of the noun phrase, it seems plausible that they directly modify the demonstrative. Note in passing that some Dutch dialects may have the pro-form in initial position of the noun phrase (cf. Corver et al. 2013:133) and that Afrikaans has the grammaticalized demonstratives hierdiethis and daardiethat, which are apparently formed by combining the Dutch demonstrative die and the locational pro-forms; cf. Donaldson (1993:142-4). We also refer the reader to Alexiadou (2007:117) for similar observations in Swedish and non-standard English, which may be different in that the pro-form immediately follows the demonstrative; cf. This here dog is a good hunter and That there cat has been with me fifteen years in certain southern U.S. dialects (Carole Boster p.c.).

The notion of distance can also be interpreted in temporal terms. The noun phrase deze weekthis week in (507a), for example, will normally include the here and now of the speaker, which is clear from the fact that the past tense leads to an infelicitous result. The noun phrase die weekthat week, on the other hand, will normally be contextually determined. For example, in a report about the 1953 flood in Zeeland, the noun phrase die week will refer to the week in which the flood occurred, which can be emphasized by the more specialized form diezelfdethe very same.

The notion of distance can also be understood more metaphorically as “relatedness” to the speaker. There seems to be a tendency for the speaker to use the distal demonstrative when referring to an object that belongs to the addressee and the proximate demonstrative when the speaker is the owner. Thus, if two people are sitting at a table with a book on it, the speaker will tend to use the distal demonstrative when he asks the addressee for permission to browse the book, as in (508a), but the proximate demonstrative when he grants the addressee permission to browse the book, as in (508b).

However, other considerations can easily override this tendency. For example, if the speaker is already holding the book, he is more likely to use the proximate pronoun to ask permission to browse the book, and if the addressee is already holding the book, he is more likely to use the distal pronoun to grant permission.

In their deictic use, demonstratives are typically used to partition the denotation set of the modified head noun. This is particularly clear in contrastive contexts such as (509), where the speaker explicitly refers to two subsets of books, but the same is true in non-contrastive contexts (although in such cases the evoked alternative referent set may be empty, as in the context sketched for the examples in (508)).

Non-interrogative demonstrative pronouns can be used as arguments, in which case the referent of the demonstrative is completely determined by the context. The form of the D-linked demonstratives is determined by the same factors as the modifiers in Subsection I: ditthis and datthat are singular and refer only to entities that would normally be referred to by a neuter noun phrase; dezethis/these and diethat/those are either singular, in which case they refer to entities that would normally be referred to by a non-neuter noun phrase, or plural. This is illustrated in Table 17, where the demonstratives function as the subject of the clause, so that their number can be determined by checking the number of the finite verb.

		singular	plural
non-neuter	proximate	Deze is leuk. this one is nice	Deze zijn leuk. these are nice
	distal	Die is leuk. that one is nice	Die zijn leuk. those are nice
neuter	proximate	Dit is mooi. this one is beautiful	Deze zijn mooi. these are beautiful
	distal	Dat is mooi. that one is beautiful	Die zijn mooi. those are beautiful

In question-answer pairs, the neuter D-linked demonstratives dit and dat can (optionally) have an -e ending in spoken language. These forms cannot be used as modifiers and are only used deictically, e.g. when pointing at the object in question. A similar “inflected” form is possible with the first-person singular personal pronoun ik; Wie is daar? Ik(ke)Who is there? Me.

The fact that the demonstratives in Table 17 are rendered in English by appealing to the singular pro-form one suggests that the Dutch examples contain an empty noun with the same function as English one. This suggestion is supported by the fact, illustrated in (512a&b), that an attributive adjective can follow the demonstrative; cf. Section A28.4 for a more detailed discussion of this kind of reduced noun phrase. Note, however, that the neuter singular demonstratives in (512c) do not have this option: for example, the neuter noun in dit/dat grote boekthis/that big book cannot be omitted. The nature of this contrast between deze/die and dit/dat is unclear.

When used deictically, the forms in Table 17 mainly refer to [-human] entities. Using these demonstratives to refer to a person generally results in a pejorative connotation: a speaker uttering examples such as (513) leaves no doubt that he does not think highly of the person he is referring to. The neutral (non-pejorative) counterparts of the examples in (513) involve a referential personal pronoun.

This pejorative meaning aspect is absent when these demonstratives are used anaphorically to refer to a person, i.e. when the referent is mentioned in the discourse immediately before the use of the demonstrative, as in (514). We will return to this use of the demonstrative in Section 19.2.3.2, sub II.

Non-D-linked demonstratives can also be used as arguments. Note, however, that the form zo’n must then be realized as zo één. As before, the two forms differ in number: zo één is singular while zulke is plural, as is clear from the number agreement with the finite verb in (515).

The examples in (516) show that the D-linked interrogative demonstrative pronoun welk(e)which can also be used as an argument, although the use of the uninflected neuter singular form leads to a marked result. The pattern in (516) is thus similar to the one we found in (512) for non-interrogative pronouns.

There are reasons to think that the non-D-linked interrogative demonstrative wat voor een cannot be used as an independent argument. The examples in (517) are acceptable, but they must contain an occurrence of what appears to be quantitative er. This analysis is supported by the fact that er cannot be used in the corresponding constructions with the overt noun boekbook. This shows that we are not dealing with a demonstrative pronoun used as an independent argument, but with a construction comparable to Jan heeft er drie [e]Jan has three of them, where quantitative er licenses the empty nominal head of the direct object; cf. Section 20.3.

We conclude from this that wat voor een is a rather isolated case, which is special in that it can only be used as a modifier. This may not be a coincidence, since it has been argued that the string wat voor een is actually not a lexical element but syntactically derived; for arguments in favor of this, we refer the reader to Bennis et al. (1998).