This section discusses the second group of determiners: pronouns. Before entering into a detailed discussion of the pronominal types, we will make some general remarks about the classification of pronouns. In most Dutch traditional (school) grammars, pronouns are divided into the subclasses given in (324); cf. Haeseryn et al. (1997: §5).

The classification in (324) is unsatisfactory, because there are several elements that could in principle belong to more than one subclass. This becomes clear when we consider the set of interrogative pronouns: this class is assumed to contain the pronouns wiewho, wienswhose and welk(e)which, based on the semantic criterion that they are all interrogative words. On syntactic grounds, however, it seems equally justified to say that wiewho is a personal pronoun because it can be used in the same positions as the pronouns hijhe and hemhim, that wienswhose is a possessive pronoun because it can be used in the same positions as the possessive pronoun zijnhis, and that welkwhich is a demonstrative pronoun because it can be used in the same positions as the demonstrative pronouns die/datthat.

Of course, there is nothing wrong with making a classification based on semantic considerations, provided that it is done in a consistent way. However, traditional grammar fails in this respect by including adverbs like wanneerwhen and hoehow not in the class of interrogative elements, but simply in the class of adverbs. This again results in a classification in which certain elements could in principle belong to more than one subclass, while some classes fail to include all relevant elements. Another example is the subclass of “indefinite” pronouns, in which Haeseryn et al. (1997) include not only pronominal quantifiers such as iemand, but also quantificational elements such as sommigesome, which seem more related to numerals such as driethree than to pronouns.

It seems that these problems are caused by the fact that the traditional classification is based on a mixture of syntactic and semantic criteria; cf. Broekhuis (2002). In order to avoid these problems, or at least to make them visible, it seems better to apply the syntactic and semantic criteria in a more consistent way. A first attempt is presented in Table 6.

In Table 6, a first division is made on the basis of the syntactic distribution of these pronouns: are they used as independent arguments or as dependent modifiers of the noun phrase? On the basis of this formal criterion, the pronouns can be divided into the three main groups in (325). This division seems to be partly reflected in the semantics of the pronouns: while the personal and possessive pronouns have a limited amount of descriptive content, such as the ability to express that their referent is [±animate] (e.g. iemandsomeone versus ietssomething) or masculine/female (e.g. hijhe versus zijshe), the demonstrative pronouns seem to lack such descriptive content; the latter are mainly deictic elements that allow the addressee to determine the referent of the noun phrase they modify.

The three groups in (325) can be divided into smaller subclasses on the basis of semantic criteria (which can also be applied to adverbial phrases): are the pronouns referential, interrogative, or quantificational, or does their reference depend on an antecedent, as is the case with relative, reflexive, and reciprocal pronouns? Since demonstrative pronouns have virtually no descriptive content, it is not surprising that they cannot be divided into as many semantic subclasses as the personal and possessive pronouns.

The following sections discuss the three main classes of pronouns in Table 6: personal pronouns are discussed in Section 18.2.1, possessive pronouns in Section 18.2.2, and demonstrative pronouns in Section 18.2.3. Reflexive, reciprocal, and at least some occurrences of referential pronouns are dependent on some syntactically expressed antecedent, and we therefore conclude our discussion of pronouns with a discussion of some restrictions on these dependencies, as expressed in the so-called binding theory in Chapter 22.