Section 19.2.2.1 has shown that possessive pronouns are in complementary distribution with articles, and we have used this fact to motivate the claim that they function as determiners of the noun phrase. Another reason for this claim is that possessive pronouns also share certain semantic properties with articles. This is most obvious in the case of referential possessive pronouns, which contribute more or less the same meaning as the definite articles. Consider Figure 1, where A represents the denotation set of the lexical part (i.e. the NP-part) of the subject and B represents the set of entities denoted by the verb phrase, where A and B are both contextually determined, i.e. dependent on the domain of discourse (domain D). The intersection A ∩ B denotes the set of entities for which the proposition expressed by the clause is claimed to be true.

Section 19.1.1.1 has argued that the core meaning of the definite article is that all entities in domain D which satisfy the description of the NP-part of the subject are included in the intersection A ∩ B, i.e. that the remainder of set A is empty; cf. (463a'). The referential possessive pronoun zijnhis in (463b) expresses a similar meaning, but additionally introduces a partitioning of set A: the assertion is not about all entities that satisfy the description of the NP, but about a subset Asub that stands in a certain relation to the referent of the possessive pronoun. Of course, the fact that the possessive pronouns imply a partitioning of set A does not necessarily imply that set A is a non-singleton set. If domain D contains only one book, the speaker can still use the noun phrase mijn boekmy book; in this case the evoked alternative referent set is empty.

Note that the relationship in question need not be one of possession, but can be of various kinds, and is largely determined by the non-linguistic context; cf. Janssen (1976:§3.1) for a relevant discussion. The referent of the possessive pronoun in (463b) may be the author or publisher of the books, but also someone who edited them or made a guess about which books would sell well.

It is not only the core meaning of the definite articles that is associated with referential possessive pronouns; other properties of definite noun phrases can also be found in noun phrases containing a referential possessive pronoun. For example, both types of noun phrases usually refer to entities in domain D that are assumed to be uniquely identifiable by the speaker; a question such as (464a) presupposes that the listener is able to identify the referent of the noun phrase mijn sleutels. And just as in the case of the definite article, noun phrases with a referential possessive pronoun can introduce new entities into domain D that are somehow anchored to a known entity in domain D. An example such as (464b) does not presuppose that the listener knows who Jan’s wife is, but that the mention of Jan is sufficient to anchor the referent of the noun phrase zijn vrouwhis wife to someone related to him.

Noun phrases with a referential possessive pronoun, like definite noun phrases, also exhibit exceptions to the general requirement that the noun phrase be uniquely referring. For example, if the noun phrase refers to a body part, like a leg or a hand, the noun phrase may be singular, leaving some vagueness as to which of the (two) hands or legs is meant; cf. example (465a). Something similar occurs with kinship nouns; an example such as (465b) does not imply that the speaker has only one nephew; apparently it is not the referent that matters here, but the relationship between the speaker and the person referred to. A similar case with a non-kinship noun is given in (465c), which expresses that the train the speaker took was delayed.

Because of the overlap in meaning between referential possessive pronouns and definite articles, the noun phrases introduced by a possessive pronoun in the primeless examples of (466) are virtually synonymous with the noun phrases in the primed examples with a definite article and a postnominal possessive van-PP. This suggests that, apart from their reference, the meaning of the referential possessive pronouns in the primeless examples consists of two parts corresponding to the meaning of the definite article and the modifying van-phrase in the primed examples: the first part involves definiteness, and the second part involves the partitioning of the set denoted by the head noun into two subsets, namely a subset that is in the relevant semantic relation with the referent of the possessive pronoun and a subset that is not.

Since referential possessive pronouns are inherently definite, possessed indefinite noun phrases usually involve the presence of an indefinite article and a postnominal possessive van-PP, as in (467). As always, the indefinite article expresses that the intersection A ∩ B has cardinality 1, without any implication for the remainder of set A, i.e. A - (A ∩ B) may or may not be empty.

For completeness’ sake, note that the complement of the preposition van is a personal object pronoun and not, as in English, an (inflected) possessive pronoun as in a book of mine/yours/his/hers.

Example (468a) shows that indefiniteness can also be inherited from the existentially quantified possessive pronoun iemandssomeoneʼs. That the whole noun phrase is indefinite is clear from the fact that the noun phrase iemands auto can occur in the expletive er-construction. The possessive pronoun in this example again introduces a partitioning of set A, but the speaker leaves open which subset of A is meant. The universal possessive pronoun also introduces a partitioning of set A, but now it is claimed that all subsets of A are subsets of B. As a result (468b) expresses more or less the same thing as the simpler sentence De/Alle auto’s staan verkeerd geparkeerdThe/All cars are wrongly parked, which perhaps explains why (468b) feels a little tortuous.

The reciprocal form elkaarseach otherʼs and the interrogative and relative form wienswhose also introduce a partitioning of the denotation set of the nominal head of their noun phrases. In (469a), the cardinality of the antecedent of the possessive pronoun is equal to the cardinality of the relevant subset of the denotation set of opstel, and the members of the antecedent and the relevant subset are reciprocally related: set A consists of three essays, each written by a different pupil, and each of the pupils admires the essays written by the other two pupils. Question (469b) can be used when the set of books is divided into subsets defined by e.g. ownership, and the speaker asks about a particular subset of books. The relative construction in (469c) allows the addressee to pick out the intended referent of the complete noun phrase on the basis of a contextually determined partitioning of the books.

For completeness, the examples in (470) show that noun phrases headed by interrogative wiens are weak indefinites, while noun phrases headed by demonstrative wiens are strong. This is clear from the fact that only (470a) can occur in expletive er-constructions; er in (470b) is only possible if it is given a locative interpretation.

The possessive pronouns owe their name to the fact that in many cases they refer to the owner of the referent of the complete noun phrase; the noun phrase mijn boekmy book typically refers to a book owned by the speaker. However, the term possessive pronoun (or, more generally, possessive noun phrase) is a misnomer, because the relation between the pronoun’s referent and the referent of the complete noun phrase need not be restricted to possession; the noun phrase mijn boek can also involve a relation of authorship. In the following subsections we will briefly discuss two systematic kinds of relation that the referent of a possessive pronoun and the referent of the complete noun phrase of which it is a part can enter into. The discussion in the following subsections is not intended to be exhaustive, as the creative powers of language users far exceed our descriptive potential.