Kinship nouns are typical examples of relational nouns: example (135a) is strange because there is no mention of a relational argument. Adding such an entity in the form of a genitive noun phrase or a PP-complement, as in (135b), makes the sentences acceptable; cf. Section 16.2.2, sub IV, for further discussion.

Body part nouns like hoofdhead and neusnose also typically take a related argument: the primed examples in (136) are odd because the inalienable possessor is not mentioned. Of course, example (136b') is possible, but not in the intended inalienable possession reading: the noun neusnose no longer functions as a relational noun, but as an ordinary noun, with the result that the noun phrase refers to someone else’s nose.

The impossibility of the indefinite article in the primed examples in (136) is due to the unique relation between the relational noun and its inalienable possessor, in the sense that the possessor has only one head/nose, so that the referent of the noun phrase can be inferred from the identity of the possessor. For the same reason, the indefinite article is excluded in the examples in (137).

In cases where the relation is not unique, as in the pair armarm and Jan in (138), the indefinite article can be used. However, despite the non-unique nature of the relation, the possessive pronoun or definite article can also be used when it is not known (or immaterial) whether we are dealing with Jan’s left or right arm.

In constructions such as (136a), the relational argument can also co-occur with the definite article; (139a) is perfectly acceptable under the same inalienable possession reading as (136a). However, this only holds when the inalienably possessed noun phrase is the complement of a locational PP; in most varieties of Dutch, (139b) is only possible with the non-inalienable possession reading.

The use of the number sign in (139b) indicates that in certain eastern and southern dialects an inalienable possession reading is possible; then the example is construed not as a transitive construction but as a semi-copular construction with the meaning “Janʼs nose is broken”; see Vxx and Axx for discussion; cf. A29.2.1, sub IB, for further discussion.

There are more relations that can be called “inherent” than the two discussed above. For example, all concrete objects have shape, size, weight, sides, and so on. Although it cannot be said that concrete objects “possess” these properties, nor that these properties are “part of” these objects, the relation between them is certainly “inherent”. Consequently, these nouns exhibit the same behavior as the relational and inalienably possessed nouns discussed above: the noun vormshape in (140a), for instance, will rarely be used in isolation from a related argument; as soon as a suitable related argument is added, as in (140b), the sentence becomes acceptable.

Note that the noun phrase in (140b) is introduced by the definite article; this is again possible thanks to the unique relation of the related argument to the relational noun, which allows us to infer the referent of the latter from the referent of the former.

Another relation that counts as “inherent” is the part-whole relation between the denotations of the nouns kaftcover and boekbook, or dakroof and gebouwbuilding: example (141) shows that these nouns behave like the inalienably possessed noun hoofdhead in (136).

Examples like (144a&b) are acceptable if the intended related argument can be inferred from the (linguistic or extra-linguistic) context; in other words, the related entity of a relational noun need not take the form of a complement. Example (135a), repeated here as (145a), is rendered acceptable by the addition of the adjunct PP met zijn zoontjewith his little son in (145b).

That the complement (van-PP or genitive noun phrase) can be left implicit if the related argument is recoverable from the context is also shown by (146). The complement of the nouns vorm or kaft expressing the related argument need not be present, since the related entity can be recovered from the preceding sentence.

The examples in (146) also show that the existence of a generally accepted (and expected) close association between two entities makes it possible for a definite article to precede a relational noun, even if the referent of the noun phrase has not previously been introduced into the discourse. The reference to the related entity of the relational noun is sufficient to ensure the identification of the noun phrase. Further examples illustrating the same point are given in (147).

In the case of kinship relations and body parts, however, the use of a possessive determiner is often preferred. This is illustrated in the (a)-examples in (148) for the kinship term vader; the use of the definite article is definitely the more marked option in the primeless examples, and is downright degraded when we know the girl (and her father), as in the primed example. With body parts, such as neusnose in the (b)-examples, the use of the definite article is odd in both contexts.

We refer the reader to Section 19.1 for a more detailed discussion of the role of definite and indefinite articles in determining reference.

Related to the case discussed in the previous subsection is the fact that the related argument can be omitted in clauses in which the relation between a relational noun and a related argument is established or explicitly denied.

Examples such as (149) are less common when the related argument is in a one-to-one relation with the relational noun: examples such as (150a) are impossible, while examples like (150b&b') carry an additional implication, namely that Jan’s father still lives/has died.

In generic contexts like (151), on the other hand, relational nouns like vaderfather and staarttail are perfectly acceptable without the related argument, because such examples express that the implied related argument stands in this unique relation to the relational noun.

Note that the hebben constructions in (149a) and (151a) differ from constructions with “possessive” hebben in that the indefinite noun phrase cannot be replaced by a definite one (e.g. with a definite article or demonstrative pronoun): compare the example in (152) with Jan heeft een/het/dat boekJan has a/the/that book.

Examples such as (150a) can be made acceptable by adding information to the noun phrase. Restrictive modification, for example, makes these non-prototypical uses meaningful by adding information that is not inherent in the basic meaning of the relational noun. This is illustrated by the examples in (153).

The use of an exclamative intonation contour has a similar effect even when the noun is not modified. This is because the exclamative contour has a similar modifying function as the restrictive modifiers in (153): (154a&b) express that the object has a remarkable feature; example (154c) seems to be fully lexicalized. Note, however, that these examples are often realized with the ethical dative meme and/or the particle toch: Jan heeft me toch een vader!

Slightly different cases are provided in (155), where the modifier seems to indicate that the referents of the noun phrases headed by vormshape and kaftcover are not identified by virtue of their relation with some uniquely related argument, but are taken from the sets of remarkable shapes and bright-yellow covers, respectively. In these examples, the relational aspect of the noun seems to have disappeared, and the nouns behave in the same way as non-relational nouns; the use of the indefinite article signals the fact that the speaker is introducing a “new” entity into the discourse domain, the reference of which cannot be inferred from the mention of a related argument in the preceding discourse.

Finally, example (156) shows that relational nouns can also occur without complement when used in predicative position.