Monadic adjectives are adjectives which take a single argument, which we assumed in Section 15.3.2, sub II, to be assigned the thematic role of referent. The adjective is predicated of this argument (from now on we will ignore the attributive use of adjectives); this is illustrated for the monadic adjectives verlegenshy and hooghigh in (431a&b).

The examples in (432) show that Ref-arguments can also be expressed in noun phrases headed by a deadjectival noun. It can be expressed either as a postnominal van-PP or as a prenominal possessive pronoun or genitive noun phrase. As always, the latter option is restricted to proper nouns and a limited set of [+human] nouns.

The marginal status of the primeless examples in (433) shows that, as with deverbal nouns, the referent argument is usually obligatory, although there are two exceptions. First, it is possible to omit the argument in generic statements such as (433a'). Second, the argument need not be expressed if it is recoverable from the (extra-)linguistic context; example (433b) would be perfectly acceptable in a conversation about a particular tower.

Dyadic adjectives are adjectives that take two arguments, one of which is assigned the thematic role of referent. The second argument is typically expressed by a postadjectival PP, but can sometimes be expressed by a preadjectival genitive or dative noun phrase. We restrict ourselves here to dyadic adjectives like gehechtattached and ingenomenpleased in (434) which take a PP theme argument: Section 15.3.2, sub IIIC, has shown that adjectives selecting a nominal complement cannot be input for nominalization.

The examples in (435) show that the derived noun inherits both arguments; as in the case of the monadic nouns, the referent argument can be expressed either by a postverbal van-PP, or as a prenominal possessive pronoun/genitive noun phrase. The second argument of the adjective is obligatory in the nominal constructions in (435), and must follow the Ref-argument if the latter occurs as a postnominal van-PP. Importantly, the second argument has the same form as in the corresponding adjectival construction.

The cases in (436) show that for some adjectives the second argument is optional. It stands to reason that the same holds for the nominalizations of these cases: it simply shows that the optionality or obligatoriness of the complement is one of the properties inherited by the deadjectival noun. Note that the Ref-argument cannot usually be omitted: this is only possible in generic contexts or when the Ref-argument is recoverable from the context, but this will not be illustrated here.

It seems that there is a difference in the productivity of the nominalization process between the two cases. It is easy to find examples of adjectives with an obligatory second argument that cannot be nominalized, as shown in (437).

Adjectives with an optional second argument are harder to find; the examples in (438) illustrate the high degree of productivity in this case.

The primeless examples in (439) show that the complements of attributively used adjectives must appear in preadjectival position; cf. Section A28.3 for discussion. However, there is no comparable position for deadjectival nouns; as the primed examples in (439) show, these PPs can only be placed in postnominal position.

Among the adjectives taking an optional prepositional complement, there are some cases where the presence or absence of the complement leads to syntactic differences regarding the pluralization of the logical subject of the adjective; cf. also Chapter A25. An example is given in (440): if the subject is singular, as in (440a), the PP-complement must be present; if it is plural, as in (440b), the PP is optional.

The adjectives in both constructions can be the input for nominalizations; interestingly, however, the nominal counterpart of the construction in (440a) selects a different preposition (metwith instead of aanto), while in the case of (440b) the plural subject now appears as PP-complement with the preposition tussenbetween. The relevant examples are given in (441).

Occasionally, adjectives can occur with three arguments, i.e. with a Ref-argument and two PP-complements. An example is (442a). The (b)-examples show that nouns derived from these adjectives can inherit all three arguments, although the result may be slightly marked. Again, the Ref-argument can occur prenominally as a genitive noun phrase/possessive pronoun or postnominally as a van-PP. The second and the third argument must occur postnominally in the same form and order as in the original adjectival construction. Furthermore, they must follow the Ref-argument if the latter occurs as a postnominal van-PP. This implies that any reordering of the PPs in (442b&b') will lead to a degraded result.

Since the complements of triadic adjectives are not always obligatorily expressed, it is to be expected that the same is true of the complements of the derived noun; one might say that the optionality of the complement is one of the features inherited from the adjective. Thus, the constructions in (443) are perfectly acceptable, with the possible exception of (443a'), which for unclear reasons seems to be slightly degraded. Recall that the Ref-argument cannot usually be omitted.

The inherited Ref-argument of deadjectival nouns must be realized either as a prenominal genitive noun phrase/possessive pronoun, or as a postnominal van-PP. The PP-complements of the input adjective are also inherited: they appear postnominally in the same form and order as the complements of the input adjective. Whether these PP-complements can be left implicit also depends on the properties of the input adjective.

Subsection I has already shown that the Ref-argument must normally be expressed: it can only be left implicit in generic contexts or if it is recoverable from the (extra-)linguistic context. We illustrate the obligatoriness of the Ref-argument again by means of a noun derived from the monadic adjective vruchtbaar.

The obligatoriness of the PP-complements to nouns derived from dyadic adjectives depends on whether they are obligatory in the corresponding adjectival construction. In other words, the optionality or obligatoriness of the complement of the adjective is inherited by the deadjectival noun. This is illustrated again in (446).

The examples in (447) illustrate again that the Ref-argument can only be left implicit if it is recoverable: the use of the possessive pronoun zijn in (447a) recalls the idea that the Ref-argument of the noun is the owner of the dog, and consequently leaving it implicit is allowed; in (447b) such a clue is missing, and leaving out the Ref-argument therefore leads to a degraded result.

Finally, as shown in example (448), triadic deadjectival nouns do not require the presence of all three arguments, a property inherited from the input adjective. Again, the Ref-argument is usually required.

Example (449) shows that the Ref-argument of a deadjectival noun cannot occur in post-copular position. This is obvious because van-PPs in post-copular position are interpreted as possessive elements, and properties, the denotation of deadjectival nouns, cannot be possessed.

For completeness, the examples in (450) show that it is also impossible to put the second or third argument of dyadic and triadic constructions in post-copular position.

The examples in (451a-c) show again that Ref-arguments behave like complements: they allow R-pronominalization. Note that example (451d) shows that the result is much worse for adjectives taking a [+human] complement: de boosheid van JanJanʼs anger; this is due to the fact that R-pronominalization is always marked when the PP contains a [+human] noun phrase.

Example (452) shows that R-pronominalization is also possible with the second argument of dyadic adjectives.

The same holds for the second and the third argument of triadic nouns such as boosheidanger, although there are additional restrictions. Example (453a) shows that R-pronominalization of the over-PP leads to a perfect result if the op-PP is left implicit. Similarly, example (453b) shows that R-pronominalization of the op-PP is significantly better when the over-PP is not expressed. The fact that the result is still marked without the presence of a second complement is due to the fact that this third argument is typically interpreted as [+human], and as such does not easily allow R-pronominalization; however, if the argument is interpreted as referring to some institution (such as the government), the example becomes more or less acceptable. Finally, example (453c) shows that pronominalization of both complements at the same time is completely impossible.

As expected, the results in (453) reflect what we find in the corresponding adjectival construction in (454). We used the strong form daar + P instead of the weak form er + P because it makes it easier to use the unsplit pattern in the adjectival construction; the judgments do not change when we use the weak form.

The PP-extraction test yields results that are far from unequivocal, although they can be characterized as rather poor overall. In the following, we consider the possibility of topicalization, relativization and questioning, PP-over-V and scrambling.

Table 15 summarizes the results of the four tests for inherited arguments of deadjectival nouns. The third and fifth columns indicate whether the results provide evidence for or against the assumption that we are dealing with complements.

The first three tests provide unequivocal evidence for the complement status of both van-PPs and PP-themes headed by other prepositions. There is a marked difference in the behavior of van-PPs and PPs headed by other prepositions with respect to the possibility of extraction: the conclusion that inherited van-themes function as complements may be supported by the extraction facts, but for PP-themes headed by other prepositions the results are negative. However, since we have seen that the PP-extraction tests are problematic in various ways and may not be suitable for establishing the complement status of PPs, it seems that we can still safely conclude that both types of PP-theme function as arguments of the derived noun.