Section 28.1 has shown that the complementive adjective appears prototypically in the copular construction, exemplified in (13); all examples express that the set referred to by the noun phrase de jongens is a subset of the set referred to by the adjective. The actual choice of the copula can add specific meaning aspects. The copula zijn is neutral and expresses a purely predicative “N is A” relation. The copulas zijnto be, wordento become and blijvento remain are aspectual in nature and indicate that a state remains stable. The copulas lijkento seem, blijkento turn out and schijnento seem are modal in nature: lijkento seem indicates that the assertion is based on the subjective perception of the speaker, blijken indicates that there is sufficient evidence to assume that the assertion is true, while schijnen indicates that that the available evidence may be insufficient.

The inchoative verb raken is sometimes used in the same function as worden, and they are often in complementary distribution. The factors determining the choice between the two options are not entirely clear. The difference between the two (a)-examples may be related to the form of the adjective: the fact that gefrustreerdfrustrated can also be used as a past/passive participle may favor raken over worden, because worden can also be used as a passive auxiliary and would introduce an unwanted ambiguity. In many cases, however, the motivation for the choice is not immediately obvious, as in the (b)-examples.

The examples in (15) show that the adjectives in the constructions with a modal verb alternate with infinitival clauses headed by an aspectual copular verb; this combines the modal and aspectual meanings found in (13).

Unlike the complementive constructions discussed in Subsections II and III, the copular construction can be used with all types of set-denoting adjectives; cf. Section 23.3.2.2 for examples.

The eastern varieties of Dutch have an alternative way of conveying the assertions expressed by the standard Dutch regular copular constructions in (13a). Typical examples of this semi-copular construction, which is also called the band-lek construction, are given in (16); cf. Van Bree (1981) and Cornips (1994: §5.2).

The dialectal semi-copular constructions, however, differ from regular copular constructions in that they typically express a possessive relationship between the nominative argument Jan/hij and the logical subject of the adjective: translated into standard Dutch, the examples in (16) would yield the sentences in (17).

Although the semi-copular constructions in (16) are unacceptable in standard Dutch (probably because the use of external subject/object possessors is more restricted in this variety; cf. Section V3.3.1.4, sub I), there are two acceptable constructions that resemble them in various ways. First, consider the examples in (18), which may be rejected as they stand by some speakers of standard Dutch, but become perfectly acceptable if certain adverbs are added; cf. the examples in (20) below.

The dialectal and standard Dutch constructions in (16) and (18) differ in at least the following two respects; cf. Broekhuis & Cornips (1994). First, unlike the examples in (16), the examples in (18) do not express a possessive relationship. This can be made visible by adding a possessive pronoun to the logical subject of the adjective: example (19) shows that this yields an unacceptable result for the dialectal construction in (16a) but a perfectly acceptable result for the standard Dutch construction in (18a). Note that example (19a) would be acceptable in standard Dutch as well with a similar (but less probable) intentionality reading as found in (19b); both the puncturing of the speaker’s tire and the cleaning of the brush are then presented as deliberate actions.

Second, the standard Dutch examples in (18) imply that the subject of the clause can affect the state of the object; this can be seen from the fact that adverbial phrases like nog nietnot yet or eindelijkfinally can be added to these examples, as in (20). These adverbial phrases express that the subject of the clause is actively involved in the process of cleaning the brush/closing the window: Jan is trying to clean the brush or open the window but has not yet/finally managed to achieve the desired result.

The involvement of the referent of the subject in obtaining the result denoted by the adjective is also clear from the fact that the verb hebben can be replaced by the (aspectual) verbs krijgento get and houdento keep, as in (21). In such cases, it is possible to use the subject-oriented adverbial phrase met moeitewith difficulty, which emphasizes the fact that Jan is actively involved in the process of cleaning the brush or opening the window by expressing that Jan experiences some difficulty in achieving the desired result.

A second standard Dutch construction that also involves hebben + adjective is given in (22a). Since het cannot be replaced by the demonstrative pronoun datthat, we have to conclude that this construction involves the non-referring element het, which we also find in constructions such as (22b), which will be discussed more in more detail in Section 28.5.2.

The verb hebben in (22a) can be replaced by krijgento get, as is shown in (23a), but the fact that the adverbial PP met moeitewith difficulty cannot be added suggests that the subject of the clause is not a controller, but rather a kind of experiencer. If we use the verb houden, as in (23b), the translation with to keep is no longer appropriate; instead, the correct translation requires the copular verb to remain; this again suggests that the subject in this construction functions as an experiencer.

In examples like those in (24), it seems possible to add the adverbial PP met moeite or to use the verb houden with the meaning “to keep”. This is only apparent, however, because example (24a) turns out to be ambiguous: in one reading, the pronoun het is a non-referring expression, just as in (22); in the second reading it is a deictic pronoun referring to some entity in the discourse domain (e.g. het gerechtthe dish), as is clear from the fact that het can be replaced by the demonstrative datthat. The examples in (24b&c) are only licensed in the second reading, which is actually the same type of construction as in (18).

Note that the pronoun het in example (24b) can also be interpreted as an anticipatory pronoun introducing an (implicit) locational phrase: Ik krijg het met moeite warm (in de kamer)I can hardly heat the room; cf. Section 28.5.2, sub III, for a discussion of this reading. Note further that, although the meanings of (22a) and (24a) are very close to the meaning of the copular constructions Ik ben benauwdI am out of breath and Ik ben warm ‘I am warm, it would be incorrect to conclude that the adjective in (22a) and (24a) is predicated of the subject ik, since a paraphrase by a copular construction is often ruled out in structurally similar examples. This is illustrated in (25); example (25a) expresses a completely different meaning from example (25b). The English paraphrases attempt to express this difference.

Consider again the primeless examples in (10), repeated here as (28a&b). Since impersonal (weather) verbs like regenento rain and vriezento freeze do not take a referential noun phrase as their subject, we conclude that the nominative noun phrase de jongen in the primed examples is in fact the logical subject of the resultative adjectives. This, in turn, implies that this noun phrase is in fact a theme-subject (i.e. an internal argument) and that the weather verbs in these examples are unaccusative in nature.

That the weather verbs in resultative constructions are unaccusative is supported by the fact that they exhibit the properties typical of unaccusative verbs. First, the singly-primed examples in (29) show that weather verbs in resultative constructions, in contrast to those in non-resultative constructions, take the auxiliary zijnto be in the perfect tense. Second, the doubly-primed examples show that the past/passive participle of the weather verb can be used attributively when it modifies a head noun corresponding to the nominative argument of the verbal resultative construction, given that the resultative adjective is present as well. Finally, the triply-primed examples show that the impersonal passive of the resultative construction is impossible.

As with the weather verbs in (28), an additional nominal argument is introduced by complementive adjectives used with an intransitive verb. Consider the primeless examples in (30). Example (30a) shows that a verb like lopen cannot take a noun phrase like zijn schoenen as a direct object. However, if the adjective kapotworn-out is added, as in (30b), this noun phrase must be present. Again, we have to conclude that the noun phrase zijn schoenen is introduced into the structure as the logical subject of the adjective (although it acts as the direct object of the verb in the sense that it is assigned accusative case by the verb). The primed examples give similar cases with the intransitive verb huilento cry.

Unaccusative resultative constructions occur not only with the impersonal verbs discussed in Subsection A, but also with regular unaccusative verbs such as vallento fall in (31a). Some unaccusative verbs, such as slibbento silt in (31b), must even occur in a resultative construction. We have ignored here that slibben also occurs as a transitive verb in jargon with a different meaning (in the jargon of dredgers).

The examples in (32) show that the verbs in (31) exhibit the typical properties of unaccusative verbs: the primeless examples show that they take the perfect auxiliary zijnto be, the singly-primed examples that the past/passive participle of the verb can be used attributively if the head noun corresponds to the nominative argument of the verbal construction (provided that the resultative adjective is also present), and the doubly-primed examples that the impersonal passive is excluded.

The unaccusative verbs in (31) differ from the intransitive verbs in (30) in that the logical subject of the adjective also satisfies the selection restrictions of the verb in the sense; the adjective cannot license an additional noun phrase with no semantic relation to the verb. The noun phrase de vaas in (33a) acts as the subject of the unaccusative verb brekento break, but it can also occur as the subject of the resultative adjective kapotbroken in (33b); the addition of an additional noun phrase such as Jan in (33c) is excluded.

The unacceptability of (33c) contrasts sharply with the acceptability of the (b)-examples in (30). The contrast can be explained in terms of case assignment: unaccusative verbs cannot assign accusative case, and the ungrammaticality of (33c) therefore results from the fact that Jan/him is not formally licensed (i.e. remains caseless); intransitive verbs can assign accusative case, so the noun phrase zijn schoenen in (30b) is formally licensed. This leaves us with the fact that intransitive verbs do not take an accusative object in the absence of a resultative adjective. This is because objects are not semantically licensed by intransitive verbs. This account is based on Chomsky’s (1981) case filter, which requires every phonetically realized noun phrase to be assigned case, and of course presupposes that case is assigned even if it has no morphological reflex.

Although it is not possible for complementive adjectives to introduce an additional noun phrase, it is possible for them to introduce a noun phrase that replaces the original subject of the unaccusative verb. Two examples are given in (34a&b); these examples show that the logical subject does not satisfy the selection restrictions of the verb springen (leaving aside the complicating fact that the verb springen is also used as a shorthand for kapot springen in (34b)). Such examples have an idiomatic flavor and are quite rare with complementive adjectives, although they are common with prepositional complementives; cf. the examples in (34c-d).

Transitive resultative constructions differ from the unaccusative ones in that logical subjects of resultative adjectives are often thematically unrelated to the verb. This is illustrated in (35); cf. Hoekstra et al. (1987). Example (35a) shows that the verb vervento paint can take the noun phrase de deurthe door as its direct object, but not the noun phrase de kwastthe brush (in the intended reading, where de kwast is not the object painted but e.g. the instrument used). Nevertheless, both noun phrases are acceptable in the resultative constructions in (35b) and (35c). However, example (35d) shows that the two noun phrases cannot be simultaneously present.

The unacceptability of (35d) may seem unexpected, since the noun phrase de deurthe door is semantically licensed by the verb vervento paint and the noun phrase de kwastthe brush is semantically licensed by the adjective kapotbroken. Again, this has to do with case assignment: if a transitive verb can assign accusative case only once, one of the two noun phrases remains caseless, which violates Chomsky’s case filter; cf. the discussion of example (33c).

Although the verb verven is used as a transitive verb in (35a), we cannot straightforwardly conclude that it is also used transitively in (35c), because this verb can also be used as a pseudo-intransitive verb: cf. Jan verftJan is painting. So it is possible that in (35c), too, we are dealing with an intransitive verb. This suggestion is supported by the paraphrases in (36): example (35b) is preferably paraphrased with the transitive verb verven, as in (36a), while (35c) must be paraphrased with the intransitive verb verven in (36b').

Example (37) provides more cases of transitive verbs with a pseudo-intransitive counterpart, in which a resultative adjective can introduce a noun phrase that is not thematically related to the verb; (37a') does not express that zijn oudershis parents are the objects being eaten, but that Jan’s parents are becoming poor because Jan eats so much. Similarly, in (37b'), de longenthe lungs are not being smoked; they are merely growing black as a result of Jan’s smoking.

The discussion suggests that, despite appearances, complementive adjectives cannot introduce an additional argument into the structure in the case of “truly” transitive (i.e. not pseudo-intransitive) verbs; this is only possible with (pseudo-)intransitive and weather verbs. In other words, when an internal argument is present, as in the case of the regular unaccusative verbs in Subsection C and the “truly” transitive verbs, this object must be construed as the logical subject of the complementive adjective in the resultative construction. This seems to be supported by the fact that, in “truly” transitive resultative constructions, the accusative object must generally be realized overtly; cf. (38).

There are, however, some exceptional constructions in which the accusative object is omitted: example (39a) is a fixed expression and example (39b) is an advertising slogan. Such examples usually have a generic interpretation; cf. Vanden Wyngaerd (1994:ch.4) and the references cited there for further discussion.

Finally, note that the accusative object of the “truly” transitive resultative construction can take the form of a reflexive. Unlike regular accusative objects, the reflexive does not have to take the complex form zichzelfhimself but can also appear in its simplex form zich; cf. Section N22.2. This is shown by the contrast between (40a) and (40b).

In (40b), the reflexive can be replaced by a regular referential noun phrase, such as Marie. Occasionally, however, this is impossible in the resultative construction; the reflexive must then appear as the simplex form zich. This is demonstrated in (41).

Some verbs are typically used in resultative constructions: in other syntactic frames, they either do not occur at all or they take on a special meaning. An example of the first case is the verb verklarento declare: example (42b) shows that omitting the complementive adjective yields an unacceptable result, regardless of whether the noun phrase het beroep is present or not.

An example of the second case is the verb of creation maken. In the resultative construction it is deprived of its normal meaning “to create/repair”, as shown by the contrast between (43a) and (43a'). Note that in examples such as (43b), where the create/repair reading is less likely, the complementive adjective must be present.

The examples with the verb houdento keep in (44) are in a sense the opposite of the resultative constructions discussed in this section; instead of expressing that the logical subject becomes part of the set denoted by A, it is expressed that the subject remains part of the set denoted by A. Examples (44a&b) are more or less lexicalized, and (44c&d) are clearly idiomatic.

Example (45) is a more or less isolated case, in which a desirable result is expressed. This construction is severely restricted in the sense that the adjective dooddead cannot easily be replaced: *?Jan wenste zijn vader ziek/gelukkigJan wished his father ill/happy. Note that in non-resultative constructions with wensen, such as Ik wens je een prettige verjaardagI wish you a happy birthday, the particle toe can be added optionally. This is not possible in (45), however.