Typically, the position of complementive adjectives is to the left of the main verb in clause-final position: placing the adjective after the main verb leads to serious unacceptability. This is illustrated in (56).

There is a divide between the northern and southern varieties of Dutch regarding the placement of the complementive when the main verb follows one or more auxiliary verbs (modal, aspectual, etc.) in clause-final position. In the northern varieties, the complementive adjective usually precedes all verbs, as in the primeless examples in (57); placement of the adjective between the auxiliary and the main verb, as in the primed examples, is generally rejected, although it should be noted that these orders are occasionally produced and some speakers tend to accept them as marked but acceptable. However, all speakers of the northern varieties of Dutch do agree that making the adjective phrases in (57) more complex by adding a complement or a degree adverb, such as ergvery, renders the primed examples completely unacceptable; the same applies to adjectives in the comparative or superlative forms.

The use of the percentage sign indicates that the requirement that the clause-final verbs be strictly adjacent to each other does not apply to the southern varieties of Dutch. This is especially clear in the varieties of Dutch spoken in Flanders, where even in formal registers several types of elements can permeate (i.e. break up) the clause-final verbs. In these varieties, the orders in the primed examples in (57) are perfectly acceptable, even if the complementive adjectives are modified. However, whenever the main verb precedes the auxiliary verb(s), a complementive cannot interrupt the verbal sequence in any variety of Dutch (or any other OV-language); cf. the contrast between the primeless and primed examples in (58). We ignore here the orders marked with % and refer the reader to Section V7.4 for further discussion of the permeation of verb clusters.

Returning to the aux-main verb orders, we should note that for some more or less fixed adjective-verb combinations, all speakers of Dutch accept the occurrence of the complementive adjective in between the auxiliary and the main verb. These combinations are sometimes considered compounds, which is reflected in the orthographic convention of writing the combination as a single word in the primeless examples of (59). That we are dealing with compounds may find support in the fact that these adjective-verb combinations can be rendered by simple English verbs: to raise, to clean and to drain. It is also suggested by the fact that modification of the adjectives by degree adverbs such as ergvery leads to an awkward result (for speakers of the northern varieties). However, the compound analysis leaves unexplained that modification is also excluded when the adjective and the verb are not adjacent, as in the primed examples (as well as in verb-second contexts).

If the complementive adjective precedes the verb(s) in clause-final position, it is usually also adjacent to it/them: the intervention of elements such as adverbial phrases leads to unacceptability in all varieties of Dutch if the sentence is pronounced with an unmarked intonation pattern (i.e. with sentence accent on the complementive). This is illustrated in (60).

The general rule for the placement of complementive adjectives can be schematized as in (61): √ indicates the position of the complementive in the “unmarked” construction, % indicates the position where the adjective cannot occur in the northern varieties, and * marks the positions where the adjective cannot occur at all; the dots indicate random strings of elements. Subsection II will show, however, that there are several exceptions to this general schema.

To be more precise, we should say that the adjectivw phrase must be adjacent to the clause-final verb(s), since otherwise an example such as (62), in which the intervening PP is an argument of the adjective (cf. Section 24.1), would pose a problem for this claim. For convenience, however, we will stick to our loose formulation in the following discussion.

In the following, we will focus on the northern varieties of Dutch and consequently ignore the position in (61) indicated by %. The northern varieties seem to obey the following left-adjacency condition: in neutral clauses, complementives are left-adjacent to the clause-final verb cluster. This condition suggests that there is a close relationship between the verb cluster and the adjective. That this is indeed the case is also clear from the fact, illustrated in (63b-c), that the adjective must be pied-piped by VP-topicalization.

The examples in (64) show that it is not impossible to separate the verb and the adjective. This frequently happens in verb-second contexts if the main verb is finite, as in (64a). Separation can also result from leftward movement of the complementive adjective. This is illustrated in (64b) for topicalization. The different types of leftward movement that can produce such separation are discussed in Subsection II.

The left-adjacency condition on complementive adjectives can be overridden by assigning contrastive focus accent to the adjective; the primed examples in (60) above become much better in this case. The order in these examples can be improved even further by placing a focus particle such as zothat before the adjective. This is illustrated in the primed examples of (65). The primeless examples show that the complementive adjective can also occupy the unmarked position.

The left-adjacency condition can also be overridden by placing the adjective in clause-initial position by wh-movement or topicalization. We illustrate this in (66) by topicalization. The corresponding wh-constructions can be constructed by the reader by placing the interrogative degree modifier hoehow before the adjective.

These examples show that we have to revise the schema in (61) as in (67): again, the dots represent a random string of elements; √ indicates the position of the complementive in the “unmarked” construction; % indicates the position in which the adjective can occur in the eastern but not in the northern varieties; and * marks the positions where the adjective cannot occur in Dutch neutral constructions. In addition, WH/TOP indicates the position of the adjective in topicalization or wh-constructions such as (66), and FOC indicates the position of the adjective in focus constructions such as (65); the asterisk below FOC indicates that this position can only be filled by contrastively focused complementive adjectives.

Leftward movement of the adjective breaks the tight relation between the complementive and the verb. We therefore expect that in such cases VP-topicalization can strand the complementive. Of course, this expectation cannot be checked when the complementive itself has undergone wh-movement or topicalization, since the prospective landing site of the VP would already be filled, but it is can be checked when the adjective has undergone focus movement, since this leaves the clause-initial position free for VP-topicalization, as in (68)

Although the markedness of (68) suggests that VP-topicalization is not possible for some reason or other, we will see in the next subsection that leftward movement of the adjective can indeed make it possible for the verb to be topicalized in isolation.

Like complementive adjectives, stranded prepositions are usually left-adjacent to the verb cluster in clause-final position, which implies that the two compete for the same position. However, they can also co-occur in a single clause. Consider example (69a), which contains the split instrument PP waar ... meewith what. At first sight, the order of the stranded preposition mee and the complementive adjective droogdry seems to be free: the stranded preposition mee can either precede or follow the complementive. Essentially the same thing holds for (69b), where the split van-PP functions as a restrictive adverbial phrase. For convenience, we have italicized the stranded preposition and its moved complement, the R-word waar.

However, the situation is more complex than the examples in (69) suggest; in fact, the general rule seems to be that the stranded preposition follows the complementive adjective. The following subsections discuss various facts that support this claim.