Section 4.6 characterizes structures exhibiting monoclausal behavior by pointing to two prototypical properties. First, such structures exhibit verb clustering/clause splitting: the verbs are placed together in clause-final position and the dependents of the most deeply embedded verb (e.g., nominal arguments and modifiers) must precede the cluster as a whole. Second, such structures exhibit the infinitivus-pro-participio (IPP) effect in perfect-tense constructions with three or more verbs.

The examples in (2) illustrate the monoclausal properties of structures containing the aspectual verb komen: example (2a) shows that the verbs cluster in clause-final position, which results in splitting the lexical projection of the main verb repareren'to repair', given in italics; example (2b) illustrates the IPP-effect.

The two properties in Table 1 are neither necessary nor sufficient, however, for assigning non-main verb status to a certain verb. The examples in (3) show that exhibiting the IPP-effect is not a necessary condition—it does not occur in passive constructions despite the fact that passive auxiliaries are frequently seen as prototypical cases of non-main verbs. Note in passing that the percentage sign in (3b) indicates that most speakers from the Netherlands omit the participle geworden in the regular passive, whereas it is often realized by, especially, Flemish speakers.

Although it is generally true for most speakers from the Netherlands that verb clustering leads to splitting of the lexical projection of the most deeply embedded verb, it is not entirely true that this always results in a structure in which the dependents of this verb precede the non-main verbs: verbal particles, for example, can remain adjacent to it, and the same holds for certain monosyllabic complementives; see the examples in (4) and the discussion in Section A6.2.2, sub I for more details.

In fact, the restriction that verb clusters are impermeable by dependents of the embedded verb is even less strict for speakers of the Flemish variety of Standard Dutch, in which the verb cluster may easily include a wide variety of complementives, indefinite objects, etc. This means that we can only maintain that verb clustering is a necessary condition for assuming non-main verb status if we replace the stronger claim that verb clustering requires splitting of the lexical projection of the main verb by the weaker one that it makes splitting possible. The discussion in this chapter will show that there is a great deal of word order variation in verb clusters, especially those that contain a past or passive participle. Although the regional variation along the north/south dimension has been an intensively studied research topic since Pauwels (1953), we still found some gaps in the available information. We were fortunate in securing native-speakers judgments from the following Flemish speakers: Evie Coussé (East-Flanders), Benny de Decker (Province of Antwerp) and Reinhild Vandekerckhove (West-Flanders).

Subsection I has shown that, although the two tests in Table 1 for establishing whether or not we are dealing with a monoclausal structure normally provide reasonably reliable results for the speakers of Standard Dutch from the Netherlands, they are not without their problems in the light of the regional variation that we find. This subsection continues to show that the occurrence in a structure exhibiting monoclausal behavior is not sufficient for concluding that the term main verb should be reserved for the most deeply embedded verb (as most grammars do that adopt the "one-main-verb-only" criterion).
      In example (5a), the verb proberen 'to try' clearly functions as a main verb semantically; it is a two-place predicate that expresses the core meaning of the main clause. That we are dealing with a two-place predicate is clear from the fact that the infinitival clause can be pronominalized, as shown by (5b).

Example (6) expresses virtually the same meaning as (5a), so that there is no semantic reason for assuming that the verb proberen functions as a non-main verb in this construction. Nevertheless, the structure exhibits monoclausal behavior, that is, verb clustering and the IPP-effect; see Section 5.2.2.3 for detailed discussion.

The examples in (5) and (6) thus show that, although the monoclausal properties in Table 1 are typically found with certain prototypical non-main verbs, it is not the case that they are restricted to these verbs. It suggests that exhibiting these properties is not sufficient for concluding that we are dealing with non-main verbs, and, for this reason, Section 4.6 proposed to simply define main verbs as n-place predicates; any verb that takes one or more arguments is a main verb.
      On the assumption that subject/object pronouns always function as arguments, pronominalization of the projection of the infinitive can be used as a test for distinguishing between main and non-main verbs: infinitival clauses can only be pronominalized if selected by a main verb. The claim that the aspectual verb gaan in example (7a) is a non-main verb can therefore be supported by the fact illustrated by the corresponding primed example that the infinitival clause (Jan) de televisie repareren cannot be pronominalized; the number sign indicates that Dat gaat is only possible in the irrelevant reading "that can be done". That the verb proberen is a main verb is clear from the fact that pronominalization of the infinitival clause de televisie te repareren is possible.

      Another difference between main and non-main verbs is that while the former can increase the number of nominal arguments in the sentence, the latter cannot. This is the reason why the two primed examples in (8) are discussed in different sections. Example (8a') is discussed in Section 5.2.3.5, that is, as a case of a main verb with an infinitival argument clause, because the use of hebben goes hand in hand with the addition of the nominal argument Jan. Example (8b'), on the other hand, is discussed in this chapter on non-main verbs because the use of hebben does not affect the number of nominal arguments in the clause, at least not on the traditional assumption that the two arguments Jan en Piet are selected by the past participle gekust (but see Section 6.2.4 for some reasons not to adopt this view).

By defining the distinction between main and non-main verbs in terms of their ability or inability to select arguments, the dividing line between the two will be drawn at a different place than in most descriptive grammars: the set of non-main verbs will be considerably reduced. This definition does not affect the set of non-main verbs selecting a participle (although Section 6.2.4 will provide reasons for assuming that perfect and passive auxiliaries are less different in this respect from their cognates with other semantic/syntactic functions than is normally assumed).

The set of non-main verbs selecting a te-infinitive, on the other hand, is substantially reduced. Whereas descriptive grammars normally assume that it includes the semi-aspectual verbs in (10a) as well as the modal verbs in (10b), the latter are excluded by our definition because they allow pronominalization of the infinitival clause and thus clearly have an argument structure: for instance, pronominalization of the infinitival clause in Jan bleek zijn fiets verkocht te hebben'Jan turned out to have sold his bike' results in Dat bleek (lit.: that turned out). We indicate our exclusion of the modal verbs in (10b) from the set of non-main verbs by marking them with the number sign #.

The set of non-main verbs selecting a bare infinitive is likewise reduced. Whereas more traditional grammars assume that this set includes at least the modal, causative and aspectual verbs in (11), our definition only includes the last category. The modal verbs are again excluded because they allow pronominalization of the infinitival clause, as will be clear from comparing Jan moet dat boek lezen 'Jan must read that book' with Jan moet dat (lit.: Jan must that). And the causative verbs are excluded because they typically add an additional nominal argument, as will be clear from comparing Jan zingt een liedje'Jan is a song' with Jan liet Marie een liedje zingen'Jan made Marie sing a song'. The number sign # indicates that we diverge from the more traditional view by excluding the verbs in (11a&b) from the set of non-main verbs.

This chapter on non-main verbs considers the verb types mentioned in (9) to (11) insofar as they are not marked by a number sign. The verbs marked by a number sign are discussed in Section 5.2 on main verbs taking an infinitival argument.