The auxiliaries hebben and zijn are temporal in the sense that the perfect-tense constructions of which they are part situate the state of affairs prior to a certain point in time. Example (66a), for instance, situates Marie’s arrival prior to the speech time (which is the default value), as is made quite clear by the fact that it can be modified by the temporal adverbial gisterenyesterday, but not by the temporal adverbial morgentomorrow. Moreover, under certain conditions, perfect-tense constructions can also have aspectual implications by expressing that the state of affairs denoted by the main verb has been completed in the sense that some logically implied endpoint has been reached: e.g. example (66b) can only be used if Jan has told the whole story. For a detailed discussion of the semantics of the perfect tense, see Section 1.5.1.

Participles are also used in combination with the auxiliaries wordento be and zijnto have been in regular passive constructions like (67a&b) and the auxiliary krijgento get in so-called krijgen-passive constructions such as (67c).

Note that the auxiliary verb zijn in (67b) is sometimes analyzed not as a passive auxiliary but as a perfect auxiliary, since the passive participle geworden can at least marginally be added to such examples. If correct, this means that worden and krijgen would exhaust the set of passive auxiliaries, but see Section 6.2.2 for a possible problem with this conclusion.

The fact that the auxiliaries discussed in this section are only instrumental in creating perfect or passive constructions explains immediately why they cannot be used as heads of clauses (although zijnto be and wordento become do occur as copulas, and hebbento have and krijgento get can also be used as main verbs of possession).

The examples in (68) show that modal and aspectual verbs like willen and gaan differ from temporal and passive auxiliaries in that they do not combine with participles, but require the main verb to take the form of a bare infinitive.

The primeless examples in (69) show that modal and aspectual verbs also differ from main verbs in that they exhibit the infinitivus-pro-participio (IPP) effect; they do not take the form of a participle in perfect-tense constructions, but of an infinitive. The primed examples have been added to show that willen and gaan do appear as participles when used as main verbs.

If modal and aspectual verbs add specific modal/aspectual information to the event expressed by the main verb, we expect that these verbs cannot be used without a main verb. This is indeed the case for aspectual verbs, but not for modal verbs: the (a)-examples in (70) show that the string een ijsje kopen can be pronominalized by hetit or datthat with modal verbs; the (b)-examples show that this is not possible with aspectual verbs (although speakers accept left dislocation constructions such as Een ijsje kopen, dat gaan we zeker!Buying an ice cream, we certainly will!; cf. Section 4.6, sub II, for reasons to assume that this does not involve pronominalization.

One could try to solve this problem with modal verbs by assuming that example (70a') actually contains a phonetically empty verb ∅ corresponding to the semantically light verb doento do in (71a), but this would leave unexplained why this verb cannot co-occur with the aspectual verb gaan.

Furthermore, this line of thinking would lead us to expect the modal verb willen to exhibit the IPP-effect regardless of whether the clause contains the verb doen or its phonetically empty counterpart ∅. The examples in (72) show that this expectation is not borne out: the effect does not occur when doen is absent.

Finally, the fact illustrated in (73) that modal verbs differ from aspectual verbs such as gaan in that they can be combined with a nominal object is problematic for the view that the former is a non-main verb.

The above examples should make it clear that it is not a priori clear that the question as to whether or not a particular verb can be used as the sole verb of a clause is a cast-iron proof for establishing whether or not that particular verb is a main verb. We will return to this issue in Section 4.6.

Semi-aspectual verbs correspond to main verbs like zittento sit, liggento lie, hangento hang and staanto stand in (74), which refer to a certain posture or position of the subject of the clause, as well as to certain verbs of movement such as lopento walk.

In their semi-aspectual use, the lexical meaning of the main verb may or may not be present; examples like those in (75) can be comfortably used when the speaker cannot observe the referent of the subject of the clause and thus cannot tell whether that referent is actually sitting or walking at the moment of speech. The primary function of the semi-aspectual verb is to indicate that we are dealing with an ongoing event; we are dealing with a progressive construction comparable to the English progressive construction, as in the renderings in (75).

The examples in (75) also show that semi-aspectual verbs differ from the modal and aspectual verbs in (68) in that they do not combine with bare infinitivals but with so-called te-infinitivals: leaving out the infinitival marker te leads to ungrammaticality. This is not the case in the corresponding perfect-tense constructions in (76), where the marker te can (sometimes preferably) be omitted. The examples in (76) also show that the semi-aspectual verbs exhibit the IPP-effect; replacing the infinitive zitten/lopen with the past participle gezeten/gelopen leads to ungrammaticality.

The previous subsections have shown that non-main verbs impose certain restrictions on the morphological form of the main verb: temporal and passive auxiliaries select participles, modal/aspectual verbs select bare infinitives, and (finite) semi-aspectual verbs select te-infinitives. What we have not yet shown is that non-main verbs obligatorily form a verbal complex with the main verb they select; the main verb refers to a state of affairs, and the non-main verbs function as modifiers providing additional information. This is clear from the fact that an embedded main verb cannot normally be the head of an independent finite clause introduced by the complementizer datthat or an infinitival clause introduced by the complementizer om. We illustrate this in (77) for the aspectual verb gaan and the semi-aspectual verb zitten; the number sign # is used to indicate that (77b) is possible when zitten is interpreted as a main verb and the infinitival clause is interpreted as an adverbial purpose clause: “Jan sits in order to read the book”.

Note that this test again shows that a modal verb such as willento want can be used as a main verb; cf. the discussion of (72) in Subsection II. We will return to this issue in Section 4.6.

All examples in the above subsections are main clauses with the non-main verb in second position. In most varieties of Dutch spoken in the Netherlands, the auxiliaries cluster with the main verb in clause-final position; the arguments of the main verb must precede the non-main verb, even if the main verb follows it. This clause splitting is illustrated in (79) for the perfect auxiliary hebbento have, the modal verb willento want, and the semi-aspectual verb zittento sit.

Note, however, that certain southern varieties of Dutch (including the standard variety spoken in Belgium) do allow the object to intervene between non-main verbs and (te-)infinitives, hence the use of the percentage sign in (79b&c). See Barbiers (2008: §2.3.1) and Chapter 7 for more information.

The previous subsections have shown that auxiliaries must be accompanied by a main verb in the same clause. Furthermore, non-main verbs impose restrictions on the form of the main verb they select: temporal and passive auxiliaries select participles, modal and aspectual verbs select bare infinitivals, and (finite) semi-aspectual verbs select te-infinitivals. Non-main verbs do not combine with clauses introduced by the complementizer dat or om, which strongly suggests that non-main verbs must form a single verbal complex with a main verb. Finally, we have seen that in the varieties of Dutch spoken in the Netherlands, clauses with non-main verbs exhibit monoclausal behavior in the sense that they induce verb clustering; this results in the splitting of the projection of the main verb when the non-main verb is in clause-final position.