The examples in (77) show that in passive clauses with two verbs (the auxiliary and the passivized main verb), the main verb always has the form of a passive participle; using an infinitival main verb leads to ungrammaticality.

In embedded clauses, the auxiliary and the passive participle are both in clause-final position, and, as might be expected on the basis of our discussion of perfect-tense constructions in Section 6.2.1, sub III, the passive participle can either precede or follow the passive auxiliary in the northern varieties of standard Dutch, although it should be noted that the order aux-part is less frequent in passive constructions than in perfect-tense constructions; cf. the studies reviewed in Haeseryn (1990: §2.2) and De Sutter (2005/2007). The percentage signs indicate that the southern varieties allow only the part-aux order.

In sequences of three verbs, passive auxiliaries never appear as finite verbs, so it is not surprising that the passivized main verb always appears as a passive participle. The examples in (79) show that, as might be expected on the basis of our discussion of perfect-tense constructions in Section 6.2.1, sub IVA, the passive participle can occupy any position in the clause-final verb cluster in the northern varieties of Dutch, although it should be noted that the placement of the participle in final position is again less frequent than in perfect-tense constructions, and that intermediate placement seems to be relatively rare.

The percentage signs again indicate that the southern varieties do not allow the participle in clause-final position. These varieties also seem to differ from the northern varieties in that they prefer the participle in the intermediate position of the verb cluster; cf. Table (24) for a similar finding for perfect-tense constructions. For the discussion in Subsection IV, it is important to add that some speakers of the southern varieties do not allow the passive participle in intermediate position, but require it to be first in the verb cluster; cf. also Haeseryn (1990: §2.3.2).

Recall that auxiliaries never appear as finite verbs in sequences of three verbs; this means that perfect passive constructions require the perfect auxiliary to appear as the finite verb and the passive auxiliary to appear as a non-finite verb. Section 6.2.1, sub IVA, has shown that such perfect-tense constructions usually exhibit the infinitivus-pro-participio (IPP) effect: the verb governed by the perfect auxiliary does not appear as a participle, but as an infinitive. This is illustrated again in (80).

It is therefore surprising that the IPP-effect does not occur in perfect passive constructions. Since Subsection II has shown that the perfect-tense version of the regular passive in the northern varieties of Dutch is special in that it (possibly) has a covert passive auxiliary, we will first illustrate this for the krijgen-passive.

The absence of the IPP-effect in (81b) is not the only remarkable property of passive constructions in the perfect tense; the placement of the passivized main verbs is also special. The examples in (82) show that while the main verb can either precede or follow the auxiliary krijgen in imperfect constructions (in the northern varieties of Dutch), the main verb must precede the auxiliary in the corresponding perfect-tense constructions; cf. Den Besten (1985).

The examples in (83) show that larger verb clusters in which the passive auxiliary appears as a past participle exhibit more or less the same behavior: the participial main verb gestuurd can in principle be placed in all positions indicated by ✓, but not in the position after the participial passive auxiliary gekregen marked with an asterisk.

Whether we find the same effect in regular passive examples like those in (84) is difficult to answer: judgments from speakers of the southern variety of Dutch are not helpful, since such speakers do not easily allow the aux-part order in (84a) anyway, and speakers of the northern varieties consider the overt expression of the perfect auxiliary geworden in (84b) to be at best marked or archaic. To the extent that (84b) is accepted by northern speakers, however, they agree that the passive participle geslagen must precede the passive auxiliary geworden; placing the passive participle after geworden leads to a completely unacceptable result.

Many speakers of the southern varieties do accept the orders in (84b) marked by two question marks, possibly with the passive auxiliary geweest instead of geworden; cf. Subsection II. As in the case of the krijgen-passive, it should be added that not all speakers of the southern varieties allow the passive participle in intermediate position; some of our Flemish informants require it to be the first verb in the verb cluster.

Clause splitting in regular passive clauses is difficult to illustrate with the internal argument of a passivized transitive verb, because it appears as the derived theme-subject of the clause; thus, in examples such as (85a), it can be expected to occupy not its underlying VP-internal position, but its derived subject position. However, Section N21.1.2 has shown that subjects presenting new information need not be moved into the regular subject position, but can remain in their underlying position in the lexical domain of the clause. With this in mind, it is interesting to note that indefinite subjects in presentational er-constructions, which always present new information, cannot be adjacent to the main verb in passive constructions such as (85b) either. If we assume that such subjects occupy their underlying VP-internal position, the fact that they must precede the passive auxiliary can be used to argue that regular passive constructions exhibit clause splitting.

The same can be shown even more clearly by nominative-dative inversion in passive constructions with ditransitive verbs, which is discussed in more detail in Section 3.2.1.3, sub IIB; the fact that the theme-subject het/een boekthe/a book follows the indirect object Jan/hem in (86a&b) clearly shows that it need not occupy the regular subject position; the fact that it nevertheless cannot permeate the verb cluster again shows that regular passive constructions exhibit clause splitting. Note in passing that the unacceptability of the nominative-dative order in (86b) supports our earlier claim that the indefinite subject in (85b) does not occupy the regular subject position. Of course, the fact that the dative object Jan/hem must also precede the passive auxiliary werd further supports that we are dealing with clause splitting.

Clause splitting in impersonal/regular passives is illustrated by the examples in (87). There is no reason to assume that the placement of the prepositional object op een brief, the complementive oranje, or the manner adverb grondig is affected by passivization, but nevertheless these elements cannot occur left-adjacent to the main verb when the latter follows the passive auxiliary. Note that the prepositional object op een brief can be extraposed as usual, i.e. it can occur in the position following the verb cluster marked with ✓.

Example (88a) shows that, as in active perfect-tense constructions, the preverbal position marked with an asterisk in (87b) becomes available when we replace the complementive oranje with a monosyllabic adjective such as roodred. In this respect, monosyllabic adjectives again behave like verbal particles such as op in (88b), which can also permeate verb clusters.

Note that the type of evidence in examples (85) to (88) is not available for those varieties of Dutch that do not allow the aux-part order (i.e. the southern varieties of standard Dutch as well as the varieties spoken in the northern part of the Netherlands).

Clause splitting can also occur when the passive participle precedes the auxiliary. This is illustrated in (89) for the verbs wachtento wait and zeggento say, which take a prepositional and a clausal complement respectively. The primeless examples first show that PP-complements can either precede or follow their main verb, while clausal complements must follow their main verb. The primed examples show that, as in perfect-tense constructions, the complement-PP/clause cannot permeate the verb cluster, i.e. it cannot be placed between the participle and the passive auxiliary. We have also indicated that the PP-complement in (89a') can precede the whole verb cluster, whereas this is excluded for the complement clause in (89b'), as usual.

For those varieties of Dutch that do not allow permeation of the clause-final verb cluster, the word-order facts in clauses with three verbs are basically the same as in clauses with two verbs. The examples in (90) show that the derived theme-subject must precede the clause-final sequence, regardless of whether it is definite or indefinite. Since we have seen in the previous subsection that (at least) the indefinite subjects in presentational er-constructions can occupy their underlying VP-internal position, the possibility of clause splitting in the (b)-examples in (90) again supports the claim that passive constructions involve a verb cluster. The subject de/een hond cannot be placed in the positions marked by an asterisk or a percentage sign.

Since speakers of the southern varieties of Dutch do not accept the modal-aux-part order, we expect them to reject any order in the primeless examples in (90). It also seems that these varieties mutually differ in whether they take the modal-part-aux or the part-modal-aux order. The crucial point here is that our informants who accept the modal-part-aux order also allow the indefinite (but not the definite) subject to permeate the clause-final verb cluster; the placement of the subject in the position marked by a percentage sign in (90b') is acceptable to such speakers, although it is considered somewhat marked compared to the alternative placement in the position preceding the auxiliary.

Clause splitting cannot be demonstrated for the northern varieties of standard Dutch in passive perfect-tense constructions, because they usually require the omission of the participle form of the passive auxiliary (cf. Subsection II); examples such as (91) are categorically rejected by northern speakers. However, it is possible for the southern varieties that allow overt expression of the passive auxiliary (i.e. those with the auxperfect-partmain-auxpassive order). If permeation of the verb cluster is also allowed, the indefinite subject in (91b) can occur in the position marked by the percentage sign.

The examples in (92) provide passivized ditransitive constructions in which the derived theme-subject follows the indirect object and thus clearly does not occupy the regular subject position. First, speakers of the northern varieties of standard Dutch require the theme-subject to precede the entire verb cluster, regardless of its word order; the fact that it is impossible to place the subject in the positions marked by an asterisk or a percentage sign again shows that passive constructions allow clause splitting. Second, southern speakers who allow the modal-part-aux order also allow the indefinite theme-subject to permeate the verb cluster, i.e. to occur in the position marked by the percentage sign in (92b).

The examples in (93) illustrate the same for passive constructions in the perfect tense for those southern speakers who prefer the verb order in (93a) to that in (93b): such speakers also allow permeation of the verb cluster by the indefinite theme-subject. The percentage signs preceding these examples again indicate that this construction type is not available to northern speakers, because they require the omission of the passive auxiliary geweest.

Clause splitting can also occur with PP-complements, complementives and manner adverbs. The judgments of northern speakers on the examples in (94) are essentially the same as on the examples in (87): PP-complements, (polysyllabic) complementives, and manner adverbs cannot permeate the verb cluster; the italicized phrases can be placed in the positions indicated by ✓, but not in the positions marked by an asterisk or a percentage sign. Note that we have not marked the positions following the main verb for the complementive oranje and the manner adverb grondig, since these elements never follow the main verb in clause-final position.

We also expect that southern speakers who accept the auxperfect-part-auxpassive order in the singly-primed examples will also accept permeation of the verb cluster. Our informants who allow this order report that the orders marked with a percentage sign are indeed acceptable (although in the case of the prepositional object op een brief, this order is judged to be marked).

The (a)-examples in (95) show that monosyllabic adjectival complementives again differ from polysyllabic ones in that they behave like verbal particles in the sense that they can at least marginally permeate the verb cluster in the northern varieties of standard Dutch, provided that they precede the main verb.

The examples in (96) show that clause splitting can also occur in perfect passive constructions for southern speakers who can overtly express the passive auxiliary geweest, i.e. accept the auxperfect-part-auxpassive order of the verb cluster. The question mark in (96a) is added to indicate that such speakers consider this order acceptable, but marked. Examples such as (96) are categorically rejected by northern speakers because of the (overt) realization of geweest.

Finally, the examples in (97) show that clause splitting can also occur when the passive participle precedes the passive auxiliary. This is especially true for constructions with prepositional/clausal complements, which may/must follow the main verb in clause-final position, but cannot permeate verb clusters. We have marked all (im)possible placements of the complement PP/clause with respect to the verbs in the cluster, but of special interest are the impossible ones within the verb cluster after the main verb gewacht/gezegd.

Subsection III has shown that in the northern varieties of standard Dutch, passive participles can follow the passive auxiliary krijgen in sequences of two verbs. The primeless examples in (98) show that this can lead to clause splitting; direct objects and VP adverbials must precede the verb cluster as a whole, whereas verbal particles can permeate the verb cluster (as long as they precede the main verb). Because the elements in question never follow the main verb in clause-final position, we have only indicated the placements that conform to this general rule.

Note that we cannot illustrate clause splitting with complementives because verbs entering the krijgen-passive are typically particle verbs, and these do not allow the addition of a complementive; cf. Section 2.2.1, sub IV, for discussion. Note also that clause splitting in cases with sequences of the verbs cannot be shown for the southern varieties of Dutch because these do not accept the aux-part order; these varieties only have the orders in the primed examples.

In imperfect krijgen-passives with three verbs, the participle can occupy any position in the verb cluster in the northern varieties of Dutch. However, the placement of the dependents of the passivized main verb is much more restricted. The examples in (99) show that the options are more or less identical to those in (98); direct objects and VP adverbials must precede the whole verb cluster, whereas verbal particles can permeate it (as long as they precede the main verb).

The southern varieties that allow permeation of the verb cluster do not accept the primeless examples because of the placement of the participle, and may differ in their preference for the singly-primed or doubly-primed examples. For those varieties that do accept the singly-primed examples, we expect the orders marked with a percentage sign to be acceptable, and our Flemish informants tell us that this expectation is indeed borne out (although this is judged to be marked with the adverb geheel).

In perfect krijgen-passives, the participle must precede the passive auxiliary gekregen, as in (100). We expect that the southern varieties that allow the participle to follow the perfect auxiliary heeft, as in the primeless examples, will also allow permeation of the verb cluster. Our Flemish informants again indicate that this expectation is borne out; they consider all orders marked with a percentage sign to be perfectly acceptable.

For completeness, example (101) shows that clause splitting also occurs when a main verb that takes a complement clause precedes the passive auxiliary. Note that the clause cannot permeate verb clusters, but must follow the entire verb cluster; this is not illustrated here.

Since krijgen-passivization is possible only with ditransitive verbs, and since we are not aware of any clear examples of ditransitive verbs taking a prepositional object (cf. Section 2.3), we cannot illustrate clause splitting with this type of complement.