• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
7.3.The linear order of verbs in verb clusters
quickinfo

Section 7.2 has discussed the hierarchical order of verbs in verb clusters, and has shown that hierarchical order does not correspond in a one-to-one fashion to linear order. For example, verb clustering may linearize the hierarchical structure in (112a) in various ways, as indicated in the (b)-examples.

112
a. Jan [moet [hebben [de film gezien]]].
b. dat Jan die film moet hebben gezien.
b'. dat Jan die film moet gezien hebben.
b''. dat Jan die film gezien moet hebben.

In order to be able to discuss in a satisfactory way the linearization of verb clusters, it is important to determine which strings of verbs do (not) constitute instantiations of such clusters; here we assume that the reader is familiar with the discussion of this issue in Section 7.1. That section also suggested that if we put aside those strings of verbs that do not make up verb clusters, the linearization of standard Dutch verb clusters can be described by means of the three generalizations in (113).

113
a. Generalization I: Past/passive participles either precede or follow their governing auxiliary.
b. Generalization II: T e-infinitives follow their governing verb.
c. Generalization III: Bare infinitives follow their governing verb (in clusters consisting of three or more verbs).

The present section investigates the linearization of verb clusters in more detail by taking these generalizations as its point of departure, and shows that they indeed provide a descriptively adequate account of the attested word order patterns found in standard Dutch, although we will also point out a number of complications.
      Subsection I starts with a description of clusters of two verbs, subsection II continues with clusters of three (and more) verbs. The literature on verb clusters normally focuses on verb clusters including a finite verb, that is, clusters in finite embedded clauses such as (114a), but we will also look at the counterparts of such clusters in (extraposed) infinitival clauses such as (114b).

114
a. Marie denkt [dat Jan dat boek probeert te lezen].
  Marie believes  that  Jan that book  tries  to read
  'Marie thinks that Jan is trying to read that book.'
b. Marie verzocht Jani [om PROi dat boek te proberen te lezen].
  Marie requested  Jan  comp  that book  to try  to read
  'Marie requested Jan to try to read that book.'

Furthermore, we will diverge from general practice by also discussing the word order of verb clusters in main clauses such as (115), that is, clauses in which the finite verb is not part of the cluster but occupies the second position of the clause. Of course, this only makes sense in structures with more than two verbs. Although it might be defensible to claim that (115) involves a clause-final cluster of no more than two verbs, we will discuss such examples in the discussion of verb clusters of three verbs for practical reasons.

115
Jan wil dat boek proberen te lezen.
  Jan wants  that book  try  to read
'Jan wants to try to read that book.'

For an introduction to the notational conventions that will be used in the discussion in the following sections, we refer the reader to Section 7.2, sub I.

readmore
[+]  I.  Clusters of two verbs

This section discusses the linearization of verb clusters of two verbs. In order to be able to evaluate the generalizations in (113), we will divide such clusters on the basis of the morphological form of the embedded main verb, as in (116). The numeral indices express the hierarchical relation between the verbs in question: Vi+1–Vi indicates that Vi+1 is superior to Vi, due to the fact that the former verb selects the projection of the latter verb as its complement.

116
Verb clusters of two verbs
a. Aux2 + past/passive participle1
b. V2 + te-infinitive1
c. V2 + bare infinitive1
[+]  A.  Aux2 + Participle1:perfect-tense and passive constructions

There are two types of verb clusters of the type Aux2 + Participle1, one with a perfect and one with a passive auxiliary. These will be discussed in separate subsections.

[+]  1.  Perfect-tense constructions

The examples in (117) show that past participles may either precede or follow the finite perfect auxiliary. When we consider the regional spread of the two word orders, it seems that the order Aux2–Part1 is only found in a restricted part of the Dutch-speaking area, which happens to include the prestigious varieties of the standard language spoken in the west/middle region of this area; the maps in Pauwels (1953), Gerritsen (1991) and Barbiers et al. (2005) all show that this order is rare in the varieties of Dutch spoken in Flanders and the more northern part of the Netherlands.

117
a. dat Jan dat boek <gelezen> heeft <%gelezen>.
  that  Jan that book    read  has
  'that Jan has read that book.'
b. dat Marie naar Utrecht <gewandeld> is <%gewandeld>.
  that  Marie to Utrecht    walked  is
  'that Marie has walked to Utrecht.'

Observe that, for ease of parlance, we will follow the general practice of describing the difference in regional distribution of these orders as a north/south or Dutch/Flemish distinction, but the reader should be aware that the varieties spoken in the more northern region of the Netherlands pattern with the southern/Flemish region in this respect.
      Speakers who allow the order Aux2–Part1 normally also allow the order Part1– Aux2. There is reason for assuming that the latter order (part1–aux2) is in fact the unmarked one for such speakers given that Barbiers et al. (2005) found that they rarely invert this order in reproduction tasks. It seems generally accepted now that the use of the Aux2–Part1 order is characteristic of written Dutch and the more formal registers of spoken Dutch (despite that it also frequently occurs in the more casual speech of many speakers); see Haeseryn (1990:ch.2) for a good review of the relevant literature on this issue. A corpus analysis by De Sutter (2005/2007) suggests that even in written Dutch the Aux2–Part1 order is a secondary one given that this order is mainly used in relatively simple sentences; there is a negative correlation between the complexity of utterances and the frequency of the Aux2–Part1 order. We refer the reader to Section 6.2.1, sub III, for further discussion of such performance factors, and simply assume that standard Dutch allows the Aux2–Part1 order as a stylistically marked option.
      The examples in (118) show that we find basically the same variation in te-infinitivals in extraposed position: both orders are acceptable (and occur frequently on the internet). It seems reasonable to assume that the Part1–Aux2 order is again the unmarked one, but to our knowledge this has not been investigated so far.

118
a. dat Jan denkt het boek al <gelezen> te hebben <%gelezen>.
  that  Jan thinks  the book  already    read  to have
  'that Jan believes to already have read that book.'
b. dat Jan denkt al van zijn ziekte <hersteld> te zijn <%hersteld >.
  that  Jan thinks  already  from his illness  recovered  to be
  'that Jan believes to already have recovered from his illness.'
[+]  2.  Passive constructions

Like past participles, passive participles may either precede or follow their auxiliary in the northern varieties of Standard Dutch, but it seems that the relative frequency of the order Aux2–Part1 is lower in passives than in perfect-tense constructions. The southern varieties are reported to allow the Part1–Aux2 order only; we indicated this in (119) by means of a percentage sign. See Haeseryn (1990: Section 2.2) and De Sutter (2005/2007) for detailed discussion.

119
a. dat er buiten <gevochten> wordt <%gevochten>.
impersonal passive
  that  there  outside    fought  is
  'that people are fighting outside.'
b. dat hij door de politie <gevolgd> wordt <%gevolgd>.
regular passive
  that  he  by the police    followed  is
  'that heʼs followed by the police.'
c. dat ze een baan <aangeboden> kreeg <%aangeboden>.
krijgen-passive
  that  she  a job    prt-offered  got
  'that she was offered a job.'

That both orders are possible is confirmed by the infinitival passive constructions in (120), which show that te-infinitivals in extraposed position allow both orders in standard Dutch. We believe that the Part1–Aux2 order is again the preferred one, especially in the case of the krijgen-passive. This seems confirmed by a Google search (6/3/2013): whereas the string [ aangeboden te krijgen] resulted in 374 hits, the string [ te krijgen aangeboden] resulted in no more than 68 hits, several of which did not instantiate the intended passive construction.

120
a. Jan beweert door de politie <gevolgd> te worden <%gevolgd>
  Jan claims  by the police  followed  to be
  'Jan claims to be followed by the police.'
b. Jan denkt snel een baan <aangeboden> te krijgen <%aangeboden>.
  Jan  thinks  soon  a job    prt-offered  to get
  'Jan believes to be offered a job soon.'

Observe in passing that infinitival impersonal passive constructions do not occur. The reason for this is not immediately clear but may be related to the fact that propositional verbs like beweren'to claim' and denken'to think' trigger subject control, that is, require there to be an overt PRO-subject in the infinitival clause.

[+]  3.  Conclusion

The findings in this section are entirely in line with generalization I in (113a): past/passive participles either precede of follow their governing auxiliary. It should be noted, however, that the Aux2–Part1 order is a stylistically marked one, which may not be part of Dutch core grammar but of the periphery (consciously learned part) of the grammar; taking this position seems consistent with the fact that this order has been promoted for a long time by normative grammarians; see Section 6.2.1, sub III, for discussion. If so, we may simplify (113a) by saying that the participle must precede the auxiliary; although we will not take this step here for the northern varieties of Standard Dutch, this indeed seems necessary in order to provide a descriptively adequate account of the variety of Standard Dutch spoken in Flanders.

[+]  B.  V2 + te-infinitive1

In clusters of the type V2 + te-infinitive1, the superior verb V2 can be a main verb like the control verb proberen'to try' or the subject raising verb schijnen'appear', or a semi-aspectual main verb like zitten'to sit'. Given that these clusters all behave in the same way when it comes to linearization, it does not seem useful to discuss these cases in separate subsections. The clusters always behave in conformity with generalization II in (113b): te-infinitives follow their governing verb.

121
a. dat Jan dat boek <*te lezen> probeert <te lezen>.
Control
  that  Jan that book      to read  tries
  'that Jan is trying to read that book.'
b. dat Jan dat boek <*te lezen> lijkt <te lezen>.
Subject Raising
  that  Jan that book      to read  appears
  'that Jan appears to be reading that book.'
c. dat Jan dat boek <*te lezen> zit <te lezen>.
Semi-aspectual
  that  Jan that book      to read  sits
  'that Jan is reading that book.'

Given this finding, it does not come as a surprise that we find the same ordering restriction in the extraposed te-infinitivals in (122). We did not include cases with schijnen: infinitival clauses with evidential modal verbs normally give rise to a semantically infelicitous result. The rare examples with schijnen'appear', lijken'seem' and blijken'turn out' that we encountered on the internet do, however, behave in conformity with generalization II.

122
a. dat Jan ontkent dat boek <*te lezen> te proberen <te lezen>.
  that  Jan denies  that book     to read  to try
  'that Jan denies to be trying to read that book.'
b. dat Jan ontkent dat boek <*te lezen> te zitten <te lezen>.
  that  Jan denies  that book     to read  to sit
  'that Jan denies to be reading that book.'
[+]  C.  V2 + bare infinitive1

Although bare infinitives normally follow their governing verb, it has been observed that this is not always the case in clusters of two verbs. This has been observed for modal verbs in Reuland (1983), Den Besten & Broekhuis (1989), Koopman (1994) and Haeseryn et al. (1997:1072-3).

123
a. dat hij het vliegtuig niet <zien> kan <zien>.
  that  he  the airplane  not    see  is.able
  'that he canʼt see the airplane.'
b. dat hij haar <spreken> moet <spreken>.
  that  he  her    speak  must
  'that he must speak to her.'

The stylistically marked Main1-Modal2 order is pervasive in especially somewhat older literary prose and poetry, but can also be found in the literary work of the last century. For example, a manual search in Vestdijk's (600 page) novel Kind tussen vier vrouwen, which was written in 1933, resulted in 24 cases for the verb kunnen'may/be able', 6 cases for moeten'must/be obliged', 3 cases for mogen'be allowed', 8 cases for willen'want', and 31 cases for zullen'will'. The same novel also provided 8 cases with the aspectual verb gaan'to go'; examples are given in (124).

124
a. ... alsof Jan Breedevoort hem knijpen ging.
Verzamelde Romans 1, 378
  as.if  Jan Breedevoort him  pinch  went
  '... as if Jan Breedevoort was going to pinch him.'
b. ... alsof hij hen [...] de keel afsnijden ging.
Verzamelde Romans 1, 473
  as.if  he  them  the throat  prt.-cut  went
  '... as if he was going to cut their throats.'

There seems to be some disagreement in the literature on the question as to whether perception verbs allow the deviant order in AcI-constructions: Reuland (1983) claims that such orders are unacceptable, Haeseryn et al. consider them archaic, and Den Besten & Broekhuis (1989) and Koopman (1994) regard them as acceptable. For this reason we marked the examples in (125), adapted from Reuland and Den Besten & Broekhuis, with a percentage sign.

125
a. dat Marie Peter de ratten <%vangen> zag <vangen>.
  that  Marie Peter the rats       catch  saw
  'that Marie saw Peter catch the rats.'
b. dat Marie hem <%lopen> zag <%lopen>.
  that  Marie him       walk  saw
  'that Marie saw him walk.'

Examples with perception verbs were not found in Vestdijk's novel (although they can be encountered elsewhere), but it does have cases of AcI-constructions with laten'to make/let': example (126a) involves permissive and (126b) causative laten. Examples of this sort are also accepted by Den Besten & Broekhuis, but Koopman (1994) claims that examples like these are acceptable with a permissive reading only; examples like these are not discussed by Reuland and Haeseryn et al.

126
a. ... zoals een poes een gewond muisje nog [...] trippelen laat.
VR 1, 226
  like  a cat  an injured mouse  still   trip  let
  '... like a cat lets an injured mouse trip for a while.'
b. Ik wil dat je het vandaag lezen laat.
VR 1, 387
  want  that  you  it  today  read  make
  'I want that you make [someone] read it today.'

That the order Main1-Modal2 is fairly special is clear from the fact that it can only occur if certain special conditions are met. Den Besten & Broekhuis note, for example, that this order is less acceptable if the object of the embedded main verb is indefinite and in a position adjacent to the verb cluster; this is illustrated in (127). They further suggest that this restriction is prosodic in nature, but since this suggestion has not been tested so far, we leave it to future research to investigate whether it is on the right track.

127
a. dat Marie dat boek waarschijnlijk lezen wil.
  that  Marie that book  probably  read  wants
  'that Marie probably wants to read that book.'
b. ? dat Marie waarschijnlijk een boek lezen wil.
  that  Marie probably  a book  read  wants
  Intended: 'that Marie probably wants to read a book.'

That the order Main1-Modal2 is special is also clear from the fact that it cannot occur in infinitival clauses. We illustrate this in (128) for clusters with a superior modal verb only.

128
a. Jan beweerde het vliegtuig niet <*zien> te kunnen <zien>.
  Jan  claimed  the airplane  not      see  to be.able
  'Jan claimed not to be able to see the airplane.'
b. Jan hield vol haar <*spreken> te moeten <spreken>.
  Jan insisted  prt.  her      speak  to had.to
  'Jan insisted on having to speak to her.'

As far as we know, it has not been investigated to what extent the stylistically marked order Main1-Modal2 occurs in spontaneous speech of speakers of Standard Dutch, and consequently it is not clear whether it should be considered part of Dutch core grammar or of its periphery. This issue is important given that it may affect our evaluation of the various theoretical accounts of verb clustering. We have to leave the issue to future research for want of relevant information. We refer the reader to Barbiers (2008: Section 1.3.1) for a discussion of the dialectal distribution of the two word orders.

[+]  D.  Summary and generalizations

The subsections above investigated the generalizations in (113), repeated here in a slightly different form as (129). The generalizations as formulated here can account for the unmarked word orders in verb clusters of two verbs.

129
a. Generalization I: Past/passive participles either precede or follow their governing auxiliary.
b. Generalization II: T e-infinitives follow their governing verb.
c. Generalization III: Bare infinitives follow their governing verb.

It should be noted that generalization I is too permissive for the southern varieties of Standard Dutch, which seem to require the participle to precede the auxiliary. The formulation of generalization III in (129c) differs from the one in (113c) in that we omitted the supplementary clause that the generalization is restricted to clusters with more than two verbs. The reason for doing this is that it is not a priori clear at this point whether the order Main1-Modal2 should be considered part of Dutch core grammar: it may be restricted to the written/formal register and thus be part of the periphery of the grammar.

[+]  II.  Clusters of three or more verbs

This section discusses the linearization of verb clusters of three (or more) verbs. In order to be able to evaluate the generalizations in (129), we will classify such clusters on the basis of the morphological form of the most deeply embedded main verb, as in (130). The numeral indices express the hierarchical relation between the verbs in question: Vi+1–Vi indicates that Vi+1 is superior to Vi since the former verb selects the projection of the latter verb as its complement.

130
Verb clusters of three verbs
a. V3 + Aux2 + past/passive participle1
b. V3 + V2 + te-infinitive1
c. V3 + V2 + bare infinitive1

It is easily possible to form verb clusters of four or more verbs, but these are relatively rare in everyday use; a more or less natural example is dat Jan dat boek zou moeten hebben kunnen lezen'that Jan should have been able to read that book'. The principles that underlie the word order of such clusters do not differ from those that underlie the order of clusters of three verbs. We will therefore not systematically discuss such larger clusters, but simply discuss some cases if expedient. The following subsections will discuss the clusters in (130) in the order given there.

[+]  A.  V3 + Aux2 + Participle1: perfect-tense and passive constructions

Past participles arise if a perfect auxiliary immediately governs the most deeply embedded main verb Main1; if a perfect auxiliary governs some higher verb Vn, where n > 1, we normally get the infinitivus-pro-participio (IPP) effect. This is illustrated in (131).

131
a. dat Jan dat boek morgen moet hebben gelezen.
Modal3-Aux2-Main1
  that  Jan  that book  tomorrow  must  have  readpart
  'that Jan has to have read that book by tomorrow.'
b. dat Jan dit boek heeft moeten/*gemoeten lezen.
Aux3-Modal2-Main1
  that  Jan this book  has  must/mustpart   read
  'that Jan has had to read that book.'

Passive participles are also found as the as the single most deeply embedded main verb (= Main1) only, for the simple reason that passivization of some higher verb Vn, where n >1, is normally not possible.

132
dat de radio moet worden gerepareerd.
Modal3-Aux2-Main1
  that  the radio  must  be  repaired
'that the radio must be repaired.'

Consequently, when discussing the linear order of verb clusters with a past/passive participle, we can focus on strings of the form V3 + Aux2 + Participle1. We will show that generalization I, according to which past/passive participles either precede of follow their governing auxiliary is correct for the variety of Standard Dutch spoken in the Netherlands, but not for that spoken in Belgium. We will further show that the participles need not be adjacent to their auxiliary but can actually occur in several positions in the cluster. We conclude with a discussion of one notable exception to the otherwise robust generalization that participles are the most deeply embedded verb in verb clusters, viz., cases in which a passive auxiliary is governed by a perfect auxiliary.

[+]  1.  Perfect-tense constructions

We start our discussion of perfect-tense constructions with main clauses, that is, structures in which the finite verb is in second position. Structures of this type do not seem to show an exceptional behavior: the examples in (133) show that the past participle may either precede or follow the auxiliary. We should, however, make the same proviso as in Subsection IA, that the Aux2–Part1 order is only found in a restricted part of the Dutch-speaking area, which happens to include the prestigious varieties of the standard language spoken in the west/middle region of this area. More generally, the Part1–Aux2 order seems to be the more common one in speech.

133
a. Jan moet dat boek morgen <gelezen> hebben <%gelezen>.
  Jan must  that book  tomorrow    readpart  have
  'Jan must have read that book by tomorrow.'
b. Els zal vanmorgen <vertrokken> zijn <%vertrokken>.
  Els will  this.morning    leftpart  be
  'Els will have left this morning.'

The examples in (134) show that the placement options of past participles in embedded clauses are a little surprising. As the participle is governed by the auxiliary, we would expect these verbs to be adjacent, but as a matter of fact they can easily be separated by the finite modal verb.

134
a. dat Jan dat boek <gelezen> moet <gelezen> hebben <%gelezen>.
  that  Jan that book    readpart  must  have
  'that Jan must have read that book.'
b. dat Els vanmorgen <vertrokken> zal <vertrokken> zijn <%vertrokken>.
  that  Els this.morning    leftpart  will  be
  'that Els will have left this morning.'

For many speakers, the three word orders are simply more or less free alternatives, with the modal3–Aux2–Part1 order moet hebben gelezen again being the stylistically most marked one. The varieties of standard Dutch spoken in the Netherlands and Belgium also seem to differ in that they exhibit different order preferences: several types of research reveal that speakers from the Netherlands prefer the part1–modal3–aux2 order gelezen moet hebben, whereas speakers from Belgium prefer the modal3–part1–aux2 order moet gelezen hebben. Other orders can be attested in some varieties of Dutch, but these are normally considered to be dialectal in nature; see Section 6.2.1, sub IV, for a more detailed discussion.
      That speakers from the Netherlands have a preference to put the participle first in the verb cluster is also clear from the extraposed te-infinitivals in (135); placement of the participle in position <2> gives rise to a degraded result for these speakers, whereas some of our Flemish informants readily accept this placement. Placement of the participle in position <1> is again restricted to the variety of Standard Dutch spoken in the Netherlands. Note that there is not much information about the regional spread of the verb orders in (135), so more careful research would be welcome.

135
a. Jan beweert dat boek morgen <gelezen> te moeten <2> hebben <1>.
  Jan claims  that book  tomorrow  readpart  to must  have
  'Jan claims to have to have read that book by tomorrow.'
b. Els zegt morgen al <vertrokken> te zullen <2> zijn <1>.
  Els says  tomorrow  already    left  to will  be
  'Els says that she will already have left tomorrow.'

The examples in (136) provide similar instances with a subject raising verb such as schijnen, which does not trigger extraposition of its infinitival complement but instead requires verb clustering; note that while (136a) is quite natural, some speakers may consider (136b) somewhat artificial due to the fact that more or less the same message can be expressed without the modal zullen. Placement of the participle in position <2> again gives rise to a degraded result for speakers from the Netherlands, whereas some of our Flemish informants have no qualms about accepting it. Placement of the participle in position <1> is again restricted to the Dutch variety of standard Dutch. Again, it should be mentioned that more careful research on the regional spread of the orders in (136) would be welcome.

136
a. Jan schijnt dat boek morgen <gelezen> te moeten <2> hebben <1>.
  Jan seems  that book  tomorrow     readpart  to must  have
  'Jan seems to have to have read that book by tomorrow.'
b. Els schijnt morgen al <vertrokken> te zullen <2> zijn <1>.
  Els seems  tomorrow  already    left  to will  be
  'It seems that Els will already have left tomorrow.'

      Clusters with more than three verbs are possible but not very common in colloquial speech. It seems that participles can appear in all positions in the cluster, as is illustrated in (137) by means of the embedded counterparts of (136a). Example (137a) and (137b) seem again restricted to the varieties of standard Dutch spoken in, respectively, the Netherlands and Flanders. The orders in (137c) and, especially, (137d) seem to be the more generally accepted ones. It goes without saying that more careful research on the regional spread of these orders would be welcome.

137
a. % dat Jan dat boek morgen schijnt te moeten hebben gelezen.
  that  Jan that book  tomorrow  seems  to must  have readpart
  'Jan seems to have to have read that book by tomorrow.'
b. % dat Jan dat boek morgen schijnt te moeten gelezen hebben.
c. dat Jan dat boek morgen schijnt gelezen te moeten hebben.
d. dat Jan dat boek morgen gelezen schijnt te moeten hebben.

Clusters with four verbs in which the superior non-finite verbs are all bare infinitives have been researched in more detail. The literature reviewed in Haeseryn (1990:70ff.) suggests that the orders in (138a&d) are the ones commonly found in the northern varieties of Standard Dutch, and that the order in (138c) is more favored than the one in (138b). In the varieties of Standard Dutch spoken in Belgium, on the other hand, the order in (138b) seems to be a common one.

138
a. % dat Jan die film zou kunnen hebben gezien.
  that  Jan that movie   wouldmodal  maymodal  haveaux  seenmain
  'that Jan could have seen that movie.'
b. dat Jan die film zou kunnen gezien hebben.
c. dat Jan die film zou gezien kunnen hebben.
d. dat Jan die film gezien zou kunnen hebben.

These acceptability judgments on the examples in (138) seem to be in line with what we found for the examples in (137), but an important difference is that all orders in (138) seem acceptable in the variety of Standard Dutch spoken in the Netherlands: while speakers of this variety consider examples such as (137b) to be degraded, examples such as (138b) are merely considered to be stylistically marked.

[+]  2.  Passive constructions

We start our discussion of passive constructions with main clauses, that is, structures in which the finite verb is in second position. Structures of this type again seem to be quite ordinary in that the examples in (139) show that the passive participle may either precede or follow the auxiliary, with the proviso that the aux-part order is only found in a restricted part of the Dutch-speaking area which happens to include the prestigious varieties of the standard language spoken in the west/middle region of this area. More generally, it seems that the part-aux order is the more common one in speech.

139
a. Er zal buiten <gevochten> worden <%gevochten>.
impersonal passive
  there  will  outside    fought  be
  'People will be fighting outside.'
b. Hij moet door Marie <geholpen> worden <%geholpen>.
regular passive
  he  must  by Marie    helped  be
  'He needs to be helped by Marie.'
c. Zij zal de baan <aangeboden> krijgen <%aangeboden>.
krijgen-passive
  she  will  the job    prt-offered  get
  'Sheʼll be offered the job.'

The examples in (140) show that in embedded clauses, the passive participle may occupy any position in the clause-final verb cluster in the northern varieties of Dutch, although placement of the participle in final position seems less frequent than in the perfect-tense construction, and that intermediate placement is relatively rare. The southern varieties do not allow the participle in final position and further seem to differ from the northern varieties in exhibiting a preference for placing the participle in the intermediate position of the verb cluster. We refer the reader to Haeseryn (1990: Section 2.3.2) for a more detailed discussion of these regional differences in frequency.

140
a. dat er buiten <gespeeld> mag < gespeeld > worden <%gespeeld >.
  that  there  outside    played  be.allowed  be
  'It will be allowed to play outside.'
b. dat hij door Marie <geholpen> moet <geholpen> worden <%geholpen>.
  that  he  by Marie   helped  must. be
  'that he needs to be helped by Marie.'
c. dat ze de baan <aangeboden> zal <aangeboden> krijgen <%aangeboden>.
  that  she  the job    prt-offered  will  get
  'that sheʼll be offered the job.'

That speakers from the Netherlands prefer to place the participle first in the verb cluster is also clear from the extraposed te-infinitivals in (141), in which placement of the participle in position <2> gives rise to a degraded result; cf. Smits (1987). Some of our Flemish informants, on the other hand, do allow placement of the participle in position <2>. Placement of the participle in position <1> is again restricted to variety of standard Dutch spoken in the Netherlands. Note that we do not provide examples of the impersonal passive as these cannot occur in infinitival clauses of this type for independent reasons; cf, subsection IA.

141
a. Jan beweert door Marie <geholpen> te moeten <2> worden <1>.
  Jan claims  by Marie    helped  to must  be
  'Jan claims that he needs to be helped by Marie.'
b. Zij denkt een baan <aangeboden> te zullen <2> krijgen <1>.
  she  thinks  a job     prt.-offered  to will  get
  'She thinks that sheʼll get offered a job.'

The examples in (141) involve the propositional verb beweren, which triggers extraposition of its infinitival complement. In (142), we find similar examples with the subject raising verb schijnen; note that whereas the (a)- and (b)-examples are quite natural, some speakers may consider the (c)-example artificial as more or less the same message can be expressed without the modal zullen. Placement of the participle in position <2> again gives rise to a degraded result for speakers from the Netherlands, whereas some of our Flemish informants are quite comfortable with this placement. Placement of the participle in position <1> is again restricted to the variety of standard Dutch spoken in the Netherlands.

142
a. Er schijnt buiten gespeeld te mogen <2> worden <1>.
  there  seems  outside  played  to be.allowed  be
  'It seems to be allowed to play outside.'
b. Jan schijnt door Marie <geholpen> te moeten <2> worden <1>.
  Jan seems  by Marie     helped  to must  be
  'It seems that Jan needs to be helped by Marie.'
c. Zij schijnt een baan <aangeboden> te zullen <2> krijgen <1>.
  she  seems  a job     prt.-offered  to will  get
  'It seems that sheʼll get offered a job.'

The embedded counterparts of (142) exhibit more or less the same pattern; we demonstrate this in (143) for the regular passive in (142b) only. The percentage signs in (143a) and (143b) again express that the marked orders are restricted to the variety of standard Dutch spoken in, respectively, the Netherlands and Flanders. The orders in (137c) and, especially, in (137d) seem to be the more generally accepted ones.

143
a. % dat Jan door Marie schijnt te moeten worden geholpen.
  that  Jan by Marie  seems  to must  be  helped
  'that Jan seems to need to be helped by Marie.'
b. % dat Jan door Marie schijnt te moeten geholpen worden.
c. dat Jan door Marie schijnt geholpen te moeten worden.
d. dat Jan door Marie geholpen schijnt te moeten worden.

The clusters in (143) contain a te-infinitive as a non-finite superior verb. Clusters with four verbs in which the superior non-finite verbs are all bare infinitives have been researched in greater detail. The literature reviewed in Haeseryn (1990:70ff.) suggests that the orders in (144a&d) are the ones commonly found in the northern varieties of Standard Dutch, and that the order in (144c) is more favored than the one in (144b). In the varieties of Standard Dutch spoken in Belgium, on the other hand, the order in (144b) seems to be a common one. This is in keeping with what we found for the examples in (143), but an important difference is that all orders in (144) seem acceptable for speakers of the variety of Standard Dutch spoken in the Netherlands: whereas such speakers consider examples such as (143b) as degraded, example (144b) is merely considered as stylistically marked.

144
a. % dat hij door Marie zou moeten worden geholpen.
  that  he  by Marie  would  must  be  helped
  'that he should be helped by Marie.'
b. dat hij door Marie zou moeten geholpen worden.
c. dat hij door Marie zou geholpen moeten worden.
d. dat hij door Marie geholpen zou moeten worden.

For completeness' sake, example (145) provides similar examples for the krijgen-passive, for which the same observations can be made as for (144).

145
a. % dat ze de baan zou moeten krijgen aangeboden.
  that  she  the job  would  must  get  prt-offered
  'that she should be offered the job.'
b. dat ze de baan zou moeten aangeboden krijgen.
c. dat ze de baan zou aangeboden moeten krijgen.
d. dat ze de baan aangeboden zou moeten krijgen.
[+]  3.  Summary

The subsections above have shown that perfect-tense and passive constructions behave in full accordance with generalization I in (129a): past participles may follow or precede the perfect auxiliary. In fact, participles seem to be able to occur in any position in the verb cluster. This is illustrated in (146), in which Vn stands for zero or more verbs in the verb cluster besides the auxiliary and the main verb.

146
Order in verb clusters of the form Vn + Aux2 + Part1
a. dat ..... <Part> Auxfinite <Part>
b. dat ..... <Part> Vfinite <Part> Aux <Part>
c. dat ..... <Part> Vfinite <Part> Vinf <Part> Aux <Part>
d. dat ..... <Part> Vfinite <Part> Vinf <Part> Vinf <Part> Aux <Part>
e. etc.

The order Aux2–Part1 seems, however, to be a stylistically marked one that is restricted to the northern varieties of standard Dutch. In the southern varieties we tend to find the pattern in (147).

147
Order in verb clusters of the form Vn + Aux2 + Part1
a. dat ..... <Part> Auxfinite
b. dat ..... <Part> Vfinite <Part> Aux
c. dat ..... <Part> Vfinite <Part> Vinf <Part> Aux
d. dat ..... <Part> Vfinite <Part> Vinf <Part> Vinf <Part> Aux
e. etc.

The northern and southern varieties further seem to differ in that the former prefers the participle to come first in the verb cluster (e.g., part1–V3–aux2), whereas the latter prefers it to be in some intermediate position (e.g., V3–part1–aux2). The northern varieties further seem to be special in that they prohibit placement of the participle between a te-infinitive and the auxiliary: * ... Vte-inf <Part> Aux.

[+]  4.  A special case: perfect passives

Passive constructions are special in that they do not exhibit the IPP-effect in the perfect tense: this implies that passive constructions constitute an exception to the general rule that verb clusters do not contain more than one participle. This is illustrated in (148) by means of a krijgen-passive; the past/passive participles are italicized.

148
a. dat Jan het boek