• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
4.2.2.3.The wat voor split
quickinfo

This section discusses in more detail the properties of the wat voor split. This split is generally assumed to be the result of moving the interrogative pronoun wat into clause-initial position, as in (246d). Splitting the wat voor phrase at some other point is excluded, as is illustrated in (246b&c).

246
a. Wat voor een boeken heb jij gelezen?
  what for a books  have  you  read
  'What kind of books did you read?'
b. * Wat voor een heb jij boeken gelezen?
c. * Wat voor heb jij eenboeken gelezen?
d. Wat heb jij voor een boeken gelezen?

It has been argued that the syntactic function of the wat voor phrase, along with its surface position in the clause, is relevant for the question as to whether the wat voor split is allowed. We will review the relevant data in I, and show that at least subjects and direct objects of various sorts of verbs allow the split, provided that they occupy their “base” position in the clause. The wat voor split is blocked not only by movement of the wat voor phrase, but also by the presence of certain other elements in the clause, such as the negative adverb niet'not'. This will be discussed in Subsection II, where we will also discuss so-called “parasitic gaps” licensed by a wat voor phrase.

readmore
[+]  I.  The syntactic function of the split phrase

Whether wat voor split is possible depends on the syntactic function of the phrase. The following subsections will show that direct objects, subjects and nominal predicates do allow the split, while indirect objects and complements of prepositional phrases do not. Further, it will be shown that the surface position of the stranded remnant of the wat voor phrase (from now on: remnant) may also bear on whether the split is possible or not. This is generally assumed to follow from the general prohibition of subextraction from a moved phrase, the so-called freezing principle.

[+]  A.  Direct objects

The examples in (247a&b) show that direct objects may undergo the wat voor split, but that the remnant must generally be left-adjacent to the verb(s) in clause-final position, that is, scrambling of the remnant, as in (247b), is excluded as an instantiation of thefreezing effect. PP-over-V also gives rise to a degraded result: speakers of Dutch may differ somewhat on their judgments of (247c), but all agree that it is marked compared to (247a).

247
a. Wati heb je gisteren [ ti voor (een) boeken] gelezen?
  what  have  you  yesterday  for a books  read
  'What kind of books did you read yesterday?'
b. * Wati heb je [ tivoor (een) boeken]j gisteren tj gelezen?
c. % Wati heb je gisteren gelezen [ tivoor (een) boeken]?

The only elements that may intervene between the remnant and the clause-final verb(s) are phrases that compete for the same position: (248a&b) provide examples involving, respectively, a verbal particle, aan, and a PP-predicate, in de kast. The latter example is perhaps slightly marked, but certainly not ungrammatical. Note in this connection that if R-extraction has taken place from the PP-predicate, as in (248b'), the result is fully acceptable.

248
a. Wati heb je de kinderen [ ti voor (een) boeken] aangeraden?
  what  have  you  the children  for a books  prt.-recommended
  'What kind of books did you read aloud to the children?'
b. (?) Wati heb je [ ti voor (een) boeken] in de kast gezet?
  what  have  you  for a books  into the bookcase  put
  'What kind of books did you put into the bookcase?'
b'. Wati heb je er [ ti voor (een) boeken] in gezet?
  what  have  you  there  for a books  into  put
  'What kind of books did you put into it?'

The examples in (249) show that inverting the order of the wat voor remnant and the particle or (the stranded preposition of) the PP-predicate gives rise to an ungrammatical result.

249
a. * Wati heb je de kinderen aan [ tivoor (een) boeken] geraden?
b. * Wati heb je in de kast [ tivoor (een) boeken] gezet?
c. * Wati heb je erin [ ti voor (een) boeken] gezet?
[+]  B.  Subjects

The data involving the nominative subject of the clause are more complex than the data involving the direct object. The following subsections will show that the wat voor split is allowed in passive constructions and clauses involving an unaccusative verb, provided that the subject occupies its base position and not the (derived) subject position. If the construction contains a transitive or an intransitive verb, the split only seems to be possible in expletive constructions.

[+]  1.  Passive constructions

The nominative subject of a passive clause corresponds to the accusative object of its active counterpart. In Dutch, the subject of a passive clause can occupy two positions: either it occupies the position that is normally taken by the direct object, or it occupies the regular subject position of the clause. This can be demonstrated relatively easily by considering the passive of a ditransitive construction: in (250b), the nominative argument het boek follows the indirect object dekinderen, which suggests that it occupies the same position as the direct object in the active construction in (250a); in (250c), on the other hand, it precedes the indirect object, which suggests that it occupies the regular subject position.

250
a. Gisteren heeft Jan de kinderen het boek voorgelezen.
  yesterday  has  Jan the children  the book  read.aloud
  'Jan read the book aloud to the children yesterday.'
b. Gisteren is de kinderen het boek voorgelezen.
  yesterday  has.been  the children  the book  read.aloud
c. Gisteren is het boek de kinderen voorgelezen.
  yesterday  has.been  the book  the children  read.aloud

If the nominative noun phrase in (250b) indeed occupies the same position as the direct object in (250a), it does not come as a surprise that the wat voor split of a subject is possible in (251b); after all, the same thing holds for the direct object in (251a). Given that the wat voor phrase in (250c) is generally assumed to have been moved into the regular subject position, the freezing principle correctly predicts that the wat voor split is impossible in (251c).

251
a. Wati heeft Jan de kinderen gisteren [ ti voor een boek] voorgelezen?
  what  has  Jan the children  yesterday  for a book  read.aloud
b. Wati is de kinderen gisteren [ ti voor een boek] voorgelezen?
  what  has.been  the children yesterday  for a book  read.aloud
c. * Wati is [ti voor een boek]j de kinderen gisteren tj voorgelezen?
  what  has.been  for a book  the children yesterday  read.aloud
[+]  2.  Unaccusative verbs

It has been argued that, just like the subject of a passive construction, the subject of an unaccusative verb is a “derived” subject. Given the discussion of the passive construction in 1 above, this can be readily shown in the case of dyadic unaccusative verbs. If the nominative argument follows the (dative) object, as in (252a), the wat voor split is possible, hence it is plausible to assume that it occupies a position that is comparable to that of a direct object. If it precedes the object, as in (252b), the wat voor split is excluded, which suggests the working of the freezing principle; in other words, the subject has been moved from its original position in (252a) into the regular subject position of the clause.

252
a. Wati zouden hem nou [NPti voor een boeken] bevallen?
  what  would  him  prt  for a books  please
  'What kind of books would please him?'
b. * Wati zouden [NPti voor een boeken]j hem nou tj bevallen?

      If we are dealing with a monadic unaccusative verb, the wat voor split is possible also, but only in the expletive construction. This can be accounted for by assuming that in expletive constructions, the regular subject position is filled by the expletive er, so the nominative argument must occupy its base position in (253a). Example (253b) is ungrammatical since er must be present if the indefinite subject remains in its base position. Example (253c), finally, is ungrammatical since the nominative argument has been moved into the regular position, and hence invokes a violation of the freezing principle.

253
a. Wati zijn er gisteren [NPti voor mensen] aangekomen?
  what are there  yesterday  for people  prt.-arrived
  'What kind of people have arrived yesterday?'
b. *? Wati zijn gisteren [NPti voor mensen] aangekomen?
c. * Wati zijn [NPti voor mensen]j gisteren tj aangekomen?

Note, however, that there is a caveat in order here. In (253) and in the examples below, we abstract away from the fact that expletive er can be dropped if certain adverbial phrases are present. A typical example involves the place adverb daar'there' in (254); see Section 8.1.4 for discussion. The fact that the wat voor remnant is placed after the adverb daar suggests that in this example the indefinite subject also occupies its base position.

254
Wat zijn (er) daar voor mensen aangekomen?
  what  are  there there  for people  prt.-arrived
'What kind of people have arrived there?'

      The wat voor split can be sensitive to the semantic type of the predicate, especially the distinction between stage-level and individual-level predicates. Whereas the former often allow the expletive construction, the latter do not due to the fact that they block an existential reading of the subject noun phrase; see Hartmann (2008: §1.4) for a review of the literature. Therefore, it is not surprising that in a copular construction (which is always an unaccusative construction), the adjectival predicate determines whether the wat voor split is possible or not. A typical stage-level predicate like beschikbaar'available' allows the wat voor split whereas an individual-level predicate like waterdicht'waterproof' does not; see Section 5.1.5.1, sub IIIA) for exceptions. Example (255b) with er is unacceptable because the individual-level predicate waterdicht does not license an existential reading of the noun phrase schoenen'shoes' and is therefore not possible in an expletive construction. Finally, (255b') without er is ungrammatical due to the freezing principle.

255
a. Wati zijn er [NPti voor boeken] beschikbaar.
  what  are  there  for  books  available
  'What kind of books are available?'
b. * Wati zijn er [NPti voor schoenen] waterdicht?
  what  are  there  for  shoes  waterproof
  'What kind of shoes are waterproof?'
b'. * Wati zijn [NPti voor schoenen]jtj waterdicht?
[+]  3.  Intransitive verbs

Den Besten (1985) has claimed that regular intransitive verbs do not allow the wat voor split. It seems, however, that this is an overgeneralization. As with monadic unaccusative verbs, the wat voor split seems possible if expletive er is present; the split in (256a) is at worst slightly marked and certainly gives rise to a much better result than the split in (256b).

256
a. (?) Wati hebben er gisteren [NPti voor (een) jongens] gevochten?
  what  have  there  yesterday  for a boys  fought
  'What kind of boys fought yesterday?'
b. * Wati hebben [NPti voor een jongens]j gisteren tj gevochten?

The contrast in (256) is not really surprising from the perspective of present-day generative grammar, given that there is a growing body of evidence in favor of the claim that the subject of an intransitive clause is not base-generated directly in the regular subject position, but in some more deeply embedded position. The fact that the nominative argument does not occupy the regular subject position in (256a) is also clear from the fact that it follows the adverbial phrase gisteren. If (256b) is indeed derived by moving the subject into the regular subject position of the clause, its unacceptability can be made to follow from the freezing principle.
      Finally, note that it has been suggested that the wat voor split is only possible if the clause contains a verb in clause-final position, especially if a modal verb like zouden in (257a) is present. Although some difference in acceptability between the examples in (257) can perhaps be detected, we think it would be an overstatement to say that (257a) is perfectly well-formed and that (257c) is completely unacceptable: all examples seem acceptable.

257
a. Wat zouden er hier voor een mensen gewoond hebben?
  what  would  there  here  for a people  lived  have
  'What kind of people would have lived here?'
b. Wat hebben er hier voor een mensen gewoond?
  what  have  there  here  for a people  lived
c. Wat wonen er hier voor een mensen?
  what  live  there  here  for a people
[+]  4.  Transitive verbs

What has been said in Subsection 3 regarding the subject of an intransitive verb also holds for the subject of a transitive clause. Although it has been claimed that the wat voor split is excluded for the subject of a transitive verb, this seems an overgeneralization. In (258a), an example is given that seems relatively good.

258
a. Wati hebben er [NPti voor een vogels] je voedertafel bezocht?
  what  have  there  for a birds  your feeding table  visited
  'What kind of birds have visited your feeding table?'
b. ? Wati hebben [NPti voor een vogels] je voedertafel bezocht?

Actually, (258b) is much better than might have been expected, as it seems to involve movement and hence should invoke a freezing effect. However, it may be the case that this example is ambiguous, because a definite direct object often makes it possible to drop the expletive er. This is shown in (259): example (259a) shows that in most varieties of Dutch the interrogative subject wie must be accompanied by the expletive. However, if a definite direct object is present, expletive er is preferably dropped; see Section 8.1.4 for more discussion.

259
a. Wie rookt %(er)?
  who smokes    there
b. Wie rookt (?er) de sigaar?
  who smokes  there  the cigar

So, in order to determine whether (258b) is excluded by the freezing principle, we have to take the placement of adverbs into account: if the subject precedes the adverb, it occupies the regular subject position, and the wat voor split is predicted to be impossible; if it follows the adverb, it is probably in its base position, and the wat voor split is predicted to be possible. As is shown by (260), the subject may actually occupy either position, so we may indeed conclude that (258b) is ambiguous. The judgments on the two examples are more or less as predicted.

260
a. ? Wati hebben gisteren [NPti voor een vogels] je voedertafel bezocht?
  what  have  yesterday  for a birds  your feeding.table  visited
  'What kid of birds visited your feeding table yesterday?'
b. * Wati hebben [NPti voor een vogels]j gisteren tj je voedertafel bezocht?
[+]  C.  Indirect objects

The primed examples in (261) show that wat voor split of nominal indirect objects always leads to a degraded result; note that for some speakers, the primeless examples are also somewhat degraded (a prepositional indirect object seems preferred by most speakers).

261
a. (?) Wat voor een meisje heb je een lolly gegeven?
  what for a girl  have  you  a lollipop  given
  'To what kind of girl did you give a lollipop?'
a'. *? Wat heb je voor een meisje een lolly gegeven?
b. (?) Wat voor een mensen heb je je artikel toegestuurd?
  what kind of people  have  you  your paper  prt.-sent
  'To what kind of people did you send your paper?'
b'. *? Wat heb je voor een mensen je stuk toegestuurd?
[+]  D.  Complements of a preposition

As was shown earlier in (230), repeated here as (262), wat voor split of the complement of a preposition is excluded as well due to the fact that subextraction from a nominal complement of a preposition is generally excluded.

262
a. [PP Op [NP wat voor een bericht]] wacht je?
  for  what for a message  wait you
  'For what kind of message are you waiting?'
b. * Wati wacht je [PP op [NPti voor een bericht]]?

It is interesting to note that the wat voor split differs in this respect from the exclamative wat-construction discussed in Section 1.2.2.1, sub IV. The two (a)-examples in (263) suggest that this construction is similar to the wat voor construction: the fact illustrated in (263a) that wat and its associated noun phrase may precede the finite verb in clause-initial position suggests that the two form a constituent, and the availability of the split pattern in (263a') suggests that wat can be extracted from this constituent by wh-movement. However, this movement analysis of (263a') runs into problems with (263b): since subextraction from a nominal complement of a preposition is normally excluded, the movement analysis wrongly predicts this example to be ungrammatical.

263
a. Wat een hoop boeken heeft hij!
  what  a lot [of] books  has  he
  'What a lot of books he has!'
a'. Wat heeft hij een hoop boeken!
b. Wat beschikt hij [PP over een hoop boeken]!
  what  has  he  a lot [of] books
  'What a lot of books he has at his disposal!'
[+]  E.  Nominal predicates

Wat voor split of a nominal predicate is fully acceptable. This is illustrated in (264).

264
a. Wat voor een jongen is Jan eigenlijk?
  what for a boy  is Jan actually
  'What kind of boy is Jan actually?'
b. Wat is Jan eigenlijk voor een jongen?
[+]  II.  The status of wat: parasitic gaps and intervention effects

In the case of wat voor split, movement of wat does of course not involve movement of an argument but of a part of an argument, viz. the complete wat voor phrase. This has several consequences, which are discussed in this subsection. We start with discussion of so-called parasitic gaps in Subsection A, followed by the discussion of several intervention effects in Subsection B. Finally, we conclude in Subsection C by pointing out a semantic difference between split and unsplit a wat voor phrases.

[+]  A.  Parasitic gaps

If wat is an argument in its own right, it may license a so-called parasitic gap in the infinitival adverbial phrase [zonder ... te lezen] in (265a). The complement of lezen need not be overtly expressed, but can be expressed by a phonetically empty parasitic gap PG, the content of which is identified by the moved wh-phrase (which is indicated by the subscript “i”). In other words, the interpretation of this example is something like “for which x, Jan threw x away without reading x”. As is shown in (265b), a parasitic gap can also be licensed if a wat voor phrase is moved into clause-initial position as a whole.

265
a. Wati gooide Jan [zonder PGi te lezen] ti weg?
  what  threw  Jan without  to read  away
  'What did Jan throw away without reading?'
b. [Wat voor een boek]i gooide Jan [zonder PGi te lezen] ti weg?
  what  for  book  threw  Jan  without  to read  away
  'What kind of book did Jan throw away without reading?'

The N1wat from the wat voor phrase, on the other hand, cannot license such a parasitic gap: it can license neither a parasitic gap with the function of direct object of the infinitival verb lezen (cf. (266a)), nor a parasitic gap that functions as an N1in a wat voor phrase functioning as the direct object of lezen (cf. (266b)). It has been assumed that this is due to the fact that parasitic gaps can be licensed by arguments only.

266
a. * Wati gooide Jan [zonder PGi te lezen] [ ti voor een boeken] weg?
  what  threw  Jan without  to read  for a books  away
b. * Wati gooide Jan [zonder [PGi (voor een tijdschriften)] te lezen] [ ti voor een boeken] weg?
  what  threw  Jan without   for a magazines  to read  for a books  away

      For completeness’ sake, note that, according to some speakers, example (267b) is acceptable as well. If this is really the case, this example is a problem for the earlier claim that scrambling induces a freezing effect. Since it is generally assumed that Dutch parasitic gaps must be licensed by a wh-moved or a scrambled phrase (cf. Bennis & Hoekstra 1984), it would follow that the wat voor phrase in (267b) has been scrambled, and, consequently, a freezing effect is wrongly predicted to arise.

267
a. [Wat voor een boek]i gooide Jan [zonder PGi te lezen] ti weg?
  what  for  book  threw  Jan  without  to read  away
  'What kind of book did Jan throw away without reading?'
b. % Wati gooide Jan [ ti voor een boek]j [zonder PGj te lezen] tj weg?

In this connection it should also be mentioned that Beermann (1997) claims that, in German, one occurrence of wat may bind the gaps in two or more wat voor phrases. Example (268) shows that this is not possible in Dutch. In fact, the examples in (268b&c) show that wat voor split is degraded anyway in these examples; the only fully acceptable option is to move the full subject into clause-initial position.

268
a. * Wati hebben (er) [ ti voor een meisjes] [ ti voor een jongens] gekust?
  what  have  there  for a girls  for a boys  kissed
  Intended meaning: 'What kind of girls kissed what kind of boys?'
b. *? Wati hebben (er) [ ti voor een meisjes] [ watvoor een jongens] gekust?
c. * Wati hebben (er) [ wat voor een meisjes] [ tivoor een jongens] gekust?
d. [Wat voor een meisjes]i hebben ti [ wat voor een jongens] gekust?
[+]  B.  Intervention effects

The previous subsection has shown that the interrogative element wat does not function as an argument; it is only the full wat voor that acts like that. This subsection shows that this conclusion is supported by the so-called intervention effect. Arguments and non-arguments differ in that the latter are more sensitive to certain intervention effects than the former. As is shown in (269), for example, an interrogative direct object can be moved across the negative adverb niet, whereas an interrogative adverbial phrase of manner cannot. Below, we will see that N1wat behaves like a non-argument in the sense that it cannot cross certain adverbs, as a result of which the wat voor split is sensitive to the presence of these adverbs.

269
a. Welke auto heb jij niet gerepareerd?
  which car  have  you  not  repaired
  'Which car didnʼt you repair?'
b. * Hoe heb jij die auto niet gerepareerd?
  how  have  you  that car  not  repaired
  '*How didnʼt you repair that car?'

      The examples in (270) show that time and place adverbs like gisteren'yesterday' and daar'there' do not have any effect on the wat voor split. The split is possible as long as the remnant follows the adverb.

270
a. Wat voor een boeken heeft hij gisteren/daar gelezen?
  what  for  books  has  he  yesterday/there  read
  'What kind of books did he read yesterday/there?'
b. Wat heeft hij gisteren/daar voor een boeken gelezen?
c. *? Wat heeft hij voor een boeken gisteren/daar gelezen?

      The situation is different, however, with manner adverb like zorgvuldig'carefully', modal adverbs like zeker'certainly', frequency adverbs like vaak'often', or the negative adverb niet'not'; the (a)- and (b)-examples in (271) to (273) show that these adverbial phrases allow movement of the complete wat voor, but block the wat voor split. Perhaps the (b)-examples become slightly better if the wat voor remnant precedes the adverbial phrase, as in the (c)-examples, but they still seem to be severely degraded; note that if one were to consider these examples grammatical, a similar problem would arise for the freezing principle, as has been pointed out for (267b).

271
a. Wat voor een boeken heeft hij zorgvuldig gelezen?
  what  for  books  has  he  carefully  read
  'What kind of books did he read carefully?'
b. * Wat heeft hij zorgvuldig voor een boeken gelezen?
c. ?? Wat heeft hij voor een boeken zorgvuldig gelezen?
272
a. Wat voor een boeken heeft hij zeker/vaak gelezen?
  what  for  books  has  he  certainly/often  read